Release 3.0.1 (#2561)

* Bump version
* Update release notes
* Return 3.6 to supported versions in setup.py

.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,48 @@
+---
+name: Bug report
+about: Create a report to help us improve
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+**Describe the bug**
+A clear and concise description of what the bug is.
+
+**To Reproduce**
+Steps to reproduce the behavior:
+
+**Expected behavior**
+A clear and concise description of what you expected to happen.
+
+**Screenshots**
+If applicable, add screenshots to help explain your problem.
+
+**Environment**
+-   Patroni version:
+-   PostgreSQL version:
+-   DCS (and its version):
+
+**Patroni configuration file**
+```
+Please copy&paste your Patroni configuration file here
+```
+
+**patronictl show-config**
+```
+Please copy&paste the output of "patronictl show-config" command here
+```
+
+**Have you checked Patroni logs?**
+Please provide a snippet of Patroni log files here
+
+**Have you checked PostgreSQL logs?**
+Please provide a snippet here
+
+**Have you tried to use GitHub issue search?**
+Maybe there is already a similar issue solved.
+
+
+**Additional context**
+Add any other context about the problem here.

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1 @@
+blank_issues_enabled: false

.github/workflows/install_deps.py

@@ -0,0 +1,145 @@
+import inspect
+import os
+import shutil
+import subprocess
+import stat
+import sys
+import tarfile
+import zipfile
+
+
+def install_requirements(what):
+    old_path = sys.path[:]
+    w = os.path.join(os.getcwd(), os.path.dirname(inspect.getfile(inspect.currentframe())))
+    sys.path.insert(0, os.path.dirname(os.path.dirname(w)))
+    try:
+        from setup import EXTRAS_REQUIRE, read
+    finally:
+        sys.path = old_path
+    requirements = ['mock>=2.0.0', 'flake8', 'pytest', 'pytest-cov'] if what == 'all' else ['behave']
+    requirements += ['coverage']
+    # try to split tests between psycopg2 and psycopg3
+    requirements += ['psycopg[binary]'] if sys.version_info > (3, 7, 0) and\
+        (sys.platform != 'darwin' or what == 'etcd3') else ['psycopg2-binary']
+    for r in read('requirements.txt').split('\n'):
+        r = r.strip()
+        if r != '':
+            extras = {e for e, v in EXTRAS_REQUIRE.items() if v and any(r.startswith(x) for x in v)}
+            if not extras or what == 'all' or what in extras:
+                requirements.append(r)
+
+    subprocess.call([sys.executable, '-m', 'pip', 'install', '--upgrade', 'pip'])
+    subprocess.call([sys.executable, '-m', 'pip', 'install', '--upgrade', 'wheel'])
+    r = subprocess.call([sys.executable, '-m', 'pip', 'install'] + requirements)
+    s = subprocess.call([sys.executable, '-m', 'pip', 'install', '--upgrade', 'setuptools'])
+    return s | r
+
+
+def install_packages(what):
+    from mapping import versions
+
+    packages = {
+        'zookeeper': ['zookeeper', 'zookeeper-bin', 'zookeeperd'],
+        'consul': ['consul'],
+    }
+    packages['exhibitor'] = packages['zookeeper']
+    packages = packages.get(what, [])
+    ver = versions.get(what)
+    if float(ver) >= 15:
+        packages += ['postgresql-{0}-citus-11.2'.format(ver)]
+    subprocess.call(['sudo', 'apt-get', 'update', '-y'])
+    return subprocess.call(['sudo', 'apt-get', 'install', '-y', 'postgresql-' + ver, 'expect-dev'] + packages)
+
+
+def get_file(url, name):
+    try:
+        from urllib.request import urlretrieve
+    except ImportError:
+        from urllib import urlretrieve
+
+    print('Downloading ' + url)
+    urlretrieve(url, name)
+
+
+def untar(archive, name):
+    with tarfile.open(archive) as tar:
+        f = tar.extractfile(name)
+        dest = os.path.basename(name)
+        with open(dest, 'wb') as d:
+            shutil.copyfileobj(f, d)
+            return dest
+
+
+def unzip(archive, name):
+    with zipfile.ZipFile(archive, 'r') as z:
+        name = z.extract(name)
+        dest = os.path.basename(name)
+        shutil.move(name, dest)
+        return dest
+
+
+def unzip_all(archive):
+    print('Extracting ' + archive)
+    with zipfile.ZipFile(archive, 'r') as z:
+        z.extractall()
+
+
+def chmod_755(name):
+    os.chmod(name, stat.S_IRUSR | stat.S_IWUSR | stat.S_IXUSR |
+             stat.S_IRGRP | stat.S_IXGRP | stat.S_IROTH | stat.S_IXOTH)
+
+
+def unpack(archive, name):
+    print('Extracting {0} from {1}'.format(name, archive))
+    func = unzip if archive.endswith('.zip') else untar
+    name = func(archive, name)
+    chmod_755(name)
+    return name
+
+
+def install_etcd():
+    version = os.environ.get('ETCDVERSION', '3.4.23')
+    platform = {'linux2': 'linux', 'win32': 'windows', 'cygwin': 'windows'}.get(sys.platform, sys.platform)
+    dirname = 'etcd-v{0}-{1}-amd64'.format(version, platform)
+    ext = 'tar.gz' if platform == 'linux' else 'zip'
+    name = '{0}.{1}'.format(dirname, ext)
+    url = 'https://github.com/etcd-io/etcd/releases/download/v{0}/{1}'.format(version, name)
+    get_file(url, name)
+    ext = '.exe' if platform == 'windows' else ''
+    return int(unpack(name, '{0}/etcd{1}'.format(dirname, ext)) is None)
+
+
+def install_postgres():
+    version = os.environ.get('PGVERSION', '15.1-1')
+    platform = {'darwin': 'osx', 'win32': 'windows-x64', 'cygwin': 'windows-x64'}[sys.platform]
+    if platform == 'osx':
+        return subprocess.call(['brew', 'install', 'expect', 'postgresql@{0}'.format(version.split('.')[0])])
+    name = 'postgresql-{0}-{1}-binaries.zip'.format(version, platform)
+    get_file('http://get.enterprisedb.com/postgresql/' + name, name)
+    unzip_all(name)
+    bin_dir = os.path.join('pgsql', 'bin')
+    for f in os.listdir(bin_dir):
+        chmod_755(os.path.join(bin_dir, f))
+    return subprocess.call(['pgsql/bin/postgres', '-V'])
+
+
+def main():
+    what = os.environ.get('DCS', sys.argv[1] if len(sys.argv) > 1 else 'all')
+
+    if what != 'all':
+        if sys.platform.startswith('linux'):
+            r = install_packages(what)
+        else:
+            r = install_postgres()
+
+        if r == 0 and what.startswith('etcd'):
+            r = install_etcd()
+
+        if r != 0:
+            return r
+
+    return install_requirements(what)
+
+
+if __name__ == '__main__':
+    sys.exit(main())

.github/workflows/mapping.py

@@ -0,0 +1 @@
+versions = {'etcd': '9.6', 'etcd3': '14', 'consul': '13', 'exhibitor': '12', 'raft': '11', 'kubernetes': '15'}

.github/workflows/release.yaml

@@ -0,0 +1,41 @@
+name: Publish Patroni distributions to PyPI and TestPyPI
+on:
+  push:
+    tags:
+      - 'v[0-9]+.[0-9]+.[0-9]+'
+  release:
+    types:
+      - published
+jobs:
+  build-n-publish:
+    name: Build and publish Patroni distributions to PyPI and TestPyPI
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@master
+
+    - name: Set up Python 3.9
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.9
+
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Build a binary wheel and a source tarball
+      run: python setup.py sdist bdist_wheel
+
+    - name: Publish distribution to Test PyPI
+      if: github.event_name == 'push'
+      uses: pypa/[email protected]
+      with:
+        password: ${{ secrets.TEST_PYPI_API_TOKEN }}
+        repository_url: https://test.pypi.org/legacy/
+
+    - name: Publish distribution to PyPI
+      if: github.event_name == 'release'
+      uses: pypa/[email protected]
+      with:
+        password: ${{ secrets.PYPI_API_TOKEN }}

.github/workflows/run_tests.py

@@ -0,0 +1,48 @@
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+
+
+def main():
+    what = os.environ.get('DCS', sys.argv[1] if len(sys.argv) > 1 else 'all')
+
+    if what == 'all':
+        flake8 = subprocess.call([sys.executable, 'setup.py', 'flake8'])
+        test = subprocess.call([sys.executable, 'setup.py', 'test'])
+        version = '.'.join(map(str, sys.version_info[:2]))
+        shutil.move('.coverage', os.path.join(tempfile.gettempdir(), '.coverage.' + version))
+        return flake8 | test
+    elif what == 'combine':
+        tmp = tempfile.gettempdir()
+        for name in os.listdir(tmp):
+            if name.startswith('.coverage.'):
+                shutil.move(os.path.join(tmp, name), name)
+        return subprocess.call([sys.executable, '-m', 'coverage', 'combine'])
+
+    env = os.environ.copy()
+    if sys.platform.startswith('linux'):
+        from mapping import versions
+
+        version = versions.get(what)
+        path = '/usr/lib/postgresql/{0}/bin:.'.format(version)
+        unbuffer = ['timeout', '900', 'unbuffer']
+    else:
+        if sys.platform == 'darwin':
+            version = os.environ.get('PGVERSION', '15.1-1')
+            path = '/usr/local/opt/postgresql@{0}/bin:.'.format(version.split('.')[0])
+            unbuffer = ['unbuffer']
+        else:
+            path = os.path.abspath(os.path.join('pgsql', 'bin'))
+            unbuffer = []
+    env['PATH'] = path + os.pathsep + env['PATH']
+    env['DCS'] = what
+    if what == 'kubernetes':
+        env['PATRONI_KUBERNETES_CONTEXT'] = 'k3d-k3s-default'
+
+    return subprocess.call(unbuffer + [sys.executable, '-m', 'behave'], env=env)
+
+
+if __name__ == '__main__':
+    sys.exit(main())

.github/workflows/tests.yaml

@@ -0,0 +1,159 @@
+name: Tests
+
+on:
+  pull_request:
+  push:
+    branches:
+    - master
+
+env:
+  CODACY_PROJECT_TOKEN: ${{ secrets.CODACY_PROJECT_TOKEN }}
+  SECRETS_AVAILABLE: ${{ secrets.CODACY_PROJECT_TOKEN != '' }}
+
+jobs:
+  unit:
+    runs-on: ${{ matrix.os }}-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu, windows, macos]
+
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: Set up Python 3.7
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.7
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Set up Python 3.8
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.8
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Set up Python 3.9
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.9
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.10'
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Set up Python 3.11
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.11
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run tests and flake8
+      run: python .github/workflows/run_tests.py
+
+    - name: Combine coverage
+      run: python .github/workflows/run_tests.py combine
+
+    - name: Install coveralls
+      run: python -m pip install coveralls
+
+    - name: Upload Coverage
+      env:
+        COVERALLS_FLAG_NAME: unit-${{ matrix.os }}
+        COVERALLS_PARALLEL: 'true'
+        GITHUB_TOKEN: ${{ secrets.github_token }}
+      run: python -m coveralls --service=github
+
+  behave:
+    runs-on: ${{ matrix.os }}-latest
+    env:
+      DCS: ${{ matrix.dcs }}
+      ETCDVERSION: 3.4.23
+      PGVERSION: 15.1-1  # for windows and macos
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu]
+        python-version: [3.7, '3.10']
+        dcs: [etcd, etcd3, consul, exhibitor, kubernetes, raft]
+        include:
+        - os: macos
+          python-version: 3.8
+          dcs: raft
+        - os: macos
+          python-version: 3.9
+          dcs: etcd
+        - os: macos
+          python-version: 3.11
+          dcs: etcd3
+
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: ${{ matrix.python-version }}
+    - uses: nolar/setup-k3d-k3s@v1
+      if: matrix.dcs == 'kubernetes'
+    - name: Add postgresql and citus apt repo
+      run: |
+        sudo apt-get update -y
+        sudo apt-get install -y wget ca-certificates gnupg debian-archive-keyring apt-transport-https
+        sudo sh -c 'echo "deb http://apt.postgresql.org/pub/repos/apt $(lsb_release -cs)-pgdg main" > /etc/apt/sources.list.d/pgdg.list'
+        sudo sh -c 'wget -qO - https://www.postgresql.org/media/keys/ACCC4CF8.asc | gpg --dearmor > /etc/apt/trusted.gpg.d/apt.postgresql.org.gpg'
+        sudo sh -c 'echo "deb [signed-by=/etc/apt/trusted.gpg.d/citusdata_community.gpg] https://repos.citusdata.com/community/ubuntu/ $(lsb_release -cs) main" > /etc/apt/sources.list.d/citusdata_community.list'
+        sudo sh -c 'wget -qO - https://repos.citusdata.com/community/gpgkey | gpg --dearmor > /etc/apt/trusted.gpg.d/citusdata_community.gpg'
+      if: matrix.os == 'ubuntu'
+    - name: Install dependencies
+      run: python .github/workflows/install_deps.py
+    - name: Run behave tests
+      run: python .github/workflows/run_tests.py
+    - name: Upload logs if behave failed
+      uses: actions/upload-artifact@v3
+      if: failure()
+      with:
+        name: behave-${{ matrix.os }}-${{ matrix.dcs }}-${{ matrix.python-version }}-logs
+        path: |
+          features/output/*_failed/*postgres?.*
+          features/output/*.log
+        if-no-files-found: error
+        retention-days: 5
+    - name: Generate coverage xml report
+      run: python -m coverage xml -o cobertura.xml
+    - name: Upload coverage to Codacy
+      run: bash <(curl -Ls https://coverage.codacy.com/get.sh) report -r cobertura.xml -l Python --partial
+      if: ${{ env.SECRETS_AVAILABLE == 'true' }}
+
+  coveralls-finish:
+    name: Finalize coveralls.io
+    needs: unit
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/setup-python@v4
+    - run: python -m pip install coveralls
+    - run: python -m coveralls --service=github --finish
+      env:
+        GITHUB_TOKEN: ${{ secrets.github_token }}
+
+  codacy-final:
+    name: Finalize Codacy
+    needs: behave
+    runs-on: ubuntu-latest
+    steps:
+    - run: bash <(curl -Ls https://coverage.codacy.com/get.sh) final
+      if: ${{ env.SECRETS_AVAILABLE == 'true' }}

.gitignore

@@ -51,3 +51,9 @@ scm-source.json
 docs/build/
 docs/source/_static/
 docs/source/_templates/
+
+# Pycharm IDE
+.idea/
+
+#VSCode IDE
+.vscode/

.travis.yml

@@ -1,114 +0,0 @@
-sudo: false
-dist: trusty
-language: python
-python:
-- "3.4"  # 2.7 and 3.5 are preinstalled by default
-env:
-  global:
-  - ETCDVERSION=3.0.17 ZKVERSION=3.4.9 CONSULVERSION=0.7.4
-  - PYVERSIONS="2.7 3.4 3.5"
-  matrix:
-  - TEST_SUITE="python setup.py"
-  - DCS="etcd" TEST_SUITE="behave"
-  - DCS="exhibitor" TEST_SUITE="behave"
-  - DCS="consul" TEST_SUITE="behave"
-cache:
-  directories:
-  - $HOME/mycache
-before_cache:
-- |
-  rm -fr $HOME/mycache/python*
-  for pv in $PYVERSIONS; do
-      if [[ $TEST_SUITE != "behave" || $pv != "3.4" ]]; then
-          fpv=$(basename $(readlink $HOME/virtualenv/python${pv}))
-          mv $HOME/virtualenv/${fpv} $HOME/mycache/${fpv}
-      fi
-  done
-install:
-  - |
-    set -e
-
-    if [[ $TEST_SUITE == "behave" ]]; then
-        function get_consul() {
-            CC=~/mycache/consul_${CONSULVERSION}
-            if [[ ! -x $CC ]]; then
-                curl -L https://releases.hashicorp.com/consul/${CONSULVERSION}/consul_${CONSULVERSION}_linux_amd64.zip \
-                    | gunzip > $CC
-                [[ ${PIPESTATUS[0]} == 0 ]] || return 1
-                chmod +x $CC
-            fi
-            ln -s $CC consul
-        }
-
-        function get_etcd() {
-            EC=~/mycache/etcd_${ETCDVERSION}
-            if [[ ! -x $EC ]]; then
-                curl -L https://github.com/coreos/etcd/releases/download/v${ETCDVERSION}/etcd-v${ETCDVERSION}-linux-amd64.tar.gz \
-                    | tar xz -C . --strip=1 --wildcards --no-anchored etcd
-                [[ ${PIPESTATUS[0]} == 0 ]] || return 1
-                mv etcd $EC
-            fi
-            ln -s $EC etcd
-        }
-
-        function get_exhibitor() {
-            ZC=~/mycache/zookeeper-${ZKVERSION}
-            if [[ ! -d $ZC ]]; then
-                curl -L http://www.apache.org/dist/zookeeper/zookeeper-${ZKVERSION}/zookeeper-${ZKVERSION}.tar.gz | tar xz
-                [[ ${PIPESTATUS[0]} == 0 ]] || return 1
-                mv zookeeper-${ZKVERSION}/conf/zoo_sample.cfg zookeeper-${ZKVERSION}/conf/zoo.cfg
-                mv zookeeper-${ZKVERSION} $ZC
-            fi
-            $ZC/bin/zkServer.sh start
-            # following lines are 'emulating' exhibitor REST API
-            while true; do
-                echo -e 'HTTP/1.0 200 OK\nContent-Type: application/json\n\n{"servers":["127.0.0.1"],"port":2181}' \
-                    | nc -l 8181 &> /dev/null
-            done&
-        }
-
-        attempt_num=1
-        until get_${DCS}; do
-            [[ $attempt_num -ge 3 ]] && exit 1
-            echo "Attempt $attempt_num failed! Trying again in $attempt_num seconds..."
-            sleep $(( attempt_num++ ))
-        done
-    fi
-
-    for pv in $PYVERSIONS; do
-        if [[ $TEST_SUITE != "behave" || $pv != "3.4" ]]; then
-            fpv=$(basename $(readlink $HOME/virtualenv/python$pv))
-            if [[ -d ~/mycache/${fpv} ]]; then
-                mv ~/virtualenv/${fpv} ~/virtualenv/${fpv}.bckp
-                mv ~/mycache/${fpv} ~/virtualenv/${fpv}
-            fi
-            source ~/virtualenv/python${pv}/bin/activate
-            # explicitly install all needed python modules to cache them
-            for p in '-r requirements.txt' 'behave codacy-coverage coverage coveralls flake8 mock pytest-cov pytest setuptools'; do
-                pip install $p --upgrade
-            done
-        fi
-    done
-script:
-  - |
-    for pv in $PYVERSIONS; do
-        source ~/virtualenv/python${pv}/bin/activate
-
-        if [[ $TEST_SUITE != "behave" ]]; then
-            echo Running unit tests using python${pv}
-            $TEST_SUITE test
-            $TEST_SUITE flake8
-        elif [[ $pv != "3.4" ]]; then
-            echo Running acceptance tests using python${pv}
-            if ! PATH=.:/usr/lib/postgresql/9.6/bin:$PATH $TEST_SUITE; then
-                # output all log files when tests are failing
-                grep . features/output/*/*postgres?.*
-                exit 1
-            fi
-        fi
-    done
-
-    set +e
-after_success:
-  - coveralls
-  - if [[ $TEST_SUITE != "behave" ]]; then python-codacy-coverage -r coverage.xml; fi

.zappr.yaml

@@ -1,13 +0,0 @@
-# for github.com
-approvals:
-  groups:
-    zalando:
-      minimum: 2
-      from:
-        orgs:
-          - "zalando"
-# team should be valid team id in team service https://teams.auth.zalando.com/api/teams/:id
-X-Zalando-Team: "acid"
-# type should be one of [code, doc, config, tools, secrets]
-# code will be the default value, if X-Zalando-Type is not found in .zappr.yml
-X-Zalando-Type: code

CODEOWNERS

@@ -0,0 +1,2 @@
+# global owners
+*        @CyberDem0n @hughcapet

Dockerfile

@@ -1,44 +1,160 @@
 ## This Dockerfile is meant to aid in the building and debugging patroni whilst developing on your local machine
 ## It has all the necessary components to play/debug with a single node appliance, running etcd
-FROM postgres:9.6
-MAINTAINER Alexander Kukushkin <[email protected]>
+ARG PG_MAJOR=15
+ARG COMPRESS=false
+ARG PGHOME=/home/postgres
+ARG PGDATA=$PGHOME/data
+ARG LC_ALL=C.UTF-8
+ARG LANG=C.UTF-8
 
-RUN echo 'APT::Install-Recommends "0";\nAPT::Install-Suggests "0";' > /etc/apt/apt.conf.d/01norecommend \
+FROM postgres:$PG_MAJOR as builder
+
+ARG PGHOME
+ARG PGDATA
+ARG LC_ALL
+ARG LANG
+
+ENV ETCDVERSION=3.3.13 CONFDVERSION=0.16.0
+
+RUN set -ex \
+    && export DEBIAN_FRONTEND=noninteractive \
+    && echo 'APT::Install-Recommends "0";\nAPT::Install-Suggests "0";' > /etc/apt/apt.conf.d/01norecommend \
     && apt-get update -y \
-    && apt-get upgrade -y \
-    && apt-get install -y curl jq haproxy python-psycopg2 python-yaml python-requests \
-        python-six python-dateutil python-urllib3 python-dnspython \
-        python-pip python-setuptools python-kazoo python-prettytable python-wheel python \
-
-    && pip install python-etcd==0.4.3 python-consul==0.7.0 click tzlocal --upgrade \
-
-    && mkdir -p /home/postgres \
-    && chown postgres:postgres /home/postgres \
-
-    # Clean up
-    && apt-get remove -y python-pip python-setuptools \
+    # postgres:10 is based on debian, which has the patroni package. We will install all required dependencies
+    && apt-cache depends patroni | sed -n -e 's/.*Depends: \(python3-.\+\)$/\1/p' \
+            | grep -Ev '^python3-(sphinx|etcd|consul|kazoo|kubernetes)' \
+            | xargs apt-get install -y vim curl less jq locales haproxy sudo \
+                            python3-etcd python3-kazoo python3-pip busybox \
+                            net-tools iputils-ping --fix-missing \
+    && pip3 install dumb-init \
+\
+    # Cleanup all locales but en_US.UTF-8
+    && find /usr/share/i18n/charmaps/ -type f ! -name UTF-8.gz -delete \
+    && find /usr/share/i18n/locales/ -type f ! -name en_US ! -name en_GB ! -name i18n* ! -name iso14651_t1 ! -name iso14651_t1_common ! -name 'translit_*' -delete \
+    && echo 'en_US.UTF-8 UTF-8' > /usr/share/i18n/SUPPORTED \
+\
+    # Make sure we have a en_US.UTF-8 locale available
+    && localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8 \
+\
+    # haproxy dummy config
+    && echo 'global\n        stats socket /run/haproxy/admin.sock mode 660 level admin' > /etc/haproxy/haproxy.cfg \
+\
+    # vim config
+    && echo 'syntax on\nfiletype plugin indent on\nset mouse-=a\nautocmd FileType yaml setlocal ts=2 sts=2 sw=2 expandtab' > /etc/vim/vimrc.local \
+\
+    # Prepare postgres/patroni/haproxy environment
+    && mkdir -p "$PGHOME/.config/patroni" /patroni /run/haproxy \
+    && ln -s ../../postgres0.yml "$PGHOME/.config/patroni/patronictl.yaml" \
+    && ln -s /patronictl.py /usr/local/bin/patronictl \
+    && sed -i "s|/var/lib/postgresql.*|$PGHOME:/bin/bash|" /etc/passwd \
+    && chown -R postgres:postgres /var/log \
+\
+    # Download etcd
+    && curl -sL "https://github.com/coreos/etcd/releases/download/v$ETCDVERSION/etcd-v$ETCDVERSION-linux-$(dpkg --print-architecture).tar.gz" \
+            | tar xz -C /usr/local/bin --strip=1 --wildcards --no-anchored etcd etcdctl \
+\
+    # Download confd
+    && curl -sL "https://github.com/kelseyhightower/confd/releases/download/v$CONFDVERSION/confd-$CONFDVERSION-linux-$(dpkg --print-architecture)" \
+            > /usr/local/bin/confd && chmod +x /usr/local/bin/confd \
+\
+    # Clean up all useless packages and some files
+    && apt-get purge -y --allow-remove-essential python3-pip gzip bzip2 util-linux e2fsprogs \
+                libmagic1 bsdmainutils login ncurses-bin libmagic-mgc e2fslibs bsdutils \
+                exim4-config gnupg-agent dirmngr libpython2.7-stdlib libpython2.7-minimal \
     && apt-get autoremove -y \
     && apt-get clean -y \
-    && rm -rf /var/lib/apt/lists/* /root/.cache
+    && rm -rf /var/lib/apt/lists/* \
+        /root/.cache \
+        /var/cache/debconf/* \
+        /etc/rc?.d \
+        /etc/systemd \
+        /docker-entrypoint* \
+        /sbin/pam* \
+        /sbin/swap* \
+        /sbin/unix* \
+        /usr/local/bin/gosu \
+        /usr/sbin/[acgipr]* \
+        /usr/sbin/*user* \
+        /usr/share/doc* \
+        /usr/share/man \
+        /usr/share/info \
+        /usr/share/i18n/locales/translit_hangul \
+        /usr/share/locale/?? \
+        /usr/share/locale/??_?? \
+        /usr/share/postgresql/*/man \
+        /usr/share/postgresql-common/pg_wrapper \
+        /usr/share/vim/vim80/doc \
+        /usr/share/vim/vim80/lang \
+        /usr/share/vim/vim80/tutor \
+#        /var/lib/dpkg/info/* \
+    && find /usr/bin -xtype l -delete \
+    && find /var/log -type f -exec truncate --size 0 {} \; \
+    && find /usr/lib/python3/dist-packages -name '*test*' | xargs rm -fr \
+    && find /lib/$(uname -m)-linux-gnu/security -type f ! -name pam_env.so ! -name pam_permit.so ! -name pam_unix.so -delete
 
-ENV ETCDVERSION 3.1.2
-RUN curl -L https://github.com/coreos/etcd/releases/download/v${ETCDVERSION}/etcd-v${ETCDVERSION}-linux-amd64.tar.gz \
-    | tar xz -C /usr/local/bin --strip=1 --wildcards --no-anchored etcd etcdctl
+# perform compression if it is necessary
+ARG COMPRESS
+RUN if [ "$COMPRESS" = "true" ]; then \
+        set -ex \
+        # Allow certain sudo commands from postgres
+        && echo 'postgres ALL=(ALL) NOPASSWD: /bin/tar xpJf /a.tar.xz -C /, /bin/rm /a.tar.xz, /bin/ln -snf dash /bin/sh' >> /etc/sudoers \
+        && ln -snf busybox /bin/sh \
+        && arch=$(uname -m) \
+        && darch=$(uname -m | sed 's/_/-/') \
+        && files="/bin/sh /usr/bin/sudo /usr/lib/sudo/sudoers.so /lib/$arch-linux-gnu/security/pam_*.so" \
+        && libs="$(ldd $files | awk '{print $3;}' | grep '^/' | sort -u) /lib/ld-linux-$darch.so.* /lib/$arch-linux-gnu/ld-linux-$darch.so.* /lib/$arch-linux-gnu/libnsl.so.* /lib/$arch-linux-gnu/libnss_compat.so.* /lib/$arch-linux-gnu/libnss_files.so.*" \
+        && (echo /var/run $files $libs | tr ' ' '\n' && realpath $files $libs) | sort -u | sed 's/^\///' > /exclude \
+        && find /etc/alternatives -xtype l -delete \
+        && save_dirs="usr lib var bin sbin etc/ssl etc/init.d etc/alternatives etc/apt" \
+        && XZ_OPT=-e9v tar -X /exclude -cpJf a.tar.xz $save_dirs \
+        # we call "cat /exclude" to avoid including files from the $save_dirs that are also among
+        # the exceptions listed in the /exclude, as "uniq -u" eliminates all non-unique lines.
+        # By calling "cat /exclude" a second time we guarantee that there will be at least two lines
+        # for each exception and therefore they will be excluded from the output passed to 'rm'.
+        && /bin/busybox sh -c "(find $save_dirs -not -type d && cat /exclude /exclude && echo exclude) | sort | uniq -u | xargs /bin/busybox rm" \
+        && /bin/busybox --install -s \
+        && /bin/busybox sh -c "find $save_dirs -type d -depth -exec rmdir -p {} \; 2> /dev/null"; \
+    fi
 
-ENV CONFDVERSION 0.11.0
-RUN curl -L https://github.com/kelseyhightower/confd/releases/download/v${CONFDVERSION}/confd-${CONFDVERSION}-linux-amd64 > /usr/local/bin/confd \
-    && chmod +x /usr/local/bin/confd
+FROM scratch
+COPY --from=builder / /
 
-ADD patronictl.py patroni.py docker/entrypoint.sh /
-ADD patroni /patroni/
-ADD extras/confd /etc/confd
-RUN ln -s /patronictl.py /usr/local/bin/patronictl
+LABEL maintainer="Alexander Kukushkin <[email protected]>"
 
-### Setting up a simple script that will serve as an entrypoint
-RUN mkdir /data/ && touch /pgpass /patroni.yml \
-    && chown postgres:postgres -R /patroni/ /data/ /pgpass /patroni.yml /etc/haproxy /var/run/ /var/lib/ /var/log/
+ARG PG_MAJOR
+ARG COMPRESS
+ARG PGHOME
+ARG PGDATA
+ARG LC_ALL
+ARG LANG
 
-EXPOSE 2379 5432 8008
+ARG PGBIN=/usr/lib/postgresql/$PG_MAJOR/bin
+
+ENV LC_ALL=$LC_ALL LANG=$LANG EDITOR=/usr/bin/editor
+ENV PGDATA=$PGDATA PATH=$PATH:$PGBIN
+
+COPY patroni /patroni/
+COPY extras/confd/conf.d/haproxy.toml /etc/confd/conf.d/
+COPY extras/confd/templates/haproxy.tmpl /etc/confd/templates/
+COPY patroni*.py docker/entrypoint.sh /
+COPY postgres?.yml $PGHOME/
+
+WORKDIR $PGHOME
+
+RUN sed -i 's/env python/&3/' /patroni*.py \
+    # "fix" patroni configs
+    && sed -i 's/^\(  connect_address:\|  - host\)/#&/' postgres?.yml \
+    && sed -i 's/^  listen: 127.0.0.1/  listen: 0.0.0.0/' postgres?.yml \
+    && sed -i "s|^\(  data_dir: \).*|\1$PGDATA|" postgres?.yml \
+    && sed -i "s|^#\(  bin_dir: \).*|\1$PGBIN|" postgres?.yml \
+    && sed -i 's/^  - encoding: UTF8/  - locale: en_US.UTF-8\n&/' postgres?.yml \
+    && sed -i 's/^\(scope\|name\|etcd\|  host\|  authentication\|  pg_hba\|  parameters\):/#&/' postgres?.yml \
+    && sed -i 's/^    \(replication\|superuser\|rewind\|unix_socket_directories\|\(\(  \)\{0,1\}\(username\|password\)\)\):/#&/' postgres?.yml \
+    && sed -i 's/^      parameters:/      pg_hba:\n      - local all all trust\n      - host replication all all md5\n      - host all all all md5\n&\n        max_connections: 100/'  postgres?.yml \
+    && if [ "$COMPRESS" = "true" ]; then chmod u+s /usr/bin/sudo; fi \
+    && chmod +s /bin/ping \
+    && chown -R postgres:postgres "$PGHOME" /run /etc/haproxy
 
-ENTRYPOINT ["/bin/bash", "/entrypoint.sh"]
 USER postgres
+
+ENTRYPOINT ["/bin/sh", "/entrypoint.sh"]

Dockerfile.citus

@@ -0,0 +1,173 @@
+## This Dockerfile is meant to aid in the building and debugging patroni whilst developing on your local machine
+## It has all the necessary components to play/debug with a single node appliance, running etcd
+ARG PG_MAJOR=15
+ARG COMPRESS=false
+ARG PGHOME=/home/postgres
+ARG PGDATA=$PGHOME/data
+ARG LC_ALL=C.UTF-8
+ARG LANG=C.UTF-8
+
+FROM postgres:$PG_MAJOR as builder
+
+ARG PGHOME
+ARG PGDATA
+ARG LC_ALL
+ARG LANG
+
+ENV ETCDVERSION=3.3.13 CONFDVERSION=0.16.0
+
+RUN set -ex \
+    && export DEBIAN_FRONTEND=noninteractive \
+    && echo 'APT::Install-Recommends "0";\nAPT::Install-Suggests "0";' > /etc/apt/apt.conf.d/01norecommend \
+    && apt-get update -y \
+    # postgres:10 is based on debian, which has the patroni package. We will install all required dependencies
+    && apt-cache depends patroni | sed -n -e 's/.*Depends: \(python3-.\+\)$/\1/p' \
+            | grep -Ev '^python3-(sphinx|etcd|consul|kazoo|kubernetes)' \
+            | xargs apt-get install -y vim curl less jq locales haproxy sudo \
+                            python3-etcd python3-kazoo python3-pip busybox \
+                            net-tools iputils-ping --fix-missing \
+    && curl https://install.citusdata.com/community/deb.sh | bash \
+    && apt-get -y install postgresql-$PG_MAJOR-citus-11.2 \
+    && pip3 install dumb-init \
+\
+    # Cleanup all locales but en_US.UTF-8
+    && find /usr/share/i18n/charmaps/ -type f ! -name UTF-8.gz -delete \
+    && find /usr/share/i18n/locales/ -type f ! -name en_US ! -name en_GB ! -name i18n* ! -name iso14651_t1 ! -name iso14651_t1_common ! -name 'translit_*' -delete \
+    && echo 'en_US.UTF-8 UTF-8' > /usr/share/i18n/SUPPORTED \
+\
+    # Make sure we have a en_US.UTF-8 locale available
+    && localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8 \
+\
+    # haproxy dummy config
+    && echo 'global\n        stats socket /run/haproxy/admin.sock mode 660 level admin' > /etc/haproxy/haproxy.cfg \
+\
+    # vim config
+    && echo 'syntax on\nfiletype plugin indent on\nset mouse-=a\nautocmd FileType yaml setlocal ts=2 sts=2 sw=2 expandtab' > /etc/vim/vimrc.local \
+\
+    # Prepare postgres/patroni/haproxy environment
+    && mkdir -p $PGHOME/.config/patroni /patroni /run/haproxy \
+    && ln -s ../../postgres0.yml $PGHOME/.config/patroni/patronictl.yaml \
+    && ln -s /patronictl.py /usr/local/bin/patronictl \
+    && sed -i "s|/var/lib/postgresql.*|$PGHOME:/bin/bash|" /etc/passwd \
+    && chown -R postgres:postgres /var/log \
+\
+    # Download etcd
+    && curl -sL https://github.com/coreos/etcd/releases/download/v${ETCDVERSION}/etcd-v${ETCDVERSION}-linux-$(dpkg --print-architecture).tar.gz \
+            | tar xz -C /usr/local/bin --strip=1 --wildcards --no-anchored etcd etcdctl \
+\
+    # Download confd
+    && curl -sL https://github.com/kelseyhightower/confd/releases/download/v${CONFDVERSION}/confd-${CONFDVERSION}-linux-$(dpkg --print-architecture) \
+            > /usr/local/bin/confd && chmod +x /usr/local/bin/confd \
+    # Prepare client cert for HAProxy
+    && cat /etc/ssl/private/ssl-cert-snakeoil.key  /etc/ssl/certs/ssl-cert-snakeoil.pem > /etc/ssl/private/ssl-cert-snakeoil.crt \
+\
+    # Clean up all useless packages and some files
+    && apt-get purge -y --allow-remove-essential python3-pip gzip bzip2 util-linux e2fsprogs \
+                libmagic1 bsdmainutils login ncurses-bin libmagic-mgc e2fslibs bsdutils \
+                exim4-config gnupg-agent dirmngr libpython2.7-stdlib libpython2.7-minimal \
+    && apt-get autoremove -y \
+    && apt-get clean -y \
+    && rm -rf /var/lib/apt/lists/* \
+        /root/.cache \
+        /var/cache/debconf/* \
+        /etc/rc?.d \
+        /etc/systemd \
+        /docker-entrypoint* \
+        /sbin/pam* \
+        /sbin/swap* \
+        /sbin/unix* \
+        /usr/local/bin/gosu \
+        /usr/sbin/[acgipr]* \
+        /usr/sbin/*user* \
+        /usr/share/doc* \
+        /usr/share/man \
+        /usr/share/info \
+        /usr/share/i18n/locales/translit_hangul \
+        /usr/share/locale/?? \
+        /usr/share/locale/??_?? \
+        /usr/share/postgresql/*/man \
+        /usr/share/postgresql-common/pg_wrapper \
+        /usr/share/vim/vim80/doc \
+        /usr/share/vim/vim80/lang \
+        /usr/share/vim/vim80/tutor \
+#        /var/lib/dpkg/info/* \
+    && find /usr/bin -xtype l -delete \
+    && find /var/log -type f -exec truncate --size 0 {} \; \
+    && find /usr/lib/python3/dist-packages -name '*test*' | xargs rm -fr \
+    && find /lib/$(uname -m)-linux-gnu/security -type f ! -name pam_env.so ! -name pam_permit.so ! -name pam_unix.so -delete
+
+# perform compression if it is necessary
+ARG COMPRESS
+RUN if [ "$COMPRESS" = "true" ]; then \
+        set -ex \
+        # Allow certain sudo commands from postgres
+        && echo 'postgres ALL=(ALL) NOPASSWD: /bin/tar xpJf /a.tar.xz -C /, /bin/rm /a.tar.xz, /bin/ln -snf dash /bin/sh' >> /etc/sudoers \
+        && ln -snf busybox /bin/sh \
+        && arch=$(uname -m) \
+        && darch=$(uname -m | sed 's/_/-/') \
+        && files="/bin/sh /usr/bin/sudo /usr/lib/sudo/sudoers.so /lib/$arch-linux-gnu/security/pam_*.so" \
+        && libs="$(ldd $files | awk '{print $3;}' | grep '^/' | sort -u) /lib/ld-linux-$darch.so.* /lib/$arch-linux-gnu/ld-linux-$darch.so.* /lib/$arch-linux-gnu/libnsl.so.* /lib/$arch-linux-gnu/libnss_compat.so.* /lib/$arch-linux-gnu/libnss_files.so.*" \
+        && (echo /var/run $files $libs | tr ' ' '\n' && realpath $files $libs) | sort -u | sed 's/^\///' > /exclude \
+        && find /etc/alternatives -xtype l -delete \
+        && save_dirs="usr lib var bin sbin etc/ssl etc/init.d etc/alternatives etc/apt" \
+        && XZ_OPT=-e9v tar -X /exclude -cpJf a.tar.xz $save_dirs \
+        # we call "cat /exclude" to avoid including files from the $save_dirs that are also among
+        # the exceptions listed in the /exclude, as "uniq -u" eliminates all non-unique lines.
+        # By calling "cat /exclude" a second time we guarantee that there will be at least two lines
+        # for each exception and therefore they will be excluded from the output passed to 'rm'.
+        && /bin/busybox sh -c "(find $save_dirs -not -type d && cat /exclude /exclude && echo exclude) | sort | uniq -u | xargs /bin/busybox rm" \
+        && /bin/busybox --install -s \
+        && /bin/busybox sh -c "find $save_dirs -type d -depth -exec rmdir -p {} \; 2> /dev/null"; \
+    else \
+        /bin/busybox --install -s; \
+    fi
+
+FROM scratch
+COPY --from=builder / /
+
+LABEL maintainer="Alexander Kukushkin <[email protected]>"
+
+ARG PG_MAJOR
+ARG COMPRESS
+ARG PGHOME
+ARG PGDATA
+ARG LC_ALL
+ARG LANG
+
+ARG PGBIN=/usr/lib/postgresql/$PG_MAJOR/bin
+
+ENV LC_ALL=$LC_ALL LANG=$LANG EDITOR=/usr/bin/editor
+ENV PGDATA=$PGDATA PATH=$PATH:$PGBIN
+
+COPY patroni /patroni/
+COPY extras/confd/conf.d/haproxy.toml /etc/confd/conf.d/
+COPY extras/confd/templates/haproxy-citus.tmpl /etc/confd/templates/haproxy.tmpl
+COPY patroni*.py docker/entrypoint.sh /
+COPY postgres?.yml $PGHOME/
+
+WORKDIR $PGHOME
+
+RUN sed -i 's/env python/&3/' /patroni*.py \
+    # "fix" patroni configs
+    && sed -i 's/^\(  connect_address:\|  - host\)/#&/' postgres?.yml \
+    && sed -i 's/^  listen: 127.0.0.1/  listen: 0.0.0.0/' postgres?.yml \
+    && sed -i "s|^\(  data_dir: \).*|\1$PGDATA|" postgres?.yml \
+    && sed -i "s|^#\(  bin_dir: \).*|\1$PGBIN|" postgres?.yml \
+    && sed -i 's/^  - encoding: UTF8/  - locale: en_US.UTF-8\n&/' postgres?.yml \
+    && sed -i 's/^scope:/log:\n  loggers:\n    patroni.postgresql.citus: DEBUG\n#&/' postgres?.yml \
+    && sed -i 's/^\(name\|etcd\|  host\|  authentication\|  pg_hba\|  parameters\):/#&/' postgres?.yml \
+    && sed -i 's/^    \(replication\|superuser\|rewind\|unix_socket_directories\|\(\(  \)\{0,1\}\(username\|password\)\)\):/#&/' postgres?.yml \
+    && sed -i 's/^postgresql:/&\n  basebackup:\n    checkpoint: fast/' postgres?.yml \
+    && sed -i 's|^      parameters:|      pg_hba:\n      - local all all trust\n      - hostssl replication all all md5 clientcert=verify-ca\n      - hostssl all all all md5 clientcert=verify-ca\n&\n        max_connections: 100\n        shared_buffers: 16MB\n        ssl: "on"\n        ssl_ca_file: /etc/ssl/certs/ssl-cert-snakeoil.pem\n        ssl_cert_file: /etc/ssl/certs/ssl-cert-snakeoil.pem\n        ssl_key_file: /etc/ssl/private/ssl-cert-snakeoil.key\n        citus.node_conninfo: "sslrootcert=/etc/ssl/certs/ssl-cert-snakeoil.pem sslkey=/etc/ssl/private/ssl-cert-snakeoil.key sslcert=/etc/ssl/certs/ssl-cert-snakeoil.pem sslmode=verify-ca"|'  postgres?.yml \
+    && sed -i 's/^#\(ctl\|  certfile\|  keyfile\)/\1/' postgres?.yml \
+    && sed -i 's|^#  cafile: .*$|  verify_client: required\n  cafile: /etc/ssl/certs/ssl-cert-snakeoil.pem|' postgres?.yml \
+    && sed -i 's|^#  cacert: .*$|  cacert: /etc/ssl/certs/ssl-cert-snakeoil.pem|' postgres?.yml \
+    && sed -i 's/^#  insecure: .*/  insecure: on/' postgres?.yml \
+    # client cert for HAProxy to access Patroni REST API
+    && if [ "$COMPRESS" = "true" ]; then chmod u+s /usr/bin/sudo; fi \
+    && chmod +s /bin/ping \
+    && chown -R postgres:postgres $PGHOME /run /etc/haproxy
+
+USER postgres
+
+ENTRYPOINT ["/bin/sh", "/entrypoint.sh"]

MAINTAINERS

@@ -1,3 +1,2 @@
-Alexander Kukushkin <alexander.kukushkin@zalando.de>
-Feike Steenbergen <feike.steenbergen@zalando.de>
-Oleksii Kliukin <[email protected]>
+Alexander Kukushkin <akukushkin@microsoft.com>
+Polina Bungina <polina.bungina@zalando.de>

README.rst

@@ -1,27 +1,31 @@
-|Build Status| |Coverage Status|
+|Tests Status| |Coverage Status|
 
 Patroni: A Template for PostgreSQL HA with ZooKeeper, etcd or Consul
-------------------------------------------------------------
+--------------------------------------------------------------------
 
 You can find a version of this documentation that is searchable and also easier to navigate at `patroni.readthedocs.io <https://patroni.readthedocs.io>`__.
 
 
 There are many ways to run high availability with PostgreSQL; for a list, see the `PostgreSQL Documentation <https://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling>`__.
 
-Patroni is a template for you to create your own customized, high-availability solution using Python and - for maximum accessibility - a distributed configuration store like `ZooKeeper <https://zookeeper.apache.org/>`__, `etcd <https://github.com/coreos/etcd>`__ or `Consul <https://github.com/hashicorp/consul>`__. Database engineers, DBAs, DevOps engineers, and SREs who are looking to quickly deploy HA PostgreSQL in the datacenter-or anywhere else-will hopefully find it useful.
+Patroni is a template for you to create your own customized, high-availability solution using Python and - for maximum accessibility - a distributed configuration store like `ZooKeeper <https://zookeeper.apache.org/>`__, `etcd <https://github.com/coreos/etcd>`__, `Consul <https://github.com/hashicorp/consul>`__ or `Kubernetes <https://kubernetes.io>`__. Database engineers, DBAs, DevOps engineers, and SREs who are looking to quickly deploy HA PostgreSQL in the datacenter-or anywhere else-will hopefully find it useful.
 
 We call Patroni a "template" because it is far from being a one-size-fits-all or plug-and-play replication system. It will have its own caveats. Use wisely.
 
-**Note to Kubernetes users**: We're currently developing Patroni to be as useful as possible for teams running Kubernetes on top of Google Compute Engine; Patroni can be the HA solution for Postgres in such an environment. To this end, there is a `Helm chart <https://github.com/kubernetes/charts/tree/master/incubator/patroni>`__ that uses Patroni and  `Spilo  <https://github.com/zalando/spilo/>`__ to provision a five-node PostgreSQL HA cluster in a Kubernetes+GCE environment. (The Helm chart deploys Spilo Docker images, not just "bare" Patroni.)
+Currently supported PostgreSQL versions: 9.3 to 15.
+
+**Note to Citus users**: Starting from 3.0 Patroni nicely integrates with the `Citus <https://github.com/citusdata/citus>`__ database extension to Postgres. Please check the `Citus support page <https://github.com/zalando/patroni/blob/master/docs/citus.rst>`__ in the Patroni documentation for more info about how to use Patroni high availability together with a Citus distributed cluster.
+
+**Note to Kubernetes users**: Patroni can run natively on top of Kubernetes. Take a look at the `Kubernetes <https://github.com/zalando/patroni/blob/master/docs/kubernetes.rst>`__ chapter of the Patroni documentation.
 
 .. contents::
     :local:
     :depth: 1
     :backlinks: none
 
-==============
+=================
 How Patroni Works
-==============
+=================
 
 Patroni originated as a fork of `Governor <https://github.com/compose/governor>`__, the project from Compose. It includes plenty of new features.
 
@@ -29,39 +33,110 @@ For an example of a Docker-based deployment with Patroni, see `Spilo <https://gi
 
 For additional background info, see:
 
-* `Elephants on Automatic: HA Clustered PostgreSQL with Helm <https://www.youtube.com/watch?v=CftcVhFMGSY>`_, talk by Josh Berkus and Oleksii Kliukin at KubeCon Berlin 2017 
+* `Elephants on Automatic: HA Clustered PostgreSQL with Helm <https://www.youtube.com/watch?v=CftcVhFMGSY>`_, talk by Josh Berkus and Oleksii Kliukin at KubeCon Berlin 2017
 * `PostgreSQL HA with Kubernetes and Patroni <https://www.youtube.com/watch?v=iruaCgeG7qs>`__, talk by Josh Berkus at KubeCon 2016 (video)
 * `Feb. 2016 Zalando Tech blog post <https://tech.zalando.de/blog/zalandos-patroni-a-template-for-high-availability-postgresql/>`__
 
-================
+==================
 Development Status
-================
+==================
 
 Patroni is in active development and accepts contributions. See our `Contributing <https://github.com/zalando/patroni/blob/master/docs/CONTRIBUTING.rst>`__ section below for more details.
 
 We report new releases information `here <https://github.com/zalando/patroni/releases>`__.
 
-===========================
-Technical Requirements/Installation
-===========================
+=========
+Community
+=========
 
-**For Mac**
+There are two places to connect with the Patroni community: `on github <https://github.com/zalando/patroni>`__, via Issues and PRs, and on channel `#patroni <https://postgresteam.slack.com/archives/C9XPYG92A>`__ in the `PostgreSQL Slack <https://postgresteam.slack.com/>`__.  If you're using Patroni, or just interested, please join us.
+
+===================================
+Technical Requirements/Installation
+===================================
+
+**Pre-requirements for Mac OS**
 
 To install requirements on a Mac, run the following:
 
 ::
 
     brew install postgresql etcd haproxy libyaml python
-    pip install psycopg2 pyyaml
 
-===================
+**Psycopg**
+
+Starting from `psycopg2-2.8 <http://initd.org/psycopg/articles/2019/04/04/psycopg-28-released/>`__ the binary version of psycopg2 will no longer be installed by default. Installing it from the source code requires C compiler and postgres+python dev packages.
+Since in the python world it is not possible to specify dependency as ``psycopg2 OR psycopg2-binary`` you will have to decide how to install it.
+
+There are a few options available:
+
+1. Use the package manager from your distro
+
+::
+
+    sudo apt-get install python-psycopg2   # install python2 psycopg2 module on Debian/Ubuntu
+    sudo apt-get install python3-psycopg2  # install python3 psycopg2 module on Debian/Ubuntu
+    sudo yum install python-psycopg2       # install python2 psycopg2 on RedHat/Fedora/CentOS
+
+2. Install psycopg2 from the binary package
+
+::
+
+    pip install psycopg2-binary
+
+3. Install psycopg2 from source
+
+::
+
+    pip install psycopg2>=2.5.4
+
+4. Use psycopg 3.0 instead of psycopg2
+
+::
+
+    pip install psycopg[binary]
+
+**General installation for pip**
+
+Patroni can be installed with pip:
+
+::
+
+    pip install patroni[dependencies]
+
+where dependencies can be either empty, or consist of one or more of the following:
+
+etcd or etcd3
+    `python-etcd` module in order to use Etcd as DCS
+consul
+    `python-consul` module in order to use Consul as DCS
+zookeeper
+    `kazoo` module in order to use Zookeeper as DCS
+exhibitor
+    `kazoo` module in order to use Exhibitor as DCS (same dependencies as for Zookeeper)
+kubernetes
+    `kubernetes` module in order to use Kubernetes as DCS in Patroni
+raft
+    `pysyncobj` module in order to use python Raft implementation as DCS
+aws
+    `boto` in order to use AWS callbacks
+
+For example, the command in order to install Patroni together with dependencies for Etcd as a DCS and AWS callbacks is:
+
+::
+
+    pip install patroni[etcd,aws]
+
+Note that external tools to call in the replica creation or custom bootstrap scripts (i.e. WAL-E) should be installed independently of Patroni.
+
+=======================
 Running and Configuring
-===================
+=======================
 
 To get started, do the following from different terminals:
 ::
 
-    > etcd --data-dir=data/etcd
+    > etcd --data-dir=data/etcd --enable-v2=true
     > ./patroni.py postgres0.yml
     > ./patroni.py postgres1.yml
 
@@ -80,9 +155,9 @@ run:
 
     > psql --host 127.0.0.1 --port 5000 postgres
 
-===============
+==================
 YAML Configuration
-===============
+==================
 
 Go `here <https://github.com/zalando/patroni/blob/master/docs/SETTINGS.rst>`__ for comprehensive information about settings for etcd, consul, and ZooKeeper. And for an example, see `postgres0.yml <https://github.com/zalando/patroni/blob/master/postgres0.yml>`__.
 
@@ -92,19 +167,19 @@ Environment Configuration
 
 Go `here <https://github.com/zalando/patroni/blob/master/docs/ENVIRONMENT.rst>`__ for comprehensive information about configuring(overriding) settings via environment variables.
 
-===============
+===================
 Replication Choices
-===============
+===================
 
 Patroni uses Postgres' streaming replication, which is asynchronous by default. Patroni's asynchronous replication configuration allows for ``maximum_lag_on_failover`` settings. This setting ensures failover will not occur if a follower is more than a certain number of bytes behind the leader. This setting should be increased or decreased based on business requirements. It's also possible to use synchronous replication for better durability guarantees. See `replication modes documentation <https://github.com/zalando/patroni/blob/master/docs/replication_modes.rst>`__ for details.
 
-===============================
+======================================
 Applications Should Not Use Superusers
-===============================
+======================================
 
 When connecting from an application, always use a non-superuser. Patroni requires access to the database to function properly. By using a superuser from an application, you can potentially use the entire connection pool, including the connections reserved for superusers, with the ``superuser_reserved_connections`` setting. If Patroni cannot access the Primary because the connection pool is full, behavior will be undesirable.
 
-.. |Build Status| image:: https://travis-ci.org/zalando/patroni.svg?branch=master
-   :target: https://travis-ci.org/zalando/patroni
+.. |Tests Status| image:: https://github.com/zalando/patroni/actions/workflows/tests.yaml/badge.svg
+   :target: https://github.com/zalando/patroni/actions/workflows/tests.yaml?query=branch%3Amaster
 .. |Coverage Status| image:: https://coveralls.io/repos/zalando/patroni/badge.svg?branch=master
-   :target: https://coveralls.io/r/zalando/patroni?branch=master
+   :target: https://coveralls.io/github/zalando/patroni?branch=master

TODO.md

@@ -1,12 +0,0 @@
-Failover
-========
-- When determining who should become master, include the minor version of PostgreSQL in the decision.
-
-Configuration
-==============
-- Provide a way to change pg_hba.conf of a running cluster on the Patroni level, without changing individual nodes.
-- Provide hooks to store and retrieve cluster-wide passwords without exposing them in a plain-text form to unauthorized users.
-
-Documentation
-==============
-- Document how to run cascading replication and possibly initialize the cluster without an access to the master node.

docker-compose-citus.yml

@@ -0,0 +1,139 @@
+# docker compose file for running a Citus cluster
+# with 3-node etcd v3 cluster as the DCS and one haproxy node.
+# The Citus cluster has a coordinator (3 nodes)
+# and two worker clusters (2 nodes).
+#
+# Before starting it up you need to build the docker image:
+# $ docker build -f Dockerfile.citus -t patroni-citus .
+# The cluster could be started as:
+# $ docker-compose -f docker-compose-citus.yml up -d
+# You can read more about it in the:
+# https://github.com/zalando/patroni/blob/master/docker/README.md#citus-cluster
+version: "2"
+
+networks:
+    demo:
+
+services:
+    etcd1: &etcd
+        image: patroni-citus
+        networks: [ demo ]
+        environment:
+            ETCDCTL_API: 3
+            ETCD_LISTEN_PEER_URLS: http://0.0.0.0:2380
+            ETCD_LISTEN_CLIENT_URLS: http://0.0.0.0:2379
+            ETCD_INITIAL_CLUSTER: etcd1=http://etcd1:2380,etcd2=http://etcd2:2380,etcd3=http://etcd3:2380
+            ETCD_INITIAL_CLUSTER_STATE: new
+            ETCD_INITIAL_CLUSTER_TOKEN: tutorial
+        container_name: demo-etcd1
+        hostname: etcd1
+        command: etcd -name etcd1 -initial-advertise-peer-urls http://etcd1:2380
+
+    etcd2:
+        <<: *etcd
+        container_name: demo-etcd2
+        hostname: etcd2
+        command: etcd -name etcd2 -initial-advertise-peer-urls http://etcd2:2380
+
+    etcd3:
+        <<: *etcd
+        container_name: demo-etcd3
+        hostname: etcd3
+        command: etcd -name etcd3 -initial-advertise-peer-urls http://etcd3:2380
+
+    haproxy:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: haproxy
+        container_name: demo-haproxy
+        ports:
+            - "5000:5000"  # Access to the coorinator primary
+            - "5001:5001"  # Load-balancing across workers primaries
+        command: haproxy
+        environment: &haproxy_env
+            ETCDCTL_API: 3
+            ETCDCTL_ENDPOINTS: http://etcd1:2379,http://etcd2:2379,http://etcd3:2379
+            PATRONI_ETCD3_HOSTS: "'etcd1:2379','etcd2:2379','etcd3:2379'"
+            PATRONI_SCOPE: demo
+            PATRONI_CITUS_GROUP: 0
+            PATRONI_CITUS_DATABASE: citus
+            PGSSLMODE: verify-ca
+            PGSSLKEY: /etc/ssl/private/ssl-cert-snakeoil.key
+            PGSSLCERT: /etc/ssl/certs/ssl-cert-snakeoil.pem
+            PGSSLROOTCERT: /etc/ssl/certs/ssl-cert-snakeoil.pem
+
+    coord1:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: coord1
+        container_name: demo-coord1
+        environment: &coord_env
+            <<: *haproxy_env
+            PATRONI_NAME: coord1
+            PATRONI_CITUS_GROUP: 0
+
+    coord2:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: coord2
+        container_name: demo-coord2
+        environment:
+            <<: *coord_env
+            PATRONI_NAME: coord2
+
+    coord3:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: coord3
+        container_name: demo-coord3
+        environment:
+            <<: *coord_env
+            PATRONI_NAME: coord3
+
+
+    work1-1:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: work1-1
+        container_name: demo-work1-1
+        environment: &work1_env
+            <<: *haproxy_env
+            PATRONI_NAME: work1-1
+            PATRONI_CITUS_GROUP: 1
+
+    work1-2:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: work1-2
+        container_name: demo-work1-2
+        environment:
+            <<: *work1_env
+            PATRONI_NAME: work1-2
+
+
+    work2-1:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: work2-1
+        container_name: demo-work2-1
+        environment: &work2_env
+            <<: *haproxy_env
+            PATRONI_NAME: work2-1
+            PATRONI_CITUS_GROUP: 2
+
+    work2-2:
+        image: patroni-citus
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: work2-2
+        container_name: demo-work2-2
+        environment:
+            <<: *work2_env
+            PATRONI_NAME: work2-2

docker-compose.yml

@@ -1,58 +1,84 @@
 # docker compose file for running a 3-node PostgreSQL cluster
-# with etcd as the SIS
+# with 3-node etcd cluster as the DCS and one haproxy node
+#
+# requires a patroni image build from the Dockerfile:
+# $ docker build -t patroni .
+# The cluster could be started as:
+# $ docker-compose up -d
+# You can read more about it in the:
+# https://github.com/zalando/patroni/blob/master/docker/README.md
+version: "2"
 
-patroni_etcd:
-    container_name: patroni_etcd
-    image: patroni
-    command: --etcd
+networks:
+    demo:
 
-dbnode1:
-    image: patroni
-    hostname: dbnode1
-    links:
-        - patroni_etcd:patroni_etcd
-    volumes:
-        - ./patroni:/patroni
-    env_file: docker/patroni-secrets.env
-    environment:
-        PATRONI_ETCD_HOST: patroni_etcd:2379
-        PATRONI_NAME: dbnode1
-        PATRONI_SCOPE: testcluster
+services:
+    etcd1: &etcd
+        image: patroni
+        networks: [ demo ]
+        environment:
+            ETCD_LISTEN_PEER_URLS: http://0.0.0.0:2380
+            ETCD_LISTEN_CLIENT_URLS: http://0.0.0.0:2379
+            ETCD_INITIAL_CLUSTER: etcd1=http://etcd1:2380,etcd2=http://etcd2:2380,etcd3=http://etcd3:2380
+            ETCD_INITIAL_CLUSTER_STATE: new
+            ETCD_INITIAL_CLUSTER_TOKEN: tutorial
+        container_name: demo-etcd1
+        hostname: etcd1
+        command: etcd -name etcd1 -initial-advertise-peer-urls http://etcd1:2380
 
-dbnode2:
-    image: patroni
-    hostname: dbnode2
-    links:
-        - patroni_etcd:patroni_etcd
-    volumes:
-        - ./patroni:/patroni
-    env_file: docker/patroni-secrets.env
-    environment:
-        PATRONI_ETCD_HOST: patroni_etcd:2379
-        PATRONI_NAME: dbnode2
-        PATRONI_SCOPE: testcluster
+    etcd2:
+        <<: *etcd
+        container_name: demo-etcd2
+        hostname: etcd2
+        command: etcd -name etcd2 -initial-advertise-peer-urls http://etcd2:2380
 
-dbnode3:
-    image: patroni
-    hostname: dbnode3
-    links:
-        - patroni_etcd:patroni_etcd
-    volumes:
-        - ./patroni:/patroni
-    env_file: docker/patroni-secrets.env
-    environment:
-        PATRONI_ETCD_HOST: patroni_etcd:2379
-        PATRONI_NAME: dbnode3
-        PATRONI_SCOPE: testcluster
+    etcd3:
+        <<: *etcd
+        container_name: demo-etcd3
+        hostname: etcd3
+        command: etcd -name etcd3 -initial-advertise-peer-urls http://etcd3:2380
 
-haproxy:
-    image: patroni
-    links:
-        - patroni_etcd:patroni_etcd
-    ports:
-        - "5000:5000"
-        - "5001:5001"
-    environment:
-        PATRONI_ETCD_HOST: patroni_etcd:2379
-        PATRONI_SCOPE: testcluster
-    command: --confd
+    haproxy:
+        image: patroni
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: haproxy
+        container_name: demo-haproxy
+        ports:
+            - "5000:5000"
+            - "5001:5001"
+        command: haproxy
+        environment: &haproxy_env
+            ETCDCTL_ENDPOINTS: http://etcd1:2379,http://etcd2:2379,http://etcd3:2379
+            PATRONI_ETCD3_HOSTS: "'etcd1:2379','etcd2:2379','etcd3:2379'"
+            PATRONI_SCOPE: demo
+
+    patroni1:
+        image: patroni
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: patroni1
+        container_name: demo-patroni1
+        environment:
+            <<: *haproxy_env
+            PATRONI_NAME: patroni1
+
+    patroni2:
+        image: patroni
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: patroni2
+        container_name: demo-patroni2
+        environment:
+            <<: *haproxy_env
+            PATRONI_NAME: patroni2
+
+    patroni3:
+        image: patroni
+        networks: [ demo ]
+        env_file: docker/patroni.env
+        hostname: patroni3
+        container_name: demo-patroni3
+        environment:
+            <<: *haproxy_env
+            PATRONI_NAME: patroni3

docker/README.md

@@ -1,47 +1,306 @@
-# Patroni Dockerfile
-You can run Patroni in a docker container using this Dockerfile, or by using one of the Docker image at
+# Dockerfile and Dockerfile.citus
+You can run Patroni in a docker container using these Dockerfiles
 
-    https://registry.opensource.zalan.do/v1/repositories/acid/patroni/tags
+They are meant in aiding development of Patroni and quick testing of features and not a production-worthy!
 
-This Dockerfile is meant in aiding development of Patroni and quick testing of features. It is not a production-worthy
-Dockerfile
+    docker build -t patroni .
+    docker build -f Dockerfile.citus -t patroni-citus .
 
 # Examples
 
 ## Standalone Patroni
 
-    docker run -d registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT
+    docker run -d patroni
 
-## Multiple Patroni's communicating with a standalone etcd inside Docker
+## Three-node Patroni cluster
 
-Basically what you would do would be: 
-
-* Run 1 container which provides etcd
-
-        docker run -d <IMAGE> --etcd-only
-
-* Run n containers running Patroni, passing the `--etcd` option to the `docker run` command
-
-        docker run -d <IMAGE> --etcd=<IP FROM etcd CONTAINER:PORT>
-
-To automate this you can run the following script:
-
-    dev_patroni_cluster.sh [OPTIONS]
-
-    Options:
-
-        --image     IMAGE    The Docker image to use for the cluster
-        --members   INT      The number of members for the cluster
-        --name      NAME     The name of the new cluster
+In addition to three Patroni containers the stack starts three containers with etcd (forming a three-node cluster), and one container with haproxy.
+The haproxy listens on ports 5000 (connects to the primary) and 5001 (does load-balancing between healthy standbys).
 
 Example session:
 
-    $ ./dev_patroni_cluster.sh --image registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT --members=2 --name=bravo
-    The etcd container is 6be871a11cb373406ca5ea1c6b39e1.0-SNAPSHOTfdde9fb1d6177212d6ad0c0d1bd9b563, ip=172.17.1.24
-    Started Patroni container 67e611f2eca7c40f9e6e0e24a4a8f2cba7e3e56d22a420e15ab9240a37a9d7a4, ip=172.17.1.25
-    Started Patroni container 47dd12ae635ab83b039f5889e250048b606ed5e48e3650b69e365e7e1d4acbcf, ip=172.17.1.26
+    $ docker-compose up -d
+    Creating demo-haproxy ...
+    Creating demo-patroni2 ...
+    Creating demo-patroni1 ...
+    Creating demo-patroni3 ...
+    Creating demo-etcd2 ...
+    Creating demo-etcd1 ...
+    Creating demo-etcd3 ...
+    Creating demo-haproxy
+    Creating demo-patroni2
+    Creating demo-patroni1
+    Creating demo-patroni3
+    Creating demo-etcd1
+    Creating demo-etcd2
+    Creating demo-etcd2 ... done
+
     $ docker ps
-    CONTAINER ID        IMAGE                                         COMMAND                CREATED             STATUS              PORTS                          NAMES
-    47dd12ae635a        registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT   "/bin/bash /entrypoi   10 seconds ago      Up 8 seconds        4001/tcp, 5432/tcp, 2380/tcp   bravo_OR64g8bx
-    67e611f2eca7        registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT   "/bin/bash /entrypoi   11 seconds ago      Up 10 seconds       2380/tcp, 4001/tcp, 5432/tcp   bravo_si9no8iz
-    6be871a11cb3        registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT   "/bin/bash /entrypoi   12 seconds ago      Up 10 seconds       4001/tcp, 5432/tcp, 2380/tcp   bravo_etcd
+    CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS                              NAMES
+    5b7a90b4cfbf        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 27 seconds                                          demo-etcd2
+    e30eea5222f2        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 27 seconds                                          demo-etcd1
+    83bcf3cb208f        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 27 seconds                                          demo-etcd3
+    922532c56e7d        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 28 seconds                                          demo-patroni3
+    14f875e445f3        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 28 seconds                                          demo-patroni2
+    110d1073b383        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 28 seconds                                          demo-patroni1
+    5af5e6e36028        patroni             "/bin/sh /entrypoint…"   29 seconds ago      Up 28 seconds       0.0.0.0:5000-5001->5000-5001/tcp   demo-haproxy
+
+    $ docker logs demo-patroni1
+    2019-02-20 08:19:32,714 INFO: Failed to import patroni.dcs.consul
+    2019-02-20 08:19:32,737 INFO: Selected new etcd server http://etcd3:2379
+    2019-02-20 08:19:35,140 INFO: Lock owner: None; I am patroni1
+    2019-02-20 08:19:35,174 INFO: trying to bootstrap a new cluster
+    ...
+    2019-02-20 08:19:39,310 INFO: postmaster pid=37
+    2019-02-20 08:19:39.314 UTC [37] LOG:  listening on IPv4 address "0.0.0.0", port 5432
+    2019-02-20 08:19:39.321 UTC [37] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
+    2019-02-20 08:19:39.353 UTC [39] LOG:  database system was shut down at 2019-02-20 08:19:36 UTC
+    2019-02-20 08:19:39.354 UTC [40] FATAL:  the database system is starting up
+    localhost:5432 - rejecting connections
+    2019-02-20 08:19:39.369 UTC [37] LOG:  database system is ready to accept connections
+    localhost:5432 - accepting connections
+    2019-02-20 08:19:39,383 INFO: establishing a new patroni connection to the postgres cluster
+    2019-02-20 08:19:39,408 INFO: running post_bootstrap
+    2019-02-20 08:19:39,432 WARNING: Could not activate Linux watchdog device: "Can't open watchdog device: [Errno 2] No such file or directory: '/dev/watchdog'"
+    2019-02-20 08:19:39,515 INFO: initialized a new cluster
+    2019-02-20 08:19:49,424 INFO: Lock owner: patroni1; I am patroni1
+    2019-02-20 08:19:49,447 INFO: Lock owner: patroni1; I am patroni1
+    2019-02-20 08:19:49,480 INFO: no action.  i am the leader with the lock
+    2019-02-20 08:19:59,422 INFO: Lock owner: patroni1; I am patroni1
+
+    $ docker exec -ti demo-patroni1 bash
+    postgres@patroni1:~$ patronictl list
+    +---------+----------+------------+--------+---------+----+-----------+
+    | Cluster |  Member  |    Host    |  Role  |  State  | TL | Lag in MB |
+    +---------+----------+------------+--------+---------+----+-----------+
+    |   demo  | patroni1 | 172.22.0.3 | Leader | running |  1 |         0 |
+    |   demo  | patroni2 | 172.22.0.7 |        | running |  1 |         0 |
+    |   demo  | patroni3 | 172.22.0.4 |        | running |  1 |         0 |
+    +---------+----------+------------+--------+---------+----+-----------+
+
+    postgres@patroni1:~$ etcdctl ls --recursive --sort -p /service/demo
+    /service/demo/config
+    /service/demo/initialize
+    /service/demo/leader
+    /service/demo/members/
+    /service/demo/members/patroni1
+    /service/demo/members/patroni2
+    /service/demo/members/patroni3
+    /service/demo/optime/
+    /service/demo/optime/leader
+
+    postgres@patroni1:~$ etcdctl member list
+    1bab629f01fa9065: name=etcd3 peerURLs=http://etcd3:2380 clientURLs=http://etcd3:2379 isLeader=false
+    8ecb6af518d241cc: name=etcd2 peerURLs=http://etcd2:2380 clientURLs=http://etcd2:2379 isLeader=true
+    b2e169fcb8a34028: name=etcd1 peerURLs=http://etcd1:2380 clientURLs=http://etcd1:2379 isLeader=false
+    postgres@patroni1:~$ exit
+
+    $ docker exec -ti demo-haproxy bash
+    postgres@haproxy:~$ psql -h localhost -p 5000 -U postgres -W
+    Password: postgres
+    psql (11.2 (Ubuntu 11.2-1.pgdg18.04+1), server 10.7 (Debian 10.7-1.pgdg90+1))
+    Type "help" for help.
+
+    localhost/postgres=# select pg_is_in_recovery();
+     pg_is_in_recovery
+    ───────────────────
+     f
+    (1 row)
+
+    localhost/postgres=# \q
+
+    $postgres@haproxy:~ psql -h localhost -p 5001 -U postgres -W
+    Password: postgres
+    psql (11.2 (Ubuntu 11.2-1.pgdg18.04+1), server 10.7 (Debian 10.7-1.pgdg90+1))
+    Type "help" for help.
+
+    localhost/postgres=# select pg_is_in_recovery();
+     pg_is_in_recovery
+    ───────────────────
+     t
+    (1 row)
+
+## Citus cluster
+
+The stack starts three containers with etcd (forming a three-node etcd cluster), seven containers with Patroni+PostgreSQL+Citus (three coordinator nodes, and two worker clusters with two nodes each), and one container with haproxy.
+The haproxy listens on ports 5000 (connects to the coordinator primary) and 5001 (does load-balancing between worker primary nodes).
+
+Example session:
+
+    $ docker-compose -f docker-compose-citus.yml up -d
+    Creating demo-work2-1 ... done
+    Creating demo-work1-1 ... done
+    Creating demo-etcd2   ... done
+    Creating demo-etcd1   ... done
+    Creating demo-coord3  ... done
+    Creating demo-etcd3   ... done
+    Creating demo-coord1  ... done
+    Creating demo-haproxy ... done
+    Creating demo-work2-2 ... done
+    Creating demo-coord2  ... done
+    Creating demo-work1-2 ... done
+
+    $ docker ps
+    CONTAINER ID   IMAGE                  COMMAND                  CREATED         STATUS         PORTS                              NAMES
+    852d8885a612   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 3 seconds                                      demo-coord3
+    cdd692f947ab   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 3 seconds                                      demo-work1-2
+    9f4e340b36da   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 3 seconds                                      demo-etcd3
+    d69c129a960a   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 4 seconds                                      demo-etcd1
+    c5849689b8cd   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 4 seconds                                      demo-coord1
+    c9d72bd6217d   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 3 seconds                                      demo-work2-1
+    24b1b43efa05   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 4 seconds                                      demo-coord2
+    cb0cc2b4ca0a   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 3 seconds                                      demo-work2-2
+    9796c6b8aad5   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 5 seconds                                      demo-work1-1
+    8baccd74dcae   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 4 seconds                                      demo-etcd2
+    353ec62a0187   patroni-citus          "/bin/sh /entrypoint…"   6 seconds ago   Up 4 seconds   0.0.0.0:5000-5001->5000-5001/tcp   demo-haproxy
+
+    $ docker logs demo-coord1
+    2023-01-05 15:09:31,295 INFO: Selected new etcd server http://172.27.0.4:2379
+    2023-01-05 15:09:31,388 INFO: Lock owner: None; I am coord1
+    2023-01-05 15:09:31,501 INFO: trying to bootstrap a new cluster
+    ...
+    2023-01-05 15:09:45,096 INFO: postmaster pid=39
+    localhost:5432 - no response
+    2023-01-05 15:09:45.137 UTC [39] LOG:  starting PostgreSQL 15.1 (Debian 15.1-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
+    2023-01-05 15:09:45.137 UTC [39] LOG:  listening on IPv4 address "0.0.0.0", port 5432
+    2023-01-05 15:09:45.152 UTC [39] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
+    2023-01-05 15:09:45.177 UTC [43] LOG:  database system was shut down at 2023-01-05 15:09:32 UTC
+    2023-01-05 15:09:45.193 UTC [39] LOG:  database system is ready to accept connections
+    localhost:5432 - accepting connections
+    localhost:5432 - accepting connections
+    2023-01-05 15:09:46,139 INFO: establishing a new patroni connection to the postgres cluster
+    2023-01-05 15:09:46,208 INFO: running post_bootstrap
+    2023-01-05 15:09:47.209 UTC [55] LOG:  starting maintenance daemon on database 16386 user 10
+    2023-01-05 15:09:47.209 UTC [55] CONTEXT:  Citus maintenance daemon for database 16386 user 10
+    2023-01-05 15:09:47,215 WARNING: Could not activate Linux watchdog device: "Can't open watchdog device: [Errno 2] No such file or directory: '/dev/watchdog'"
+    2023-01-05 15:09:47.446 UTC [41] LOG:  checkpoint starting: immediate force wait
+    2023-01-05 15:09:47,466 INFO: initialized a new cluster
+    2023-01-05 15:09:47,594 DEBUG: query(SELECT nodeid, groupid, nodename, nodeport, noderole FROM pg_catalog.pg_dist_node WHERE noderole = 'primary', ())
+    2023-01-05 15:09:47,594 INFO: establishing a new patroni connection to the postgres cluster
+    2023-01-05 15:09:47,467 INFO: Lock owner: coord1; I am coord1
+    2023-01-05 15:09:47,613 DEBUG: query(SELECT pg_catalog.citus_set_coordinator_host(%s, %s, 'primary', 'default'), ('172.27.0.6', 5432))
+    2023-01-05 15:09:47,924 INFO: no action. I am (coord1), the leader with the lock
+    2023-01-05 15:09:51.282 UTC [41] LOG:  checkpoint complete: wrote 1086 buffers (53.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.029 s, sync=3.746 s, total=3.837 s; sync files=280, longest=0.028 s, average=0.014 s; distance=8965 kB, estimate=8965 kB
+    2023-01-05 15:09:51.283 UTC [41] LOG:  checkpoint starting: immediate force wait
+    2023-01-05 15:09:51.495 UTC [41] LOG:  checkpoint complete: wrote 18 buffers (0.9%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.044 s, sync=0.091 s, total=0.212 s; sync files=15, longest=0.015 s, average=0.007 s; distance=67 kB, estimate=8076 kB
+    2023-01-05 15:09:57,467 INFO: Lock owner: coord1; I am coord1
+    2023-01-05 15:09:57,569 INFO: Assigning synchronous standby status to ['coord3']
+    server signaled
+    2023-01-05 15:09:57.574 UTC [39] LOG:  received SIGHUP, reloading configuration files
+    2023-01-05 15:09:57.580 UTC [39] LOG:  parameter "synchronous_standby_names" changed to "coord3"
+    2023-01-05 15:09:59,637 INFO: Synchronous standby status assigned to ['coord3']
+    2023-01-05 15:09:59,638 DEBUG: query(SELECT pg_catalog.citus_add_node(%s, %s, %s, 'primary', 'default'), ('172.27.0.2', 5432, 1))
+    2023-01-05 15:09:59.690 UTC [67] LOG:  standby "coord3" is now a synchronous standby with priority 1
+    2023-01-05 15:09:59.690 UTC [67] STATEMENT:  START_REPLICATION SLOT "coord3" 0/3000000 TIMELINE 1
+    2023-01-05 15:09:59,694 INFO: no action. I am (coord1), the leader with the lock
+    2023-01-05 15:09:59,704 DEBUG: query(SELECT pg_catalog.citus_add_node(%s, %s, %s, 'primary', 'default'), ('172.27.0.8', 5432, 2))
+    2023-01-05 15:10:07,625 INFO: no action. I am (coord1), the leader with the lock
+    2023-01-05 15:10:17,579 INFO: no action. I am (coord1), the leader with the lock
+
+    $ docker exec -ti demo-haproxy bash
+    postgres@haproxy:~$ etcdctl member list
+    1bab629f01fa9065, started, etcd3, http://etcd3:2380, http://172.27.0.10:2379
+    8ecb6af518d241cc, started, etcd2, http://etcd2:2380, http://172.27.0.4:2379
+    b2e169fcb8a34028, started, etcd1, http://etcd1:2380, http://172.27.0.7:2379
+
+    postgres@haproxy:~$ etcdctl get --keys-only --prefix /service/demo
+    /service/demo/0/config
+    /service/demo/0/initialize
+    /service/demo/0/leader
+    /service/demo/0/members/coord1
+    /service/demo/0/members/coord2
+    /service/demo/0/members/coord3
+    /service/demo/0/status
+    /service/demo/0/sync
+    /service/demo/1/config
+    /service/demo/1/initialize
+    /service/demo/1/leader
+    /service/demo/1/members/work1-1
+    /service/demo/1/members/work1-2
+    /service/demo/1/status
+    /service/demo/1/sync
+    /service/demo/2/config
+    /service/demo/2/initialize
+    /service/demo/2/leader
+    /service/demo/2/members/work2-1
+    /service/demo/2/members/work2-2
+    /service/demo/2/status
+    /service/demo/2/sync
+
+    postgres@haproxy:~$ psql -h localhost -p 5000 -U postgres -d citus
+    Password for user postgres: postgres
+    psql (15.1 (Debian 15.1-1.pgdg110+1))
+    SSL connection (protocol: TLSv1.3, cipher: TLS_AES_256_GCM_SHA384, compression: off)
+    Type "help" for help.
+
+    citus=# select pg_is_in_recovery();
+     pg_is_in_recovery
+    -------------------
+     f
+    (1 row)
+
+    citus=# table pg_dist_node;
+     nodeid | groupid |  nodename  | nodeport | noderack | hasmetadata | isactive | noderole | nodecluster | metadatasynced | shouldhaveshards
+    --------+---------+------------+----------+----------+-------------+----------+----------+-------------+----------------+------------------
+          1 |       0 | 172.27.0.6 |     5432 | default  | t           | t        | primary  | default     | t              | f
+          2 |       1 | 172.27.0.2 |     5432 | default  | t           | t        | primary  | default     | t              | t
+          3 |       2 | 172.27.0.8 |     5432 | default  | t           | t        | primary  | default     | t              | t
+    (3 rows)
+
+    citus=# \q
+
+    postgres@haproxy:~$ patronictl list
+    + Citus cluster: demo ----------+--------------+---------+----+-----------+
+    | Group | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    |     0 | coord1  | 172.27.0.6  | Leader       | running |  1 |           |
+    |     0 | coord2  | 172.27.0.5  | Replica      | running |  1 |         0 |
+    |     0 | coord3  | 172.27.0.9  | Sync Standby | running |  1 |         0 |
+    |     1 | work1-1 | 172.27.0.2  | Leader       | running |  1 |           |
+    |     1 | work1-2 | 172.27.0.12 | Sync Standby | running |  1 |         0 |
+    |     2 | work2-1 | 172.27.0.11 | Sync Standby | running |  1 |         0 |
+    |     2 | work2-2 | 172.27.0.8  | Leader       | running |  1 |           |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+
+    postgres@haproxy:~$ patronictl switchover --group 2 --force
+    Current cluster topology
+    + Citus cluster: demo (group: 2, 7185185529556963355) +-----------+
+    | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +---------+-------------+--------------+---------+----+-----------+
+    | work2-1 | 172.27.0.11 | Sync Standby | running |  1 |         0 |
+    | work2-2 | 172.27.0.8  | Leader       | running |  1 |           |
+    +---------+-------------+--------------+---------+----+-----------+
+    2023-01-05 15:29:29.54204 Successfully switched over to "work2-1"
+    + Citus cluster: demo (group: 2, 7185185529556963355) -------+
+    | Member  | Host        | Role    | State   | TL | Lag in MB |
+    +---------+-------------+---------+---------+----+-----------+
+    | work2-1 | 172.27.0.11 | Leader  | running |  1 |           |
+    | work2-2 | 172.27.0.8  | Replica | stopped |    |   unknown |
+    +---------+-------------+---------+---------+----+-----------+
+
+    postgres@haproxy:~$ patronictl list
+    + Citus cluster: demo ----------+--------------+---------+----+-----------+
+    | Group | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    |     0 | coord1  | 172.27.0.6  | Leader       | running |  1 |           |
+    |     0 | coord2  | 172.27.0.5  | Replica      | running |  1 |         0 |
+    |     0 | coord3  | 172.27.0.9  | Sync Standby | running |  1 |         0 |
+    |     1 | work1-1 | 172.27.0.2  | Leader       | running |  1 |           |
+    |     1 | work1-2 | 172.27.0.12 | Sync Standby | running |  1 |         0 |
+    |     2 | work2-1 | 172.27.0.11 | Leader       | running |  2 |           |
+    |     2 | work2-2 | 172.27.0.8  | Sync Standby | running |  2 |         0 |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+
+    postgres@haproxy:~$ psql -h localhost -p 5000 -U postgres -d citus
+    Password for user postgres: postgres
+    psql (15.1 (Debian 15.1-1.pgdg110+1))
+    SSL connection (protocol: TLSv1.3, cipher: TLS_AES_256_GCM_SHA384, compression: off)
+    Type "help" for help.
+
+    citus=# table pg_dist_node;
+     nodeid | groupid |  nodename   | nodeport | noderack | hasmetadata | isactive | noderole | nodecluster | metadatasynced | shouldhaveshards
+    --------+---------+-------------+----------+----------+-------------+----------+----------+-------------+----------------+------------------
+          1 |       0 | 172.27.0.6  |     5432 | default  | t           | t        | primary  | default     | t              | f
+          3 |       2 | 172.27.0.11 |     5432 | default  | t           | t        | primary  | default     | t              | t
+          2 |       1 | 172.27.0.2  |     5432 | default  | t           | t        | primary  | default     | t              | t
+    (3 rows)

docker/dev_patroni_cluster.sh

@@ -1,104 +0,0 @@
-#!/bin/bash
-
-DOCKER_IMAGE="registry.opensource.zalan.do/acid/patroni:1.0-SNAPSHOT"
-MEMBERS=3
-
-
-function usage()
-{
-    cat <<__EOF__
-Usage: $0 
-
-Options:
-
-    --image     IMAGE    The Docker image to use for the cluster
-    --members   INT      The number of members for the cluster
-    --name      NAME     The name of the new cluster
-
-Examples:
-
-    $0 --image ${DOCKER_IMAGE}
-    $0
-    $0 --image ${DOCKER_IMAGE} --members=2
-__EOF__
-}
-
-
-optspec=":-:"
-while getopts "$optspec" optchar; do
-    case "${optchar}" in
-        -)
-            case "${OPTARG}" in
-                help)
-                    usage
-                    exit 0
-                    ;;
-                name)
-                    PATRONI_SCOPE="${!OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
-                    ;;
-                name=*)
-                    PATRONI_SCOPE="${OPTARG#*=}"
-                    ;;
-                image)
-                    DOCKER_IMAGE="${!OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
-                    ;;
-                image=*)
-                    DOCKER_IMAGE="${OPTARG#*=}"
-                    ;;
-                members)
-                    MEMBERS="${!OPTIND}"; OPTIND=$(( $OPTIND + 1 ))
-                    ;;
-                members=*)
-                    MEMBERS="${OPTARG#*=}"
-                    ;;
-                *)
-                    if [ "$OPTERR" = 1 ] && [ "${optspec:0:1}" != ":" ]; then
-                        echo "Unknown option --${OPTARG}" >&2
-                    fi
-                    ;;
-            esac;;
-        *)
-            if [ "$OPTERR" != 1 ] || [ "${optspec:0:1}" = ":" ]; then
-                echo "Non-option argument: '-${OPTARG}'" >&2
-                usage
-                exit 1
-            fi
-            ;;
-    esac
-done
-
-if [ -z ${PATRONI_SCOPE} ]; then
-    PATRONI_SCOPE=$(cat /dev/urandom | LC_ALL=C tr -dc 'a-z0-9' | head -c 8)
-fi
-
-function docker_run()
-{
-    local name=$1
-    shift
-    container=$(docker run -d --name=$name $*)
-    container_ip=$(docker inspect --format '{{ .NetworkSettings.IPAddress }}' ${container})
-    echo "Started container ${name}, ip=${container_ip}"
-}
-
-
-ETCD_CONTAINER="${PATRONI_SCOPE}_etcd"
-docker_run ${ETCD_CONTAINER} ${DOCKER_IMAGE} --etcd
-
-DOCKER_ARGS="--link=${ETCD_CONTAINER}:${ETCD_CONTAINER} -e PATRONI_SCOPE=${PATRONI_SCOPE} -e PATRONI_ETCD_HOST=${ETCD_CONTAINER}:2379"
-PATRONI_ENV=$(sed 's/#.*//g' docker/patroni-secrets.env | sed -n 's/^PATRONI_.*$/-e &/p' | tr '\n' ' ')
-PATRONI_VOLUME="-v $(dirname $(dirname $(realpath $0)))/patroni:/patroni"
-
-for i in $(seq 1 "${MEMBERS}"); do
-    container_name=postgres${i}
-    docker_run "${PATRONI_SCOPE}_${container_name}" \
-        $PATRONI_VOLUME \
-        $DOCKER_ARGS \
-        $PATRONI_ENV \
-        -e PATRONI_NAME=${container_name} \
-        ${DOCKER_IMAGE}
-done
-
-docker_run "${PATRONI_SCOPE}_haproxy" \
-    -p=5000 -p=5001 \
-    $DOCKER_ARGS \
-    ${DOCKER_IMAGE} --confd

docker/entrypoint.sh

@@ -1,115 +1,75 @@
-#!/bin/bash
+#!/bin/sh
 
-function usage()
-{
-    cat <<__EOF__
-Usage: $0
-
-Options:
-
-    --etcd      Do not run Patroni, run a standalone etcd
-    --confd     Do not run Patroni, run a standalone confd
-    --zookeeper Do not run Patroni, run a standalone zookeeper
-
-Examples:
-
-    $0 --etcd
-    $0 --confd
-    $0 --zookeeper
-    $0
-__EOF__
-}
+if [ -f /a.tar.xz ]; then
+    echo "decompressing image..."
+    sudo tar xpJf /a.tar.xz -C / > /dev/null 2>&1
+    sudo rm /a.tar.xz
+    sudo ln -snf dash /bin/sh
+fi
 
+readonly PATRONI_SCOPE="${PATRONI_SCOPE:-batman}"
+PATRONI_NAMESPACE="${PATRONI_NAMESPACE:-/service}"
+readonly PATRONI_NAMESPACE="${PATRONI_NAMESPACE%/}"
 DOCKER_IP=$(hostname --ip-address)
-PATRONI_SCOPE=${PATRONI_SCOPE:-batman}
-ETCD_ARGS="--data-dir /tmp/etcd.data -advertise-client-urls=http://${DOCKER_IP}:2379 -listen-client-urls=http://0.0.0.0:2379 -listen-peer-urls=http://0.0.0.0:2380"
+readonly DOCKER_IP
 
-optspec=":vh-:"
-while getopts "$optspec" optchar; do
-    case "${optchar}" in
-        -)
-            case "${OPTARG}" in
-                confd)
-                    haproxy -f /etc/haproxy/haproxy.cfg -p /var/run/haproxy.pid -D
-                    CONFD="confd -prefix=${PATRONI_NAMESPACE:-/service}/$PATRONI_SCOPE -interval=10 -backend"
-                    if [ ! -z ${PATRONI_ZOOKEEPER_HOSTS} ]; then
-                        while ! /usr/share/zookeeper/bin/zkCli.sh -server ${PATRONI_ZOOKEEPER_HOSTS} ls /; do
-                            sleep 1
-                        done
-                        exec $CONFD zookeeper -node ${PATRONI_ZOOKEEPER_HOSTS}
-                    else
-                        while ! curl -s ${PATRONI_ETCD_HOST}/v2/members | jq -r '.members[0].clientURLs[0]' | grep -q http; do
-                            sleep 1
-                        done
-                        exec $CONFD etcd -node $PATRONI_ETCD_HOST
-                    fi
-                    ;;
-                etcd)
-                    exec etcd $ETCD_ARGS
-                    ;;
-                zookeeper)
-                    exec /usr/share/zookeeper/bin/zkServer.sh start-foreground
-                    ;;
-                cheat)
-                    CHEAT=1
-                    ;;
-                help)
-                    usage
-                    exit 0
-                    ;;
-                *)
-                    if [ "$OPTERR" = 1 ] && [ "${optspec:0:1}" != ":" ]; then
-                        echo "Unknown option --${OPTARG}" >&2
-                    fi
-                    ;;
-            esac;;
-        *)
-            if [ "$OPTERR" != 1 ] || [ "${optspec:0:1}" = ":" ]; then
-                echo "Non-option argument: '-${OPTARG}'" >&2
-                usage
-                exit 1
-            fi
-            ;;
-    esac
-done
+case "$1" in
+    haproxy)
+        haproxy -f /etc/haproxy/haproxy.cfg -p /var/run/haproxy.pid -D
+        set -- confd "-prefix=$PATRONI_NAMESPACE/$PATRONI_SCOPE" -interval=10 -backend
+        if [ -n "$PATRONI_ZOOKEEPER_HOSTS" ]; then
+            while ! /usr/share/zookeeper/bin/zkCli.sh -server "$PATRONI_ZOOKEEPER_HOSTS" ls /; do
+                sleep 1
+            done
+            set -- "$@" zookeeper -node "$PATRONI_ZOOKEEPER_HOSTS"
+        else
+            while ! etcdctl member list 2> /dev/null; do
+                sleep 1
+            done
+            set -- "$@" etcdv3
+            while IFS='' read -r line; do
+                set -- "$@" -node "$line"
+            done <<-EOT
+$(echo "$ETCDCTL_ENDPOINTS" | sed 's/,/\n/g')
+EOT
+        fi
+        exec dumb-init "$@"
+        ;;
+    etcd)
+        exec "$@" -advertise-client-urls "http://$DOCKER_IP:2379"
+        ;;
+    zookeeper)
+        exec /usr/share/zookeeper/bin/zkServer.sh start-foreground
+        ;;
+esac
 
 ## We start an etcd
-if [[ -z ${PATRONI_ETCD_HOST} && -z ${PATRONI_ZOOKEEPER_HOSTS} ]]; then
-    etcd $ETCD_ARGS > /var/log/etcd.log 2> /var/log/etcd.err &
-    export PATRONI_ETCD_HOST="127.0.0.1:2379"
+if [ -z "$PATRONI_ETCD3_HOSTS" ] && [ -z "$PATRONI_ZOOKEEPER_HOSTS" ]; then
+    export PATRONI_ETCD_URL="http://127.0.0.1:2379"
+    etcd --data-dir /tmp/etcd.data -advertise-client-urls=$PATRONI_ETCD_URL -listen-client-urls=http://0.0.0.0:2379 > /var/log/etcd.log 2> /var/log/etcd.err &
 fi
 
 export PATRONI_SCOPE
-export PATRONI_NAME="${PATRONI_NAME:-${HOSTNAME}}"
-export PATRONI_RESTAPI_CONNECT_ADDRESS="${DOCKER_IP}:8008"
+export PATRONI_NAMESPACE
+export PATRONI_NAME="${PATRONI_NAME:-$(hostname)}"
+export PATRONI_RESTAPI_CONNECT_ADDRESS="$DOCKER_IP:8008"
 export PATRONI_RESTAPI_LISTEN="0.0.0.0:8008"
-export PATRONI_admin_PASSWORD="${PATRONI_admin_PASSWORD:=admin}"
+export PATRONI_admin_PASSWORD="${PATRONI_admin_PASSWORD:-admin}"
 export PATRONI_admin_OPTIONS="${PATRONI_admin_OPTIONS:-createdb, createrole}"
-export PATRONI_POSTGRESQL_CONNECT_ADDRESS="${DOCKER_IP}:5432"
+export PATRONI_POSTGRESQL_CONNECT_ADDRESS="$DOCKER_IP:5432"
 export PATRONI_POSTGRESQL_LISTEN="0.0.0.0:5432"
-export PATRONI_POSTGRESQL_DATA_DIR="data/${PATRONI_SCOPE}"
+export PATRONI_POSTGRESQL_DATA_DIR="${PATRONI_POSTGRESQL_DATA_DIR:-$PGDATA}"
 export PATRONI_REPLICATION_USERNAME="${PATRONI_REPLICATION_USERNAME:-replicator}"
-export PATRONI_REPLICATION_PASSWORD="${PATRONI_REPLICATION_PASSWORD:-abcd}"
+export PATRONI_REPLICATION_PASSWORD="${PATRONI_REPLICATION_PASSWORD:-replicate}"
 export PATRONI_SUPERUSER_USERNAME="${PATRONI_SUPERUSER_USERNAME:-postgres}"
 export PATRONI_SUPERUSER_PASSWORD="${PATRONI_SUPERUSER_PASSWORD:-postgres}"
-export PATRONI_POSTGRESQL_PGPASS="$HOME/.pgpass"
+export PATRONI_REPLICATION_SSLMODE="${PATRONI_REPLICATION_SSLMODE:-$PGSSLMODE}"
+export PATRONI_REPLICATION_SSLKEY="${PATRONI_REPLICATION_SSLKEY:-$PGSSLKEY}"
+export PATRONI_REPLICATION_SSLCERT="${PATRONI_REPLICATION_SSLCERT:-$PGSSLCERT}"
+export PATRONI_REPLICATION_SSLROOTCERT="${PATRONI_REPLICATION_SSLROOTCERT:-$PGSSLROOTCERT}"
+export PATRONI_SUPERUSER_SSLMODE="${PATRONI_SUPERUSER_SSLMODE:-$PGSSLMODE}"
+export PATRONI_SUPERUSER_SSLKEY="${PATRONI_SUPERUSER_SSLKEY:-$PGSSLKEY}"
+export PATRONI_SUPERUSER_SSLCERT="${PATRONI_SUPERUSER_SSLCERT:-$PGSSLCERT}"
+export PATRONI_SUPERUSER_SSLROOTCERT="${PATRONI_SUPERUSER_SSLROOTCERT:-$PGSSLROOTCERT}"
 
-cat > /patroni.yml <<__EOF__
-bootstrap:
-  dcs:
-    postgresql:
-      use_pg_rewind: true
-
-  pg_hba:
-  - host all all 0.0.0.0/0 md5
-  - host replication replicator ${DOCKER_IP}/16    md5
-__EOF__
-
-mkdir -p "$HOME/.config/patroni"
-[ -h "$HOME/.config/patroni/patronictl.yaml" ] || ln -s /patroni.yml "$HOME/.config/patroni/patronictl.yaml"
-
-[ -z $CHEAT ] && exec python /patroni.py /patroni.yml
-
-while true; do
-    sleep 60
-done
+exec python3 /patroni.py postgres0.yml

docs/CONTRIBUTING.rst

@@ -5,6 +5,44 @@ Contributing guidelines
 
 Wanna contribute to Patroni? Yay - here is how!
 
+Chatting
+--------
+
+Just want to chat with other Patroni users?  Looking for interactive troubleshooting help? Join us on channel `#patroni <https://postgresteam.slack.com/archives/C9XPYG92A>`__ in the `PostgreSQL Slack <https://postgresteam.slack.com/>`__.
+
+Running tests
+-------------
+
+Requirements for running behave tests:
+
+1. PostgreSQL packages need to be installed.
+2. PostgreSQL binaries must be available in your `PATH`. You may need to add them to the path with something like `PATH=/usr/lib/postgresql/11/bin:$PATH python -m behave`.
+3. If you'd like to test with external DCSs (e.g., Etcd, Consul, and Zookeeper) you'll need the packages installed and respective services running and accepting unencrypted/unprotected connections on localhost and default port. In the case of Etcd or Consul, the behave test suite could start them up if binaries are available in the `PATH`.
+
+ Install dependencies:
+
+.. code-block:: bash
+
+    # You may want to use Virtualenv or specify pip3.
+    pip install -r requirements.txt
+    pip install -r requirements.dev.txt
+
+After you have all dependencies installed, you can run the various test suites:
+
+.. code-block:: bash
+
+    # You may want to use Virtualenv or specify python3.
+
+    # Run flake8 to check syntax and formatting:
+    python setup.py flake8
+
+    # Run the pytest suite in tests/:
+    python setup.py test
+
+    # Run the behave (https://behave.readthedocs.io/en/latest/) test suite in features/;
+    # modify DCS as desired (raft has no dependencies so is the easiest to start with):
+    DCS=raft python -m behave
+
 Reporting issues
 ----------------
 

docs/ENVIRONMENT.rst

@@ -1,6 +1,5 @@
 .. _environment:
 
-==================================
 Environment Configuration Settings
 ==================================
 
@@ -13,6 +12,18 @@ Global/Universal
 -  **PATRONI\_NAMESPACE**: path within the configuration store where Patroni will keep information about the cluster. Default value: "/service"
 -  **PATRONI\_SCOPE**: cluster name
 
+Log
+---
+-  **PATRONI\_LOG\_LEVEL**: sets the general logging level. Default value is **INFO** (see `the docs for Python logging <https://docs.python.org/3.6/library/logging.html#levels>`_)
+-  **PATRONI\_LOG\_TRACEBACK\_LEVEL**: sets the level where tracebacks will be visible. Default value is **ERROR**. Set it to **DEBUG** if you want to see tracebacks only if you enable **PATRONI\_LOG\_LEVEL=DEBUG**.
+-  **PATRONI\_LOG\_FORMAT**: sets the log formatting string. Default value is **%(asctime)s %(levelname)s: %(message)s** (see `the LogRecord attributes <https://docs.python.org/3.6/library/logging.html#logrecord-attributes>`_)
+-  **PATRONI\_LOG\_DATEFORMAT**: sets the datetime formatting string. (see the `formatTime() documentation <https://docs.python.org/3.6/library/logging.html#logging.Formatter.formatTime>`_)
+-  **PATRONI\_LOG\_MAX\_QUEUE\_SIZE**: Patroni is using two-step logging. Log records are written into the in-memory queue and there is a separate thread which pulls them from the queue and writes to stderr or file. The maximum size of the internal queue is limited by default by **1000** records, which is enough to keep logs for the past 1h20m.
+-  **PATRONI\_LOG\_DIR**: Directory to write application logs to. The directory must exist and be writable by the user executing Patroni. If you set this env variable, the application will retain 4 25MB logs by default. You can tune those retention values with `PATRONI_LOG_FILE_NUM` and `PATRONI_LOG_FILE_SIZE` (see below).
+-  **PATRONI\_LOG\_FILE\_NUM**: The number of application logs to retain.
+-  **PATRONI\_LOG\_FILE\_SIZE**: Size of patroni.log file (in bytes) that triggers a log rolling.
+-  **PATRONI\_LOG\_LOGGERS**: Redefine logging level per python module. Example ``PATRONI_LOG_LOGGERS="{patroni.postmaster: WARNING, urllib3: DEBUG}"``
+
 Bootstrap configuration
 -----------------------
 It is possible to create new database users right after the successful initialization of a new cluster. This process is defined by the following variables:
@@ -22,46 +33,164 @@ It is possible to create new database users right after the successful initializ
 
 Example: defining ``PATRONI_admin_PASSWORD=strongpasswd`` and ``PATRONI_admin_OPTIONS='createrole,createdb'`` will cause creation of the user **admin** with the password **strongpasswd** that is allowed to create other users and databases.
 
+Citus
+-----
+Enables integration Patroni with `Citus <https://docs.citusdata.com>`__. If configured, Patroni will take care of registering Citus worker nodes on the coordinator. You can find more information about Citus support :ref:`here <citus>`.
+
+-  **PATRONI\_CITUS\_GROUP**: the Citus group id, integer. Use ``0`` for coordinator and ``1``, ``2``, etc... for workers
+-  **PATRONI\_CITUS\_DATABASE**: the database where ``citus`` extension should be created. Must be the same on the coordinator and all workers. Currently only one database is supported.
+
 Consul
 ------
--  **PATRONI\_CONSUL\_HOST**: the host:port for the Consul endpoint.
+-  **PATRONI\_CONSUL\_HOST**: the host:port for the Consul local agent.
+-  **PATRONI\_CONSUL\_URL**: url for the Consul local agent, in format: http(s)://host:port
+-  **PATRONI\_CONSUL\_PORT**: (optional) Consul port
+-  **PATRONI\_CONSUL\_SCHEME**: (optional) **http** or **https**, defaults to **http**
+-  **PATRONI\_CONSUL\_TOKEN**: (optional) ACL token
+-  **PATRONI\_CONSUL\_VERIFY**: (optional) whether to verify the SSL certificate for HTTPS requests
+-  **PATRONI\_CONSUL\_CACERT**: (optional) The ca certificate. If present it will enable validation.
+-  **PATRONI\_CONSUL\_CERT**: (optional) File with the client certificate
+-  **PATRONI\_CONSUL\_KEY**: (optional) File with the client key. Can be empty if the key is part of certificate.
+-  **PATRONI\_CONSUL\_DC**: (optional) Datacenter to communicate with. By default the datacenter of the host is used.
+-  **PATRONI\_CONSUL\_CONSISTENCY**: (optional) Select consul consistency mode. Possible values are ``default``, ``consistent``, or ``stale`` (more details in `consul API reference <https://www.consul.io/api/features/consistency.html/>`__)
+-  **PATRONI\_CONSUL\_CHECKS**: (optional) list of Consul health checks used for the session. By default an empty list is used.
+-  **PATRONI\_CONSUL\_REGISTER\_SERVICE**: (optional) whether or not to register a service with the name defined by the scope parameter and the tag master, primary, replica, or standby-leader depending on the node's role. Defaults to **false**
+-  **PATRONI\_CONSUL\_SERVICE\_TAGS**: (optional) additional static tags to add to the Consul service apart from the role (``master``/``primary``/``replica``/``standby-leader``). By default an empty list is used.
+-  **PATRONI\_CONSUL\_SERVICE\_CHECK\_INTERVAL**: (optional) how often to perform health check against registered url
+-  **PATRONI\_CONSUL\_SERVICE\_CHECK\_TLS\_SERVER\_NAME**: (optional) overide SNI host when connecting via TLS, see also `consul agent check API reference <https://www.consul.io/api-docs/agent/check#tlsservername>`__.
 
 Etcd
 ----
--  **PATRONI\_ETCD\_HOST**: the host:port for the etcd endpoint.
--  **PATRONI\_ETCD\_URL**: url for the etcd, in format: http(s)://(username:password@)host:port
+
 -  **PATRONI\_ETCD\_PROXY**: proxy url for the etcd. If you are connecting to the etcd using proxy, use this parameter instead of **PATRONI\_ETCD\_URL**
--  **PATRONI\_ETCD\_SRV**: Domain to search the SRV record(s) for cluster autodiscovery.
--  **PATRONI\_ETCD\_CACERT**: The ca certificate. If pressent it will enable validation.
--  **PATRONI\_ETCD\_CERT**: File with the client certificate
+-  **PATRONI\_ETCD\_URL**: url for the etcd, in format: http(s)://(username:password@)host:port
+-  **PATRONI\_ETCD\_HOSTS**: list of etcd endpoints in format 'host1:port1','host2:port2',etc...
+-  **PATRONI\_ETCD\_USE\_PROXIES**: If this parameter is set to true, Patroni will consider **hosts** as a list of proxies and will not perform a topology discovery of etcd cluster but stick to a fixed list of **hosts**.
+-  **PATRONI\_ETCD\_PROTOCOL**: http or https, if not specified http is used. If the **url** or **proxy** is specified - will take protocol from them.
+-  **PATRONI\_ETCD\_HOST**: the host:port for the etcd endpoint.
+-  **PATRONI\_ETCD\_SRV**: Domain to search the SRV record(s) for cluster autodiscovery. Patroni will try to query these SRV service names for specified domain (in that order until first success): ``_etcd-client-ssl``, ``_etcd-client``, ``_etcd-ssl``, ``_etcd``, ``_etcd-server-ssl``, ``_etcd-server``. If SRV records for ``_etcd-server-ssl`` or ``_etcd-server`` are retrieved then ETCD peer protocol is used do query ETCD for available members. Otherwise hosts from SRV records will be used.
+-  **PATRONI\_ETCD\_SRV\_SUFFIX**: Configures a suffix to the SRV name that is queried during discovery. Use this flag to differentiate between multiple etcd clusters under the same domain. Works only with conjunction with **PATRONI\_ETCD\_SRV**. For example, if ``PATRONI_ETCD_SRV_SUFFIX=foo`` and ``PATRONI_ETCD_SRV=example.org`` are set, the following DNS SRV query is made:``_etcd-client-ssl-foo._tcp.example.com`` (and so on for every possible ETCD SRV service name).
+-  **PATRONI\_ETCD\_USERNAME**: username for etcd authentication.
+-  **PATRONI\_ETCD\_PASSWORD**: password for etcd authentication.
+-  **PATRONI\_ETCD\_CACERT**: The ca certificate. If present it will enable validation.
+-  **PATRONI\_ETCD\_CERT**: File with the client certificate.
 -  **PATRONI\_ETCD\_KEY**: File with the client key. Can be empty if the key is part of certificate.
 
+Etcdv3
+------
+Environment names for Etcdv3 are similar as for Etcd, you just need to use ``ETCD3`` instead of ``ETCD`` in the variable name. Example: ``PATRONI_ETCD3_HOST``, ``PATRONI_ETCD3_CACERT``, and so on.
+
+.. warning::
+    Keys created with protocol version 2 are not visible with protocol version 3 and the other way around, therefore it is not possible to switch from Etcd to Etcdv3 just by updating Patroni configuration.
+
+
+ZooKeeper
+---------
+-  **PATRONI\_ZOOKEEPER\_HOSTS**: Comma separated list of ZooKeeper cluster members: "'host1:port1','host2:port2','etc...'". It is important to quote every single entity!
+-  **PATRONI\_ZOOKEEPER\_USE\_SSL**: (optional) Whether SSL is used or not. Defaults to ``false``. If set to ``false``, all SSL specific parameters are ignored.
+-  **PATRONI\_ZOOKEEPER\_CACERT**: (optional) The CA certificate. If present it will enable validation.
+-  **PATRONI\_ZOOKEEPER\_CERT**: (optional) File with the client certificate.
+-  **PATRONI\_ZOOKEEPER\_KEY**: (optional) File with the client key.
+-  **PATRONI\_ZOOKEEPER\_KEY\_PASSWORD**: (optional) The client key password.
+-  **PATRONI\_ZOOKEEPER\_VERIFY**: (optional) Whether to verify certificate or not. Defaults to ``true``.
+-  **PATRONI\_ZOOKEEPER\_SET\_ACLS**: (optional) If set, configure Kazoo to apply a default ACL to each ZNode that it creates. ACLs will assume 'x509' schema and should be specified as a dictionary with the principal as the key and one or more permissions as a list in the value.  Permissions may be one of ``CREATE``, ``READ``, ``WRITE``, ``DELETE`` or ``ADMIN``.  For example, ``set_acls: {CN=principal1: [CREATE, READ], CN=principal2: [ALL]}``.
+
+.. note::
+    It is required to install ``kazoo>=2.6.0`` to support SSL.
+
+
 Exhibitor
 ---------
 -  **PATRONI\_EXHIBITOR\_HOSTS**: initial list of Exhibitor (ZooKeeper) nodes in format: 'host1,host2,etc...'. This list updates automatically whenever the Exhibitor (ZooKeeper) cluster topology changes.
 -  **PATRONI\_EXHIBITOR\_PORT**: Exhibitor port.
 
+.. _kubernetes_environment:
+
+Kubernetes
+----------
+-  **PATRONI\_KUBERNETES\_BYPASS\_API\_SERVICE**: (optional) When communicating with the Kubernetes API, Patroni is usually relying on the `kubernetes` service, the address of which is exposed in the pods via the `KUBERNETES_SERVICE_HOST` environment variable. If `PATRONI_KUBERNETES_BYPASS_API_SERVICE` is set to ``true``, Patroni will resolve the list of API nodes behind the service and connect directly to them.
+-  **PATRONI\_KUBERNETES\_NAMESPACE**: (optional) Kubernetes namespace where the Patroni pod is running. Default value is `default`.
+-  **PATRONI\_KUBERNETES\_LABELS**: Labels in format ``{label1: value1, label2: value2}``. These labels will be used to find existing objects (Pods and either Endpoints or ConfigMaps) associated with the current cluster. Also Patroni will set them on every object (Endpoint or ConfigMap) it creates.
+-  **PATRONI\_KUBERNETES\_SCOPE\_LABEL**: (optional) name of the label containing cluster name. Default value is `cluster-name`.
+-  **PATRONI\_KUBERNETES\_ROLE\_LABEL**: (optional) name of the label containing Postgres role (`master` or `replica`). Patroni will set this label on the pod it is running in. Default value is `role`.
+-  **PATRONI\_KUBERNETES\_USE\_ENDPOINTS**: (optional) if set to true, Patroni will use Endpoints instead of ConfigMaps to run leader elections and keep cluster state.
+-  **PATRONI\_KUBERNETES\_POD\_IP**: (optional) IP address of the pod Patroni is running in. This value is required when `PATRONI_KUBERNETES_USE_ENDPOINTS` is enabled and is used to populate the leader endpoint subsets when the pod's PostgreSQL is promoted.
+-  **PATRONI\_KUBERNETES\_PORTS**: (optional) if the Service object has the name for the port, the same name must appear in the Endpoint object, otherwise service won't work. For example, if your service is defined as ``{Kind: Service, spec: {ports: [{name: postgresql, port: 5432, targetPort: 5432}]}}``, then you have to set ``PATRONI_KUBERNETES_PORTS='[{"name": "postgresql", "port": 5432}]'`` and Patroni will use it for updating subsets of the leader Endpoint. This parameter is used only if `PATRONI_KUBERNETES_USE_ENDPOINTS` is set.
+-  **PATRONI\_KUBERNETES\_CACERT**: (optional) Specifies the file with the CA_BUNDLE file with certificates of trusted CAs to use while verifying Kubernetes API SSL certs. If not provided, patroni will use the value provided by the ServiceAccount secret.
+
+Raft (deprecated)
+-----------------
+
+-  **PATRONI\_RAFT\_SELF\_ADDR**: ``ip:port`` to listen on for Raft connections. The ``self_addr`` must be accessible from other nodes of the cluster. If not set, the node will not participate in consensus.
+-  **PATRONI\_RAFT\_BIND\_ADDR**: (optional) ``ip:port`` to listen on for Raft connections. If not specified the ``self_addr`` will be used.
+-  **PATRONI\_RAFT\_PARTNER\_ADDRS**: list of other Patroni nodes in the cluster in format ``"'ip1:port1','ip2:port2'"``. It is important to quote every single entity!
+-  **PATRONI\_RAFT\_DATA\_DIR**: directory where to store Raft log and snapshot. If not specified the current working directory is used.
+-  **PATRONI\_RAFT\_PASSWORD**: (optional) Encrypt Raft traffic with a specified password, requires ``cryptography`` python module.
+
 PostgreSQL
 ----------
 -  **PATRONI\_POSTGRESQL\_LISTEN**: IP address + port that Postgres listens to. Multiple comma-separated addresses are permitted, as long as the port component is appended after to the last one with a colon, i.e. ``listen: 127.0.0.1,127.0.0.2:5432``. Patroni will use the first address from this list to establish local connections to the PostgreSQL node.
 -  **PATRONI\_POSTGRESQL\_CONNECT\_ADDRESS**: IP address + port through which Postgres is accessible from other nodes and applications.
+-  **PATRONI\_POSTGRESQL\_PROXY\_ADDRESS**: IP address + port through which a connection pool (e.g. pgbouncer) running next to Postgres is accessible. The value is written to the member key in DCS as ``proxy_url`` and could be used/useful for service discovery.
 -  **PATRONI\_POSTGRESQL\_DATA\_DIR**: The location of the Postgres data directory, either existing or to be initialized by Patroni.
+-  **PATRONI\_POSTGRESQL\_CONFIG\_DIR**: The location of the Postgres configuration directory, defaults to the data directory. Must be writable by Patroni.
 -  **PATRONI\_POSTGRESQL\_BIN_DIR**: Path to PostgreSQL binaries. (pg_ctl, pg_rewind, pg_basebackup, postgres) The  default value is an empty string meaning that PATH environment variable will be used to find the executables.
 -  **PATRONI\_POSTGRESQL\_PGPASS**: path to the `.pgpass <https://www.postgresql.org/docs/current/static/libpq-pgpass.html>`__ password file. Patroni creates this file before executing pg\_basebackup and under some other circumstances. The location must be writable by Patroni.
--  **PATRONI\_REPLICATION\_USERNAME**: replication username; the user will be created during initialization. Replicas will use this user to access master via streaming replication
+-  **PATRONI\_REPLICATION\_USERNAME**: replication username; the user will be created during initialization. Replicas will use this user to access the replication source via streaming replication
 -  **PATRONI\_REPLICATION\_PASSWORD**: replication password; the user will be created during initialization.
+-  **PATRONI\_REPLICATION\_SSLMODE**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+-  **PATRONI\_REPLICATION\_SSLKEY**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+-  **PATRONI\_REPLICATION\_SSLPASSWORD**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``PATRONI_REPLICATION_SSLKEY``.
+-  **PATRONI\_REPLICATION\_SSLCERT**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+-  **PATRONI\_REPLICATION\_SSLROOTCERT**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+-  **PATRONI\_REPLICATION\_SSLCRL**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_REPLICATION\_SSLCRLDIR**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_REPLICATION\_GSSENCMODE**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+-  **PATRONI\_REPLICATION\_CHANNEL\_BINDING**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
 -  **PATRONI\_SUPERUSER\_USERNAME**: name for the superuser, set during initialization (initdb) and later used by Patroni to connect to the postgres. Also this user is used by pg_rewind.
 -  **PATRONI\_SUPERUSER\_PASSWORD**: password for the superuser, set during initialization (initdb).
+-  **PATRONI\_SUPERUSER\_SSLMODE**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+-  **PATRONI\_SUPERUSER\_SSLKEY**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+-  **PATRONI\_SUPERUSER\_SSLPASSWORD**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``PATRONI_SUPERUSER_SSLKEY``.
+-  **PATRONI\_SUPERUSER\_SSLCERT**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+-  **PATRONI\_SUPERUSER\_SSLROOTCERT**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+-  **PATRONI\_SUPERUSER\_SSLCRL**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_SUPERUSER\_SSLCRLDIR**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_SUPERUSER\_GSSENCMODE**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+-  **PATRONI\_SUPERUSER\_CHANNEL\_BINDING**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
+-  **PATRONI\_REWIND\_USERNAME**: name for the user for ``pg_rewind``; the user will be created during initialization of postgres 11+ and all necessary `permissions <https://www.postgresql.org/docs/11/app-pgrewind.html#id-1.9.5.8.8>`__ will be granted.
+-  **PATRONI\_REWIND\_PASSWORD**: password for the user for ``pg_rewind``; the user will be created during initialization.
+-  **PATRONI\_REWIND\_SSLMODE**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+-  **PATRONI\_REWIND\_SSLKEY**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+-  **PATRONI\_REWIND\_SSLPASSWORD**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``PATRONI_REWIND_SSLKEY``.
+-  **PATRONI\_REWIND\_SSLCERT**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+-  **PATRONI\_REWIND\_SSLROOTCERT**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+-  **PATRONI\_REWIND\_SSLCRL**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_REWIND\_SSLCRLDIR**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+-  **PATRONI\_REWIND\_GSSENCMODE**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+-  **PATRONI\_REWIND\_CHANNEL\_BINDING**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
 
 REST API
--------- 
+--------
 -  **PATRONI\_RESTAPI\_CONNECT\_ADDRESS**: IP address and port to access the REST API.
 -  **PATRONI\_RESTAPI\_LISTEN**: IP address and port that Patroni will listen to, to provide health-check information for HAProxy.
 -  **PATRONI\_RESTAPI\_USERNAME**: Basic-auth username to protect unsafe REST API endpoints.
 -  **PATRONI\_RESTAPI\_PASSWORD**: Basic-auth password to protect unsafe REST API endpoints.
 -  **PATRONI\_RESTAPI\_CERTFILE**: Specifies the file with the certificate in the PEM format. If the certfile is not specified or is left empty, the API server will work without SSL.
 -  **PATRONI\_RESTAPI\_KEYFILE**: Specifies the file with the secret key in the PEM format.
+-  **PATRONI\_RESTAPI\_KEYFILE\_PASSWORD**: Specifies a password for decrypting the keyfile.
+-  **PATRONI\_RESTAPI\_CAFILE**: Specifies the file with the CA_BUNDLE with certificates of trusted CAs to use while verifying client certs.
+-  **PATRONI\_RESTAPI\_CIPHERS**: (optional) Specifies the permitted cipher suites (e.g. "ECDHE-RSA-AES256-GCM-SHA384:DHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-GCM-SHA256:DHE-RSA-AES128-GCM-SHA256:!SSLv1:!SSLv2:!SSLv3:!TLSv1:!TLSv1.1")
+-  **PATRONI\_RESTAPI\_VERIFY\_CLIENT**: ``none`` (default), ``optional`` or ``required``. When ``none`` REST API will not check client certificates. When ``required`` client certificates are required for all REST API calls. When ``optional`` client certificates are required for all unsafe REST API endpoints. When ``required`` is used, then client authentication succeeds, if the certificate signature verification succeeds. For ``optional`` the client cert will only be checked for ``PUT``, ``POST``, ``PATCH``, and ``DELETE`` requests.
+-  **PATRONI\_RESTAPI\_ALLOWLIST**: (optional): Specifies the set of hosts that are allowed to call unsafe REST API endpoints. The single element could be a host name, an IP address or a network address using CIDR notation. By default ``allow all`` is used. In case if ``allowlist`` or ``allowlist_include_members`` are set, anything that is not included is rejected.
+-  **PATRONI\_RESTAPI\_ALLOWLIST\_INCLUDE\_MEMBERS**: (optional): If set to ``true`` it allows accessing unsafe REST API endpoints from other cluster members registered in DCS (IP address or hostname is taken from the members ``api_url``). Be careful, it might happen that OS will use a different IP for outgoing connections.
+-  **PATRONI\_RESTAPI\_HTTP\_EXTRA\_HEADERS**: (optional) HTTP headers let the REST API server pass additional information with an HTTP response.
+-  **PATRONI\_RESTAPI\_HTTPS\_EXTRA\_HEADERS**: (optional) HTTPS headers let the REST API server pass additional information with an HTTP response when TLS is enabled. This will also pass additional information set in ``http_extra_headers``.
 
-ZooKeeper
----------
--  **PATRONI\_ZOOKEEPER\_HOSTS**: comma separated list of ZooKeeper cluster members: "'host1:port1','host2:port2','etc...'". It is important to quote every single entity!
+CTL
+---
+-  **PATRONICTL\_CONFIG\_FILE**: location of the configuration file.
+-  **PATRONI\_CTL\_INSECURE**: Allow connections to REST API without verifying SSL certs.
+-  **PATRONI\_CTL\_CACERT**: Specifies the file with the CA_BUNDLE file or directory with certificates of trusted CAs to use while verifying REST API SSL certs. If not provided patronictl will use the value provided for REST API "cafile" parameter.
+-  **PATRONI\_CTL\_CERTFILE**: Specifies the file with the client certificate in the PEM format. If not provided patronictl will use the value provided for REST API "certfile" parameter.
+-  **PATRONI\_CTL\_KEYFILE**: Specifies the file with the client secret key in the PEM format. If not provided patronictl will use the value provided for REST API "keyfile" parameter.

docs/README.rst

@@ -1,8 +1,8 @@
 .. _readme:
 
-=================
-How Patroni Works
-=================
+============
+Introduction
+============
 
 Patroni originated as a fork of `Governor <https://github.com/compose/governor>`__, the project from Compose. It includes plenty of new features.
 
@@ -13,35 +13,116 @@ For additional background info, see:
 * `PostgreSQL HA with Kubernetes and Patroni <https://www.youtube.com/watch?v=iruaCgeG7qs>`__, talk by Josh Berkus at KubeCon 2016 (video)
 * `Feb. 2016 Zalando Tech blog post <https://tech.zalando.de/blog/zalandos-patroni-a-template-for-high-availability-postgresql/>`__
 
-==================
+
 Development Status
-==================
+------------------
 
 Patroni is in active development and accepts contributions. See our :ref:`Contributing <contributing>` section below for more details.
 
 We report new releases information :ref:`here <releases>`.
 
-===================================
-Technical Requirements/Installation
-===================================
 
-**For Mac**
+Technical Requirements/Installation
+-----------------------------------
+
+**Pre-requirements for Mac OS**
 
 To install requirements on a Mac, run the following:
 
 ::
 
     brew install postgresql etcd haproxy libyaml python
-    pip install psycopg2 pyyaml
 
-=======================
+.. _psycopg2_install_options:
+
+**Psycopg**
+
+Starting from `psycopg2-2.8 <http://initd.org/psycopg/articles/2019/04/04/psycopg-28-released/>`__ the binary version of psycopg2 will no longer be installed by default. Installing it from the source code requires C compiler and postgres+python dev packages.
+Since in the python world it is not possible to specify dependency as ``psycopg2 OR psycopg2-binary`` you will have to decide how to install it.
+
+There are a few options available:
+
+1. Use the package manager from your distro
+
+::
+
+    sudo apt-get install python-psycopg2   # install python2 psycopg2 module on Debian/Ubuntu
+    sudo apt-get install python3-psycopg2  # install python3 psycopg2 module on Debian/Ubuntu
+    sudo yum install python-psycopg2       # install python2 psycopg2 on RedHat/Fedora/CentOS
+
+2. Install psycopg2 from the binary package
+
+::
+
+    pip install psycopg2-binary
+
+3. Install psycopg2 from source
+
+::
+
+    pip install psycopg2>=2.5.4
+
+4. Use psycopg 3.0 instead of psycopg2
+
+::
+
+    pip install psycopg[binary]>=3.0.0
+
+**General installation for pip**
+
+Patroni can be installed with pip:
+
+::
+
+    pip install patroni[dependencies]
+
+where dependencies can be either empty, or consist of one or more of the following:
+
+etcd or etcd3
+    `python-etcd` module in order to use Etcd as Distributed Configuration Store (DCS)
+consul
+    `python-consul` module in order to use Consul as DCS
+zookeeper
+    `kazoo` module in order to use Zookeeper as DCS
+exhibitor
+    `kazoo` module in order to use Exhibitor as DCS (same dependencies as for Zookeeper)
+kubernetes
+    `kubernetes` module in order to use Kubernetes as DCS in Patroni
+raft
+    `pysyncobj` module in order to use python Raft implementation as DCS
+aws
+    `boto` in order to use AWS callbacks
+
+For example, the command in order to install Patroni together with dependencies for Etcd as a DCS and AWS callbacks is:
+
+::
+
+    pip install patroni[etcd,aws]
+
+Note that external tools to call in the replica creation or custom bootstrap scripts (i.e. WAL-E) should be installed
+independently of Patroni.
+
+
+.. _running_configuring:
+
+Planning the Number of PostgreSQL Nodes
+---------------------------------------
+
+Patroni/PostgreSQL nodes are decoupled from DCS nodes (except when Patroni implements RAFT on its own) and therefore
+there is no requirement on the minimal number of nodes. Running a cluster consisting of one primary and one standby is
+perfectly fine. You can add more standby nodes later.
+
 Running and Configuring
-=======================
+-----------------------
+
+The following section assumes Patroni repository as being cloned from https://github.com/zalando/patroni. Namely, you
+will need example configuration files `postgres0.yml` and `postgres1.yml`. If you installed Patroni with pip, you can
+obtain those files from the git repository and replace `./patroni.py` below with `patroni` command.
 
 To get started, do the following from different terminals:
 ::
 
-    > etcd --data-dir=data/etcd
+    > etcd --data-dir=data/etcd --enable-v2=true
     > ./patroni.py postgres0.yml
     > ./patroni.py postgres1.yml
 
@@ -60,27 +141,27 @@ run:
 
     > psql --host 127.0.0.1 --port 5000 postgres
 
-==================
+
 YAML Configuration
-==================
+------------------
 
 Go :ref:`here <settings>` for comprehensive information about settings for etcd, consul, and ZooKeeper. And for an example, see `postgres0.yml <https://github.com/zalando/patroni/blob/master/postgres0.yml>`__.
 
-=========================
+
 Environment Configuration
-=========================
+-------------------------
 
 Go :ref:`here <environment>` for comprehensive information about configuring(overriding) settings via environment variables.
 
-===================
+
 Replication Choices
-===================
+-------------------
 
 Patroni uses Postgres' streaming replication, which is asynchronous by default. Patroni's asynchronous replication configuration allows for ``maximum_lag_on_failover`` settings. This setting ensures failover will not occur if a follower is more than a certain number of bytes behind the leader. This setting should be increased or decreased based on business requirements. It's also possible to use synchronous replication for better durability guarantees. See :ref:`replication modes documentation <replication_modes>` for details.
 
-======================================
+
 Applications Should Not Use Superusers
-======================================
+--------------------------------------
 
 When connecting from an application, always use a non-superuser. Patroni requires access to the database to function properly. By using a superuser from an application, you can potentially use the entire connection pool, including the connections reserved for superusers, with the ``superuser_reserved_connections`` setting. If Patroni cannot access the Primary because the connection pool is full, behavior will be undesirable.
 
@@ -88,3 +169,20 @@ When connecting from an application, always use a non-superuser. Patroni require
    :target: https://travis-ci.org/zalando/patroni
 .. |Coverage Status| image:: https://coveralls.io/repos/zalando/patroni/badge.svg?branch=master
    :target: https://coveralls.io/r/zalando/patroni?branch=master
+
+Testing Your HA Solution
+--------------------------------------
+Testing an HA solution is a time consuming process, with many variables. This is particularly true considering a cross-platform application. You need a trained system administrator or a consultant to do this work. It is not something we can cover in depth in the documentation.
+
+That said, here are some pieces of your infrastructure you should be sure to test:
+
+* Network (the network in front of your system as well as the NICs [physical or virtual] themselves)
+* Disk IO
+* file limits (nofile in Linux)
+* RAM. Even if you have oomkiller turned off as suggested, the unavailability of RAM could cause issues.
+* CPU
+* Virtualization Contention (overcommitting the hypervisor)
+* Any cgroup limitation (likely to be related to the above)
+* ``kill -9`` of any postgres process (except postmaster!). This is a decent simulation of a segfault.
+
+One thing that you should not do is run ``kill -9`` on a postmaster process. This is because doing so does not mimic any real life scenario. If you are concerned your infrastructure is insecure and an attacker could run ``kill -9``, no amount of HA process is going to fix that. The attacker will simply kill the process again, or cause chaos in another way.

docs/SETTINGS.rst

@@ -4,107 +4,406 @@
 YAML Configuration Settings
 ===========================
 
+.. _dynamic_configuration_settings:
+
+Dynamic configuration settings
+------------------------------
+
+Dynamic configuration is stored in the DCS (Distributed Configuration Store) and applied on all cluster nodes. Some parameters, like **loop_wait**, **ttl**, **postgresql.parameters.max_connections**, **postgresql.parameters.max_worker_processes** and so on could be set only in the dynamic configuration. Some other parameters like **postgresql.listen**, **postgresql.data_dir** could be set only locally, i.e. in the Patroni config file or via :ref:`configuration <environment>` variable. In most cases the local configuration will override the dynamic configuration. In order to change the dynamic configuration you can use either ``patronictl edit-config`` tool or Patroni :ref:`REST API <rest_api>`.
+
+-  **loop\_wait**: the number of seconds the loop will sleep. Default value: 10
+-  **ttl**: the TTL to acquire the leader lock (in seconds). Think of it as the length of time before initiation of the automatic failover process. Default value: 30
+-  **retry\_timeout**: timeout for DCS and PostgreSQL operation retries (in seconds). DCS or network issues shorter than this will not cause Patroni to demote the leader. Default value: 10
+-  **maximum\_lag\_on\_failover**: the maximum bytes a follower may lag to be able to participate in leader election.
+-  **maximum\_lag\_on\_syncnode**: the maximum bytes a synchronous follower may lag before it is considered as an unhealthy candidate and swapped by healthy asynchronous follower. Patroni utilize the max replica lsn if there is more than one follower, otherwise it will use leader's current wal lsn. Default is -1, Patroni will not take action to swap synchronous unhealthy follower when the value is set to 0 or below. Please set the value high enough so Patroni won't swap synchrounous follower fequently during high transaction volume.
+-  **max\_timelines\_history**: maximum number of timeline history items kept in DCS.  Default value: 0. When set to 0, it keeps the full history in DCS.
+-  **primary\_start\_timeout**: the amount of time a primary is allowed to recover from failures before failover is triggered (in seconds). Default is 300 seconds. When set to 0 failover is done immediately after a crash is detected if possible. When using asynchronous replication a failover can cause lost transactions. Worst case failover time for primary failure is: loop\_wait + primary\_start\_timeout + loop\_wait, unless primary\_start\_timeout is zero, in which case it's just loop\_wait. Set the value according to your durability/availability tradeoff.
+- **primary\_stop\_timeout**: The number of seconds Patroni is allowed to wait when stopping Postgres and effective only when synchronous_mode is enabled. When set to > 0 and the synchronous_mode is enabled, Patroni sends SIGKILL to the postmaster if the stop operation is running for more than the value set by primary\_stop\_timeout. Set the value according to your durability/availability tradeoff. If the parameter is not set or set <= 0, primary\_stop\_timeout does not apply.
+-  **synchronous\_mode**: turns on synchronous replication mode. In this mode a replica will be chosen as synchronous and only the latest leader and synchronous replica are able to participate in leader election. Synchronous mode makes sure that successfully committed transactions will not be lost at failover, at the cost of losing availability for writes when Patroni cannot ensure transaction durability. See :ref:`replication modes documentation <replication_modes>` for details.
+-  **synchronous\_mode\_strict**: prevents disabling synchronous replication if no synchronous replicas are available, blocking all client writes to the primary. See :ref:`replication modes documentation <replication_modes>` for details.
+-  **failsafe\_mode**: Enables :ref:`DCS Failsafe Mode <dcs_failsafe_mode>`. Defaults to `false`.
+-  **postgresql**:
+    -  **use\_pg\_rewind**: whether or not to use pg_rewind. Defaults to `false`.
+    -  **use\_slots**: whether or not to use replication slots. Defaults to `true` on PostgreSQL 9.4+.
+    -  **recovery\_conf**: additional configuration settings written to recovery.conf when configuring follower. There is no recovery.conf anymore in PostgreSQL 12, but you may continue using this section, because Patroni handles it transparently.
+    -  **parameters**: list of configuration settings for Postgres.
+-  **standby\_cluster**: if this section is defined, we want to bootstrap a standby cluster.
+    -  **host**: an address of remote node
+    -  **port**: a port of remote node
+    -  **primary\_slot\_name**: which slot on the remote node to use for replication. This parameter is optional, the default value is derived from the instance name (see function `slot_name_from_member_name`).
+    -  **create\_replica\_methods**: an ordered list of methods that can be used to bootstrap standby leader from the remote primary, can be different from the list defined in :ref:`postgresql_settings`
+    -  **restore\_command**: command to restore WAL records from the remote primary to nodes in a standby cluster, can be different from the list defined in :ref:`postgresql_settings`
+    -  **archive\_cleanup\_command**: cleanup command for standby leader
+    -  **recovery\_min\_apply\_delay**: how long to wait before actually apply WAL records on a standby leader
+-  **slots**: define permanent replication slots. These slots will be preserved during switchover/failover. The logical slots are copied from the primary to a standby with restart, and after that their position advanced every **loop_wait** seconds (if necessary). Copying logical slot files performed via ``libpq`` connection and using either rewind or superuser credentials (see **postgresql.authentication** section). There is always a chance that the logical slot position on the replica is a bit behind the former primary, therefore application should be prepared that some messages could be received the second time after the failover. The easiest way of doing so - tracking ``confirmed_flush_lsn``. Enabling permanent logical replication slots requires **postgresql.use_slots** to be set and will also automatically enable the ``hot_standby_feedback``. Since the failover of logical replication slots is unsafe on PostgreSQL 9.6 and older and PostgreSQL version 10 is missing some important functions, the feature only works with PostgreSQL 11+.
+    -  **my_slot_name**: the name of replication slot. If the permanent slot name matches with the name of the current primary it will not be created. Everything else is the responsibility of the operator to make sure that there are no clashes in names between replication slots automatically created by Patroni for members and permanent replication slots.
+        -  **type**: slot type. Could be ``physical`` or ``logical``. If the slot is logical, you have to additionally define ``database`` and ``plugin``.
+        -  **database**: the database name where logical slots should be created.
+        -  **plugin**: the plugin name for the logical slot.
+-  **ignore_slots**: list of sets of replication slot properties for which Patroni should ignore matching slots. This configuration/feature/etc. is useful when some replication slots are managed outside of Patroni. Any subset of matching properties will cause a slot to be ignored.
+    -  **name**: the name of the replication slot.
+    -  **type**: slot type. Can be ``physical`` or ``logical``. If the slot is logical, you may additionally define ``database`` and/or ``plugin``.
+    -  **database**: the database name (when matching a ``logical`` slot).
+    -  **plugin**: the logical decoding plugin (when matching a ``logical`` slot).
+
+Note: **slots** is a hashmap while **ignore_slots** is an array. For example:
+
+.. code:: YAML
+
+        slots:
+          permanent_logical_slot_name:
+            type: logical
+            database: my_db
+            plugin: test_decoding
+          permanent_physical_slot_name:
+            type: physical
+          ...
+        ignore_slots:
+          - name: ignored_logical_slot_name
+            type: logical
+            database: my_db
+            plugin: test_decoding
+          - name: ignored_physical_slot_name
+            type: physical
+          ...
+
 Global/Universal
 ----------------
 -  **name**: the name of the host. Must be unique for the cluster.
 -  **namespace**: path within the configuration store where Patroni will keep information about the cluster. Default value: "/service"
 -  **scope**: cluster name
 
+Log
+---
+-  **level**: sets the general logging level. Default value is **INFO** (see `the docs for Python logging <https://docs.python.org/3.6/library/logging.html#levels>`_)
+-  **traceback\_level**: sets the level where tracebacks will be visible. Default value is **ERROR**. Set it to **DEBUG** if you want to see tracebacks only if you enable **log.level=DEBUG**.
+-  **format**: sets the log formatting string. Default value is **%(asctime)s %(levelname)s: %(message)s** (see `the LogRecord attributes <https://docs.python.org/3.6/library/logging.html#logrecord-attributes>`_)
+-  **dateformat**: sets the datetime formatting string. (see the `formatTime() documentation <https://docs.python.org/3.6/library/logging.html#logging.Formatter.formatTime>`_)
+-  **max\_queue\_size**: Patroni is using two-step logging. Log records are written into the in-memory queue and there is a separate thread which pulls them from the queue and writes to stderr or file. The maximum size of the internal queue is limited by default by **1000** records, which is enough to keep logs for the past 1h20m.
+-  **dir**: Directory to write application logs to. The directory must exist and be writable by the user executing Patroni. If you set this value, the application will retain 4 25MB logs by default. You can tune those retention values with `file_num` and `file_size` (see below).
+-  **file\_num**: The number of application logs to retain.
+-  **file\_size**: Size of patroni.log file (in bytes) that triggers a log rolling.
+-  **loggers**: This section allows redefining logging level per python module
+    -  **patroni.postmaster: WARNING**
+    -  **urllib3: DEBUG**
+
+.. _bootstrap_settings:
+
 Bootstrap configuration
 -----------------------
--  **dcs**: This section will be written into `/<namespace>/<scope>/config` of a given configuration store after initializing of new cluster. This is the global configuration for the cluster. If you want to change some parameters for all cluster nodes - just do it in DCS (or via Patroni API) and all nodes will apply this configuration.
-    -  **loop\_wait**: the number of seconds the loop will sleep. Default value: 10
-    -  **ttl**: the TTL to acquire the leader lock. Think of it as the length of time before initiation of the automatic failover process. Default value: 30
-    -  **maximum\_lag\_on\_failover**: the maximum bytes a follower may lag to be able to participate in leader election.
-    -  **master\_start\_timeout**: the amount of time a master is allowed to recover from failures before failover is triggered. Default is 300 seconds. When set to 0 failover is done immediately after a crash is detected if possible. When using asynchronous replication a failover can cause lost transactions. Best worst case failover time for master failure is: loop\_wait + master\_start\_timeout + loop\_wait, unless master\_start\_timeout is zero, in which case it's just loop\_wait. Set the value according to your durability/availability tradeoff.
-    -  **synchronous\_mode**: turns on synchronous replication mode. In this mode a replica will be chosen as synchronous and only the latest leader and synchronous replica are able to participate in leader election. Synchronous mode makes sure that succesfully committed transactions will not be lost at failover, at the cost of losing availability for writes when Patroni cannot ensure transaction durability. See `replication modes documentation <https://github.com/zalando/patroni/blob/master/docs/replication_modes.rst>`__ for details.
-    -  **postgresql**:
-        -  **use\_pg\_rewind**:whether or not to use pg_rewind
-        -  **use\_slots**: whether or not to use replication_slots. Must be False for PostgreSQL 9.3. You should comment out max_replication_slots before it becomes ineligible for leader status.
-        -  **recovery\_conf**: additional configuration settings written to recovery.conf when configuring follower. 
-        -  **parameters**: list of configuration settings for Postgres. Many of these are required for replication to work.
--  **initdb**: List options to be passed on to initdb.
-        -  **- data-checksums**: Must be enabled when pg_rewind is needed on 9.3.
-        -  **- encoding: UTF8**: default encoding for new databases.
-        -  **- locale: UTF8**: default locale for new databases.
--  **pg\_hba**: list of lines that you should add to pg\_hba.conf.
-        -  **- host all all 0.0.0.0/0 md5**.
-        -  **- host replication replicator 127.0.0.1/32 md5**: A line like this is required for replication.
--  **users**: Some additional users users which needs to be created after initializing new cluster
-    -  **admin**: the name of user
-        -  **password: zalando**:
-        -  **options**: list of options for CREATE USER statement
-            -  **- createrole**
-            -  **- createdb**
--  **post_init**: An additional script that will be executed after initializing the cluster. The script receives a connection string URL (with the cluster superuser as a user name). The PGPASSFILE variable is set to the location of pgpass file.
+-  **bootstrap**:
+    -  **dcs**: This section will be written into `/<namespace>/<scope>/config` of the given configuration store after initializing of new cluster. The global dynamic configuration for the cluster. Under the ``bootstrap.dcs`` you can put any of the parameters described in the :ref:`Dynamic Configuration settings <dynamic_configuration_settings>` and after Patroni initialized (bootstrapped) the new cluster, it will write this section into `/<namespace>/<scope>/config` of the configuration store. All later changes of ``bootstrap.dcs`` will not take any effect! If you want to change them please use either ``patronictl edit-config`` or Patroni :ref:`REST API <rest_api>`.
+    -  **method**: custom script to use for bootstrapping this cluster.
+       See :ref:`custom bootstrap methods documentation <custom_bootstrap>` for details.
+       When ``initdb`` is specified revert to the default ``initdb`` command. ``initdb`` is also triggered when no ``method``
+       parameter is present in the configuration file.
+    -  **initdb**: List options to be passed on to initdb.
+            -  **- data-checksums**: Must be enabled when pg_rewind is needed on 9.3.
+            -  **- encoding: UTF8**: default encoding for new databases.
+            -  **- locale: UTF8**: default locale for new databases.
+    -  **pg\_hba**: list of lines that you should add to pg\_hba.conf.
+            -  **- host all all 0.0.0.0/0 md5**.
+            -  **- host replication replicator 127.0.0.1/32 md5**: A line like this is required for replication.
+    -  **users**: Some additional users which need to be created after initializing new cluster
+        -  **admin**: the name of user
+            -  **password: zalando**:
+            -  **options**: list of options for CREATE USER statement
+                -  **- createrole**
+                -  **- createdb**
+    -  **post\_bootstrap** or **post\_init**: An additional script that will be executed after initializing the cluster. The script receives a connection string URL (with the cluster superuser as a user name). The PGPASSFILE variable is set to the location of pgpass file.
+
+.. _citus_settings:
+
+Citus
+-----
+Enables integration Patroni with `Citus <https://docs.citusdata.com>`__. If configured, Patroni will take care of registering Citus worker nodes on the coordinator. You can find more information about Citus support :ref:`here <citus>`.
+
+-  **group**: the Citus group id, integer. Use ``0`` for coordinator and ``1``, ``2``, etc... for workers
+-  **database**: the database where ``citus`` extension should be created. Must be the same on the coordinator and all workers. Currently only one database is supported.
+
+.. _consul_settings:
 
 Consul
 ------
--  **host**: the host:port for the Consul endpoint.
+Most of the parameters are optional, but you have to specify one of the **host** or **url**
+
+-  **host**: the host:port for the Consul local agent.
+-  **url**: url for the Consul local agent, in format: http(s)://host:port.
+-  **port**: (optional) Consul port.
+-  **scheme**: (optional) **http** or **https**, defaults to **http**.
+-  **token**: (optional) ACL token.
+-  **verify**: (optional) whether to verify the SSL certificate for HTTPS requests.
+-  **cacert**: (optional) The ca certificate. If present it will enable validation.
+-  **cert**: (optional) file with the client certificate.
+-  **key**: (optional) file with the client key. Can be empty if the key is part of **cert**.
+-  **dc**: (optional) Datacenter to communicate with. By default the datacenter of the host is used.
+-  **consistency**: (optional) Select consul consistency mode. Possible values are ``default``, ``consistent``, or ``stale`` (more details in `consul API reference <https://www.consul.io/api/features/consistency.html/>`__)
+-  **checks**: (optional) list of Consul health checks used for the session. By default an empty list is used.
+-  **register\_service**: (optional) whether or not to register a service with the name defined by the scope parameter and the tag master, primary, replica, or standby-leader depending on the node's role. Defaults to **false**.
+-  **service\_tags**: (optional) additional static tags to add to the Consul service apart from the role (``master``/``primary``/``replica``/``standby-leader``). By default an empty list is used.
+-  **service\_check\_interval**: (optional) how often to perform health check against registered url. Defaults to '5s'.
+-  **service\_check\_tls\_server\_name**: (optional) overide SNI host when connecting via TLS, see also `consul agent check API reference <https://www.consul.io/api-docs/agent/check#tlsservername>`__.
+
+The ``token`` needs to have the following ACL permissions:
+
+::
+
+    service_prefix "${scope}" {
+        policy = "write"
+    }
+    key_prefix "${namespace}/${scope}" {
+        policy = "write"
+    }
+    session_prefix "" {
+        policy = "write"
+    }
 
 Etcd
 ----
-Most of the parameters are optional, but you have to specify one of the **host**, **url**, **proxy** or **srv**
+Most of the parameters are optional, but you have to specify one of the **host**, **hosts**, **url**, **proxy** or **srv**
+
 -  **host**: the host:port for the etcd endpoint.
--  **url**: url for the etcd
--  **proxy**: proxy url for the etcd. If you are connecting to the etcd using proxy, use this parameter instead of **url**
--  **srv**: Domain to search the SRV record(s) for cluster autodiscovery.
+-  **hosts**: list of etcd endpoint in format host1:port1,host2:port2,etc... Could be a comma separated string or an actual yaml list.
+-  **use\_proxies**: If this parameter is set to true, Patroni will consider **hosts** as a list of proxies and will not perform a topology discovery of etcd cluster.
+-  **url**: url for the etcd.
+-  **proxy**: proxy url for the etcd. If you are connecting to the etcd using proxy, use this parameter instead of **url**.
+-  **srv**: Domain to search the SRV record(s) for cluster autodiscovery. Patroni will try to query these SRV service names for specified domain (in that order until first success): ``_etcd-client-ssl``, ``_etcd-client``, ``_etcd-ssl``, ``_etcd``, ``_etcd-server-ssl``, ``_etcd-server``. If SRV records for ``_etcd-server-ssl`` or ``_etcd-server`` are retrieved then ETCD peer protocol is used do query ETCD for available members. Otherwise hosts from SRV records will be used.
+-  **srv\_suffix**: Configures a suffix to the SRV name that is queried during discovery. Use this flag to differentiate between multiple etcd clusters under the same domain. Works only with conjunction with **srv**. For example, if ``srv_suffix: foo`` and ``srv: example.org`` are set, the following DNS SRV query is made:``_etcd-client-ssl-foo._tcp.example.com`` (and so on for every possible ETCD SRV service name).
 -  **protocol**: (optional) http or https, if not specified http is used. If the **url** or **proxy** is specified - will take protocol from them.
--  **username**: (optional) username for etcd authentication
+-  **username**: (optional) username for etcd authentication.
 -  **password**: (optional) password for etcd authentication.
--  **cacert**: (optional) The ca certificate. If pressent it will enable validation.
--  **cert**: (optional) file with the client certificate
+-  **cacert**: (optional) The ca certificate. If present it will enable validation.
+-  **cert**: (optional) file with the client certificate.
 -  **key**: (optional) file with the client key. Can be empty if the key is part of **cert**.
 
+Etcdv3
+------
+If you want that Patroni works with Etcd cluster via protocol version 3, you need to use the ``etcd3`` section in the Patroni configuration file. All configuration parameters are the same as for ``etcd``.
+
+.. warning::
+    Keys created with protocol version 2 are not visible with protocol version 3 and the other way around, therefore it is not possible to switch from ``etcd`` to ``etcd3`` just by updating Patroni config file.
+
+
+ZooKeeper
+----------
+-  **hosts**: List of ZooKeeper cluster members in format: ['host1:port1', 'host2:port2', 'etc...'].
+-  **use_ssl**: (optional) Whether SSL is used or not. Defaults to ``false``. If set to ``false``, all SSL specific parameters are ignored.
+-  **cacert**: (optional) The CA certificate. If present it will enable validation.
+-  **cert**: (optional) File with the client certificate.
+-  **key**: (optional) File with the client key.
+-  **key_password**: (optional) The client key password.
+-  **verify**: (optional) Whether to verify certificate or not. Defaults to ``true``.
+-  **set_acls**: (optional) If set, configure Kazoo to apply a default ACL to each ZNode that it creates. ACLs will assume 'x509' schema and should be specified as a dictionary with the principal as the key and one or more permissions as a list in the value.  Permissions may be one of ``CREATE``, ``READ``, ``WRITE``, ``DELETE`` or ``ADMIN``.  For example, ``set_acls: {CN=principal1: [CREATE, READ], CN=principal2: [ALL]}``.
+
+.. note::
+    It is required to install ``kazoo>=2.6.0`` to support SSL.
+
+
 Exhibitor
 ---------
 -  **hosts**: initial list of Exhibitor (ZooKeeper) nodes in format: 'host1,host2,etc...'. This list updates automatically whenever the Exhibitor (ZooKeeper) cluster topology changes.
--  **poll\_interval**: how often the list of ZooKeeper and Exhibitor nodes should be updated from Exhibitor
+-  **poll\_interval**: how often the list of ZooKeeper and Exhibitor nodes should be updated from Exhibitor.
 -  **port**: Exhibitor port.
 
+.. _kubernetes_settings:
+
+Kubernetes
+----------
+-  **bypass\_api\_service**: (optional) When communicating with the Kubernetes API, Patroni is usually relying on the `kubernetes` service, the address of which is exposed in the pods via the `KUBERNETES_SERVICE_HOST` environment variable. If `bypass_api_service` is set to ``true``, Patroni will resolve the list of API nodes behind the service and connect directly to them.
+-  **namespace**: (optional) Kubernetes namespace where Patroni pod is running. Default value is `default`.
+-  **labels**: Labels in format ``{label1: value1, label2: value2}``. These labels will be used to find existing objects (Pods and either Endpoints or ConfigMaps) associated with the current cluster. Also Patroni will set them on every object (Endpoint or ConfigMap) it creates.
+-  **scope\_label**: (optional) name of the label containing cluster name. Default value is `cluster-name`.
+-  **role\_label**: (optional) name of the label containing role (master or replica). Patroni will set this label on the pod it runs in. Default value is ``role``.
+-  **use\_endpoints**: (optional) if set to true, Patroni will use Endpoints instead of ConfigMaps to run leader elections and keep cluster state.
+-  **pod\_ip**: (optional) IP address of the pod Patroni is running in. This value is required when `use_endpoints` is enabled and is used to populate the leader endpoint subsets when the pod's PostgreSQL is promoted.
+-  **ports**: (optional) if the Service object has the name for the port, the same name must appear in the Endpoint object, otherwise service won't work. For example, if your service is defined as ``{Kind: Service, spec: {ports: [{name: postgresql, port: 5432, targetPort: 5432}]}}``, then you have to set ``kubernetes.ports: [{"name": "postgresql", "port": 5432}]`` and Patroni will use it for updating subsets of the leader Endpoint. This parameter is used only if `kubernetes.use_endpoints` is set.
+-  **cacert**: (optional) Specifies the file with the CA_BUNDLE file with certificates of trusted CAs to use while verifying Kubernetes API SSL certs. If not provided, patroni will use the value provided by the ServiceAccount secret.
+
+
+.. _raft_settings:
+
+Raft (deprecated)
+-----------------
+-  **self\_addr**: ``ip:port`` to listen on for Raft connections. The ``self_addr`` must be accessible from other nodes of the cluster. If not set, the node will not participate in consensus.
+-  **bind\_addr**: (optional) ``ip:port`` to listen on for Raft connections. If not specified the ``self_addr`` will be used.
+-  **partner\_addrs**: list of other Patroni nodes in the cluster in format: ['ip1:port', 'ip2:port', 'etc...']
+-  **data\_dir**: directory where to store Raft log and snapshot. If not specified the current working directory is used.
+-  **password**: (optional) Encrypt Raft traffic with a specified password, requires ``cryptography`` python module.
+
+  Short FAQ about Raft implementation
+
+  - Q: How to list all the nodes providing consensus?
+
+    A: ``syncobj_admin -conn host:port -status`` where the host:port is the address of one of the cluster nodes
+
+  - Q: Node that was a part of consensus and has gone and I can't reuse the same IP for other node. How to remove this node from the consensus?
+
+    A: ``syncobj_admin -conn host:port -remove host2:port2`` where the ``host2:port2`` is the address of the node you want to remove from consensus.
+
+  - Q: Where to get the ``syncobj_admin`` utility?
+
+    A: It is installed together with ``pysyncobj`` module (python RAFT implementation), which is Patroni dependency.
+
+  - Q: it is possible to run Patroni node without adding in to the consensus?
+
+    A: Yes, just comment out or remove ``raft.self_addr`` from Patroni configuration.
+
+  - Q: It is possible to run Patroni and PostgreSQL only on two nodes?
+
+    A: Yes, on the third node you can run ``patroni_raft_controller`` (without Patroni and PostgreSQL). In such a setup, one can temporarily lose one node without affecting the primary.
+
+
 .. _postgresql_settings:
 
 PostgreSQL
 ----------
--  **authentication**:
-    -  **superuser**:
-        -  **username**: name for the superuser, set during initialization (initdb) and later used by Patroni to connect to the postgres.
-        -  **password**: password for the superuser, set during initialization (initdb).
-    -  **replication**:
-        -  **username**: replication username; the user will be created during initialization. Replicas will use this user to access master via streaming replication
-        -  **password**: replication password; the user will be created during initialization.
--  **callbacks**: callback scripts to run on certain actions. Patroni will pass the action, role and cluster name. (See scripts/aws.py as an example of how to write them.)
-        -  **on\_reload**: run this script when configuration reload is triggered.
-        -  **on\_restart**: run this script when the cluster restarts.
-        -  **on\_role\_change**: run this script when the cluster is being promoted or demoted.
-        -  **on\_start**: run this script when the cluster starts.
-        -  **on\_stop**: run this script when the cluster stops.
--  **connect\_address**: IP address + port through which Postgres is accessible from other nodes and applications.
--  **create\_replica\_methods**: an ordered list of the create methods for turning a Patroni node into a new replica. "basebackup" is the default method; other methods are assumed to refer to scripts, each of which is configured as its own config item.
--  **data\_dir**: The location of the Postgres data directory, either existing or to be initialized by Patroni.
--  **bin\_dir**: Path to PostgreSQL binaries. (pg_ctl, pg_rewind, pg_basebackup, postgres) The  default value is an empty string meaning that PATH environment variable will be used to find the executables.
--  **listen**: IP address + port that Postgres listens to; must be accessible from other nodes in the cluster, if you're using streaming replication. Multiple comma-separated addresses are permitted, as long as the port component is appended after to the last one with a colon, i.e. ``listen: 127.0.0.1,127.0.0.2:5432``. Patroni will use the first address from this list to establish local connections to the PostgreSQL node.
--  **pgpass**: path to the `.pgpass <https://www.postgresql.org/docs/current/static/libpq-pgpass.html>`__ password file. Patroni creates this file before executing pg\_basebackup, the post_init script and under some other circumstances. The location must be writable by Patroni.
--  **recovery\_conf**: additional configuration settings written to recovery.conf when configuring follower.
--  **custom_conf** : path to an optional custom ``postgresql.conf`` file, that will be used in place of ``postgresql.base.conf``. The file must exist on all cluster nodes, be readable by PostgreSQL and will be included from its location on the real ``postgresql.conf``. Note that Patroni will not monitor this file for changes, nor backup it. However, its settings can still be overriden by Patroni's own configuration facilities - see `dynamic configuration <https://github.com/zalando/patroni/blob/master/docs/dynamic_configuration.rst>`__ for details.
--  **parameters**: list of configuration settings for Postgres. Many of these are required for replication to work.
--  **pg\_ctl\_timeout**: How long should pg_ctl wait when doing ``start``, ``stop`` or ``restart``. Default value is 60 seconds.
--  **use\_pg\_rewind**: try to use pg\_rewind on the former leader when it joins cluster as a replica.
--  **remove\_data\_directory\_on\_rewind\_failure**: If this option is enabled, Patroni will remove postgres data directory and recreate replica. Otherwise it will try to follow the new leader. Default value is **false**.
--  **replica\_method** for each create_replica_method other than basebackup, you would add a configuration section of the same name. At a minimum, this should include "command" with a full path to the actual script to be executed.  Other configuration parameters will be passed along to the script in the form "parameter=value".
+-  **postgresql**:
+    -  **authentication**:
+        -  **superuser**:
+            -  **username**: name for the superuser, set during initialization (initdb) and later used by Patroni to connect to the postgres.
+            -  **password**: password for the superuser, set during initialization (initdb).
+            -  **sslmode**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+            -  **sslkey**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+            -  **sslpassword**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``sslkey``.
+            -  **sslcert**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+            -  **sslrootcert**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+            -  **sslcrl**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **sslcrldir**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **gssencmode**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+            -  **channel_binding**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
+        -  **replication**:
+            -  **username**: replication username; the user will be created during initialization. Replicas will use this user to access the replication source via streaming replication
+            -  **password**: replication password; the user will be created during initialization.
+            -  **sslmode**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+            -  **sslkey**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+            -  **sslpassword**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``sslkey``.
+            -  **sslcert**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+            -  **sslrootcert**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+            -  **sslcrl**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **sslcrldir**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **gssencmode**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+            -  **channel_binding**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
+        -  **rewind**:
+            -  **username**: name for the user for ``pg_rewind``; the user will be created during initialization of postgres 11+ and all necessary `permissions <https://www.postgresql.org/docs/11/app-pgrewind.html#id-1.9.5.8.8>`__ will be granted.
+            -  **password**: password for the user for ``pg_rewind``; the user will be created during initialization.
+            -  **sslmode**: (optional) maps to the `sslmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLMODE>`__ connection parameter, which allows a client to specify the type of TLS negotiation mode with the server. For more information on how each mode works, please visit the `PostgreSQL documentation <https://www.postgresql.org/docs/current/libpq-ssl.html#LIBPQ-SSL-SSLMODE-STATEMENTS>`__. The default mode is ``prefer``.
+            -  **sslkey**: (optional) maps to the `sslkey <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLKEY>`__ connection parameter, which specifies the location of the secret key used with the client's certificate.
+            -  **sslpassword**: (optional) maps to the `sslpassword <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLPASSWORD>`__ connection parameter, which specifies the password for the secret key specified in ``sslkey``.
+            -  **sslcert**: (optional) maps to the `sslcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCERT>`__ connection parameter, which specifies the location of the client certificate.
+            -  **sslrootcert**: (optional) maps to the `sslrootcert <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLROOTCERT>`__ connection parameter, which specifies the location of a file containing one ore more certificate authorities (CA) certificates that the client will use to verify a server's certificate.
+            -  **sslcrl**: (optional) maps to the `sslcrl <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRL>`__ connection parameter, which specifies the location of a file containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **sslcrldir**: (optional) maps to the `sslcrldir <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-SSLCRLDIR>`__ connection parameter, which specifies the location of a directory with files containing a certificate revocation list. A client will reject connecting to any server that has a certificate present in this list.
+            -  **gssencmode**: (optional) maps to the `gssencmode <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-GSSENCMODE>`__ connection parameter, which determines whether or with what priority a secure GSS TCP/IP connection will be negotiated with the server
+            -  **channel_binding**: (optional) maps to the `channel_binding <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNECT-CHANNEL-BINDING>`__ connection parameter, which controls the client's use of channel binding.
+    -  **callbacks**: callback scripts to run on certain actions. Patroni will pass the action, role and cluster name. (See scripts/aws.py as an example of how to write them.)
+            -  **on\_reload**: run this script when configuration reload is triggered.
+            -  **on\_restart**: run this script when the postgres restarts (without changing role).
+            -  **on\_role\_change**: run this script when the postgres is being promoted or demoted.
+            -  **on\_start**: run this script when the postgres starts.
+            -  **on\_stop**: run this script when the postgres stops.
+    -  **connect\_address**: IP address + port through which Postgres is accessible from other nodes and applications.
+    -  **proxy\_address**: IP address + port through which a connection pool (e.g. pgbouncer) running next to Postgres is accessible. The value is written to the member key in DCS as ``proxy_url`` and could be used/useful for service discovery.
+    -  **create\_replica\_methods**: an ordered list of the create methods for turning a Patroni node into a new replica.
+       "basebackup" is the default method; other methods are assumed to refer to scripts, each of which is configured as its
+       own config item. See :ref:`custom replica creation methods documentation <custom_replica_creation>` for further explanation.
+    -  **data\_dir**: The location of the Postgres data directory, either :ref:`existing <existing_data>` or to be initialized by Patroni.
+    -  **config\_dir**: The location of the Postgres configuration directory, defaults to the data directory. Must be writable by Patroni.
+    -  **bin\_dir**: Path to PostgreSQL binaries (pg_ctl, pg_rewind, pg_basebackup, postgres). The default value is an empty string meaning that PATH environment variable will be used to find the executables.
+    -  **listen**: IP address + port that Postgres listens to; must be accessible from other nodes in the cluster, if you're using streaming replication. Multiple comma-separated addresses are permitted, as long as the port component is appended after to the last one with a colon, i.e. ``listen: 127.0.0.1,127.0.0.2:5432``. Patroni will use the first address from this list to establish local connections to the PostgreSQL node.
+    -  **use\_unix\_socket**: specifies that Patroni should prefer to use unix sockets to connect to the cluster. Default value is ``false``. If ``unix_socket_directories`` is defined, Patroni will use the first suitable value from it to connect to the cluster and fallback to tcp if nothing is suitable. If ``unix_socket_directories`` is not specified in ``postgresql.parameters``, Patroni will assume that the default value should be used and omit ``host`` from the connection parameters.
+    -  **use\_unix\_socket\_repl**: specifies that Patroni should prefer to use unix sockets for replication user cluster connection. Default value is ``false``. If ``unix_socket_directories`` is defined, Patroni will use the first suitable value from it to connect to the cluster and fallback to tcp if nothing is suitable. If ``unix_socket_directories`` is not specified in ``postgresql.parameters``, Patroni will assume that the default value should be used and omit ``host`` from the connection parameters.
+    -  **pgpass**: path to the `.pgpass <https://www.postgresql.org/docs/current/static/libpq-pgpass.html>`__ password file. Patroni creates this file before executing pg\_basebackup, the post_init script and under some other circumstances. The location must be writable by Patroni.
+    -  **recovery\_conf**: additional configuration settings written to recovery.conf when configuring follower.
+    -  **custom\_conf** : path to an optional custom ``postgresql.conf`` file, that will be used in place of ``postgresql.base.conf``. The file must exist on all cluster nodes, be readable by PostgreSQL and will be included from its location on the real ``postgresql.conf``. Note that Patroni will not monitor this file for changes, nor backup it. However, its settings can still be overridden by Patroni's own configuration facilities - see :ref:`dynamic configuration <dynamic_configuration>` for details.
+    -  **parameters**: list of configuration settings for Postgres. Many of these are required for replication to work.
+    -  **pg\_hba**: list of lines that Patroni will use to generate ``pg_hba.conf``. This parameter has higher priority than ``bootstrap.pg_hba``. Together with :ref:`dynamic configuration <dynamic_configuration>` it simplifies management of ``pg_hba.conf``.
+            -  **- host all all 0.0.0.0/0 md5**.
+            -  **- host replication replicator 127.0.0.1/32 md5**: A line like this is required for replication.
+    -  **pg\_ident**: list of lines that Patroni will use to generate ``pg_ident.conf``. Together with :ref:`dynamic configuration <dynamic_configuration>` it simplifies management of ``pg_ident.conf``.
+            -  **- mapname1 systemname1 pguser1**.
+            -  **- mapname1 systemname2 pguser2**.
+    -  **pg\_ctl\_timeout**: How long should pg_ctl wait when doing ``start``, ``stop`` or ``restart``. Default value is 60 seconds.
+    -  **use\_pg\_rewind**: try to use pg\_rewind on the former leader when it joins cluster as a replica.
+    -  **remove\_data\_directory\_on\_rewind\_failure**: If this option is enabled, Patroni will remove the PostgreSQL data directory and recreate the replica. Otherwise it will try to follow the new leader. Default value is **false**.
+    -  **remove\_data\_directory\_on\_diverged\_timelines**: Patroni will remove the PostgreSQL data directory and recreate the replica if it notices that timelines are diverging and the former primary can not start streaming from the new primary. This option is useful when ``pg_rewind`` can not be used. While performing timelines divergence check on PostgreSQL v10 and older Patroni will try to connect with replication credential to the "postgres" database. Hence, such access should be allowed in the  pg_hba.conf. Default value is **false**.
+    -  **replica\_method**: for each create_replica_methods other than basebackup, you would add a configuration section of the same name. At a minimum, this should include "command" with a full path to the actual script to be executed. Other configuration parameters will be passed along to the script in the form "parameter=value".
+    -  **pre\_promote**: a fencing script that executes during a failover after acquiring the leader lock but before promoting the replica. If the script exits with a non-zero code, Patroni does not promote the replica and removes the leader key from DCS.
+
+.. _restapi_settings:
 
 REST API
--------- 
--  **connect\_address**: IP address and port to access the REST API.
--  **listen**: IP address and port that Patroni will listen to, to provide health-check information for HAProxy.
--  **Optional**:
-        -  **authentication**:
+--------
+-  **restapi**:
+        -  **connect\_address**: IP address (or hostname) and port, to access the Patroni's :ref:`REST API <rest_api>`. All the members of the cluster must be able to connect to this address, so unless the Patroni setup is intended for a demo inside the localhost, this address must be a non "localhost" or loopback address (ie: "localhost" or "127.0.0.1"). It can serve as an endpoint for HTTP health checks (read below about the "listen" REST API parameter), and also for user queries (either directly or via the REST API), as well as for the health checks done by the cluster members during leader elections (for example, to determine whether the leader is still running, or if there is a node which has a WAL position that is ahead of the one doing the query; etc.) The connect_address is put in the member key in DCS, making it possible to translate the member name into the address to connect to its REST API.
+
+        -  **listen**: IP address (or hostname) and port that Patroni will listen to for the REST API - to provide also the same health checks and cluster messaging between the participating nodes, as described above. to provide health-check information for HAProxy (or any other load balancer capable of doing a HTTP "OPTION" or "GET" checks).
+
+        -  **authentication**: (optional)
             -  **username**: Basic-auth username to protect unsafe REST API endpoints.
             -  **password**: Basic-auth password to protect unsafe REST API endpoints.
+        -  **certfile**: (optional): Specifies the file with the certificate in the PEM format. If the certfile is not specified or is left empty, the API server will work without SSL.
+        -  **keyfile**: (optional): Specifies the file with the secret key in the PEM format.
+        -  **keyfile\_password**: (optional): Specifies a password for decrypting the keyfile.
+        -  **cafile**: (optional): Specifies the file with the CA_BUNDLE with certificates of trusted CAs to use while verifying client certs.
+        -  **ciphers**: (optional): Specifies the permitted cipher suites (e.g. "ECDHE-RSA-AES256-GCM-SHA384:DHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-GCM-SHA256:DHE-RSA-AES128-GCM-SHA256:!SSLv1:!SSLv2:!SSLv3:!TLSv1:!TLSv1.1")
+        -  **verify\_client**: (optional): ``none`` (default), ``optional`` or ``required``. When ``none`` REST API will not check client certificates. When ``required`` client certificates are required for all REST API calls. When ``optional`` client certificates are required for all unsafe REST API endpoints. When ``required`` is used, then client authentication succeeds, if the certificate signature verification succeeds.  For ``optional`` the client cert will only be checked for ``PUT``, ``POST``, ``PATCH``, and ``DELETE`` requests.
+        -  **allowlist**: (optional): Specifies the set of hosts that are allowed to call unsafe REST API endpoints. The single element could be a host name, an IP address or a network address using CIDR notation. By default ``allow all`` is used. In case if ``allowlist`` or ``allowlist_include_members`` are set, anything that is not included is rejected.
+        -  **allowlist\_include\_members**: (optional): If set to ``true`` it allows accessing unsafe REST API endpoints from other cluster members registered in DCS (IP address or hostname is taken from the members ``api_url``). Be careful, it might happen that OS will use a different IP for outgoing connections.
+        -  **http\_extra\_headers**: (optional): HTTP headers let the REST API server pass additional information with an HTTP response.
+        -  **https\_extra\_headers**: (optional): HTTPS headers let the REST API server pass additional information with an HTTP response when TLS is enabled. This will also pass additional information set in ``http_extra_headers``.
 
-        -  **certfile**: Specifies the file with the certificate in the PEM format. If the certfile is not specified or is left empty, the API server will work without SSL.
-        -  **keyfile**: Specifies the file with the secret key in the PEM format.
+Here is an example of both **http_extra_headers** and **https_extra_headers**:
 
-ZooKeeper
-----------
--  **hosts**: list of ZooKeeper cluster members in format: ['host1:port1', 'host2:port2', 'etc...'].
+.. code:: YAML
+
+        restapi:
+          listen: <listen>
+          connect_address: <connect_address>
+          authentication:
+            username: <username>
+            password: <password>
+          http_extra_headers:
+            'X-Frame-Options': 'SAMEORIGIN'
+            'X-XSS-Protection': '1; mode=block'
+            'X-Content-Type-Options': 'nosniff'
+          cafile: <ca file>
+          certfile: <cert>
+          keyfile: <key>
+          https_extra_headers:
+            'Strict-Transport-Security': 'max-age=31536000; includeSubDomains'
+
+.. _patronictl_settings:
+
+CTL
+---
+- **ctl**: (optional)
+    -  **insecure**: Allow connections to REST API without verifying SSL certs.
+    -  **cacert**: Specifies the file with the CA_BUNDLE file or directory with certificates of trusted CAs to use while verifying REST API SSL certs. If not provided patronictl will use the value provided for REST API "cafile" parameter.
+    -  **certfile**: Specifies the file with the client certificate in the PEM format. If not provided patronictl will use the value provided for REST API "certfile" parameter.
+    -  **keyfile**: Specifies the file with the client secret key in the PEM format. If not provided patronictl will use the value provided for REST API "keyfile" parameter.
+    -  **keyfile\_password**: Specifies a password for decrypting the keyfile. If not provided patronictl will use the value provided for REST API "keyfile\_password" parameter.
+
+Watchdog
+--------
+- **mode**: ``off``, ``automatic`` or ``required``. When ``off`` watchdog is disabled. When ``automatic`` watchdog will be used if available, but ignored if it is not. When ``required`` the node will not become a leader unless watchdog can be successfully enabled.
+- **device**: Path to watchdog device. Defaults to ``/dev/watchdog``.
+- **safety_margin**: Number of seconds of safety margin between watchdog triggering and leader key expiration.
+
+.. _tags_settings:
+
+Tags
+----
+- **nofailover**: ``true`` or ``false``, controls whether this node is allowed to participate in the leader race and become a leader. Defaults to ``false``
+- **clonefrom**: ``true`` or ``false``. If set to ``true`` other nodes might prefer to use this node for bootstrap (take ``pg_basebackup`` from). If there are several nodes with ``clonefrom`` tag set to ``true`` the node to bootstrap from will be chosen randomly. The default value is ``false``.
+- **noloadbalance**: ``true`` or ``false``. If set to ``true`` the node will return HTTP Status Code 503 for the ``GET /replica`` REST API health-check and therefore will be excluded from the load-balancing. Defaults to ``false``.
+- **replicatefrom**: The IP address/hostname of another replica. Used to support cascading replication.
+- **nosync**: ``true`` or ``false``. If set to ``true`` the node will never be selected as a synchronous replica.
+
+In addition to these predefined tags, you can also add your own ones:
+
+- **key1**: ``true``
+- **key2**: ``false``
+- **key3**: ``1.4``
+- **key4**: ``"RandomString"``
+
+Tags are visible in the :ref:`REST API <rest_api>` and ``patronictl list`` You can also check for an instance health using these tags. If the tag isn't defined for an instance, or if the respective value doesn't match the querying value, it will return HTTP Status Code 503.

docs/citus.rst

@@ -0,0 +1,355 @@
+.. _citus:
+
+Citus support
+=============
+
+Patroni makes it extremely simple to deploy `Multi-Node Citus`__ clusters.
+
+__ https://docs.citusdata.com/en/stable/installation/multi_node.html
+
+TL;DR
+-----
+
+There are only a few simple rules you need to follow:
+
+1. `Citus <https://github.com/citusdata/citus>`__ database extension to
+   PostgreSQL must be available on all nodes.  Absolute minimum supported Citus
+   version is 10.0, but, to take all benefits from transparent switchovers and
+   restarts of workers we recommend using at least Citus 11.2.
+2. Cluster name (``scope``) must be the same for all Citus nodes!
+3. Superuser credentials must be the same on coordinator and all worker
+   nodes, and ``pg_hba.conf`` should allow superuser access between all nodes.
+4. :ref:`REST API <restapi_settings>` access should be allowed from worker
+   nodes to the coordinator. E.g., credentials should be the same and if
+   configured, client certificates from worker nodes must be accepted by the
+   coordinator.
+5. Add the following section to the ``patroni.yaml``:
+
+.. code:: YAML
+
+        citus:
+          group: X  # 0 for coordinator and 1, 2, 3, etc for workers
+          database: citus  # must be the same on all nodes
+
+
+After that you just need to start Patroni and it will handle the rest:
+
+1. ``citus`` extension will be automatically added to ``shared_preload_libraries``.
+2. If ``max_prepared_transactions`` isn't explicitly set in the global
+   :ref:`dynamic configuration <dynamic_configuration>` Patroni will
+   automatically set it to ``2*max_connections``.
+3. The ``citus.database`` will be automatically created followed by ``CREATE EXTENSION citus``.
+4. Current superuser :ref:`credentials <postgresql_settings>` will be added to the ``pg_dist_authinfo``
+   table to allow cross-node communication. Don't forget to update them if
+   later you decide to change superuser username/password/sslcert/sslkey!
+5. The coordinator primary node will automatically discover worker primary
+   nodes and add them to the ``pg_dist_node`` table using the
+   ``citus_add_node()`` function.
+6. Patroni will also maintain ``pg_dist_node`` in case failover/switchover
+   on the coordinator or worker clusters occurs.
+
+patronictl
+----------
+
+Coordinator and worker clusters are physically different PostgreSQL/Patroni
+clusters that are just logically groupped together using the
+`Citus <https://github.com/citusdata/citus>`__ database extension to
+PostgreSQL. Therefore in most cases it is not possible to manage them as a
+single entity.
+
+It results in two major differences in ``patronictl`` behaviour when
+``patroni.yaml`` has the ``citus`` section comparing with the usual:
+
+1. The ``list`` and the ``topology`` by default output all members of the Citus
+   formation (coordinators and workers). The new column ``Group`` indicates
+   which Citus group they belong to.
+2. For all ``patronictl`` commands the new option is introduced, named
+   ``--group``. For some commands the default value for the group might be
+   taken from the ``patroni.yaml``. For example, ``patronictl pause`` will
+   enable the maintenance mode by default for the ``group`` that is set in the
+   ``citus`` section, but for example for ``patronictl  switchover`` or
+   ``patronictl remove`` the group must be explicitly specified.
+
+An example of ``patronictl list`` output for the Citus cluster::
+
+    postgres@coord1:~$ patronictl list demo
+    + Citus cluster: demo ----------+--------------+---------+----+-----------+
+    | Group | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    |     0 | coord1  | 172.27.0.10 | Replica      | running |  1 |         0 |
+    |     0 | coord2  | 172.27.0.6  | Sync Standby | running |  1 |         0 |
+    |     0 | coord3  | 172.27.0.4  | Leader       | running |  1 |           |
+    |     1 | work1-1 | 172.27.0.8  | Sync Standby | running |  1 |         0 |
+    |     1 | work1-2 | 172.27.0.2  | Leader       | running |  1 |           |
+    |     2 | work2-1 | 172.27.0.5  | Sync Standby | running |  1 |         0 |
+    |     2 | work2-2 | 172.27.0.7  | Leader       | running |  1 |           |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+
+If we add the ``--group`` option, the output will change to::
+
+    postgres@coord1:~$ patronictl list demo --group 0
+    + Citus cluster: demo (group: 0, 7179854923829112860) -----------+
+    | Member | Host        | Role         | State   | TL | Lag in MB |
+    +--------+-------------+--------------+---------+----+-----------+
+    | coord1 | 172.27.0.10 | Replica      | running |  1 |         0 |
+    | coord2 | 172.27.0.6  | Sync Standby | running |  1 |         0 |
+    | coord3 | 172.27.0.4  | Leader       | running |  1 |           |
+    +--------+-------------+--------------+---------+----+-----------+
+
+    postgres@coord1:~$ patronictl list demo --group 1
+    + Citus cluster: demo (group: 1, 7179854923881963547) -----------+
+    | Member  | Host       | Role         | State   | TL | Lag in MB |
+    +---------+------------+--------------+---------+----+-----------+
+    | work1-1 | 172.27.0.8 | Sync Standby | running |  1 |         0 |
+    | work1-2 | 172.27.0.2 | Leader       | running |  1 |           |
+    +---------+------------+--------------+---------+----+-----------+
+
+Citus worker switchover
+-----------------------
+
+When a switchover is orchestrated for a Citus worker node, Citus offers the
+opportunity to make the switchover close to transparent for an application.
+Because the application connects to the coordinator, which in turn connects to
+the worker nodes, then it is possible with Citus to `pause` the SQL traffic on
+the coordinator for the shards hosted on a worker node. The switchover then
+happens while the traffic is kept on the coordinator, and resumes as soon as a
+new primary worker node is ready to accept read-write queries.
+
+An example of ``patronictl switchover`` on the worker cluster::
+
+    postgres@coord1:~$ patronictl switchover demo
+    + Citus cluster: demo ----------+--------------+---------+----+-----------+
+    | Group | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    |     0 | coord1  | 172.27.0.10 | Replica      | running |  1 |         0 |
+    |     0 | coord2  | 172.27.0.6  | Sync Standby | running |  1 |         0 |
+    |     0 | coord3  | 172.27.0.4  | Leader       | running |  1 |           |
+    |     1 | work1-1 | 172.27.0.8  | Leader       | running |  1 |           |
+    |     1 | work1-2 | 172.27.0.2  | Sync Standby | running |  1 |         0 |
+    |     2 | work2-1 | 172.27.0.5  | Sync Standby | running |  1 |         0 |
+    |     2 | work2-2 | 172.27.0.7  | Leader       | running |  1 |           |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    Citus group: 2
+    Primary [work2-2]:
+    Candidate ['work2-1'] []:
+    When should the switchover take place (e.g. 2022-12-22T08:02 )  [now]:
+    Current cluster topology
+    + Citus cluster: demo (group: 2, 7179854924063375386) -----------+
+    | Member  | Host       | Role         | State   | TL | Lag in MB |
+    +---------+------------+--------------+---------+----+-----------+
+    | work2-1 | 172.27.0.5 | Sync Standby | running |  1 |         0 |
+    | work2-2 | 172.27.0.7 | Leader       | running |  1 |           |
+    +---------+------------+--------------+---------+----+-----------+
+    Are you sure you want to switchover cluster demo, demoting current primary work2-2? [y/N]: y
+    2022-12-22 07:02:40.33003 Successfully switched over to "work2-1"
+    + Citus cluster: demo (group: 2, 7179854924063375386) ------+
+    | Member  | Host       | Role    | State   | TL | Lag in MB |
+    +---------+------------+---------+---------+----+-----------+
+    | work2-1 | 172.27.0.5 | Leader  | running |  1 |           |
+    | work2-2 | 172.27.0.7 | Replica | stopped |    |   unknown |
+    +---------+------------+---------+---------+----+-----------+
+
+    postgres@coord1:~$ patronictl list demo
+    + Citus cluster: demo ----------+--------------+---------+----+-----------+
+    | Group | Member  | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+    |     0 | coord1  | 172.27.0.10 | Replica      | running |  1 |         0 |
+    |     0 | coord2  | 172.27.0.6  | Sync Standby | running |  1 |         0 |
+    |     0 | coord3  | 172.27.0.4  | Leader       | running |  1 |           |
+    |     1 | work1-1 | 172.27.0.8  | Leader       | running |  1 |           |
+    |     1 | work1-2 | 172.27.0.2  | Sync Standby | running |  1 |         0 |
+    |     2 | work2-1 | 172.27.0.5  | Leader       | running |  2 |           |
+    |     2 | work2-2 | 172.27.0.7  | Sync Standby | running |  2 |         0 |
+    +-------+---------+-------------+--------------+---------+----+-----------+
+
+And this is how it looks on the coordinator side::
+
+    # The worker primary notifies the coordinator that it is going to execute "pg_ctl stop".
+    2022-12-22 07:02:38,636 DEBUG: query("BEGIN")
+    2022-12-22 07:02:38,636 DEBUG: query("SELECT pg_catalog.citus_update_node(3, '172.27.0.7-demoted', 5432, true, 10000)")
+    # From this moment all application traffic on the coordinator to the worker group 2 is paused.
+
+    # The future worker primary notifies the coordinator that it acquired the leader lock in DCS and about to run "pg_ctl promote".
+    2022-12-22 07:02:40,085 DEBUG: query("SELECT pg_catalog.citus_update_node(3, '172.27.0.5', 5432)")
+
+    # The new worker primary just finished promote and notifies coordinator that it is ready to accept read-write traffic.
+    2022-12-22 07:02:41,485 DEBUG: query("COMMIT")
+    # From this moment the application traffic on the coordinator to the worker group 2 is unblocked.
+
+Peek into DCS
+-------------
+
+The Citus cluster (coordinator and workers) are stored in DCS as a fleet of
+Patroni clusters logically grouped together::
+
+    /service/batman/              # scope=batman
+    /service/batman/0/            # citus.group=0, coordinator
+    /service/batman/0/initialize
+    /service/batman/0/leader
+    /service/batman/0/members/
+    /service/batman/0/members/m1
+    /service/batman/0/members/m2
+    /service/batman/1/            # citus.group=1, worker
+    /service/batman/1/initialize
+    /service/batman/1/leader
+    /service/batman/1/members/
+    /service/batman/1/members/m3
+    /service/batman/1/members/m4
+    ...
+
+Such an approach was chosen because for most DCS it becomes possible to fetch
+the entire Citus cluster with a single recursive read request. Only Citus
+coordinator nodes are reading the whole tree, because they have to discover
+worker nodes. Worker nodes are reading only the subtree for their own group and
+in some cases they could read the subtree of the coordinator group.
+
+Citus on Kubernetes
+-------------------
+
+Since Kubernetes doesn't support hierarchical structures we had to include the
+citus group to all K8s objects Patroni creates::
+
+    batman-0-leader  # the leader config map for the coordinator
+    batman-0-config  # the config map holding initialize, config, and history "keys"
+    ...
+    batman-1-leader  # the leader config map for worker group 1
+    batman-1-config
+    ...
+
+I.e., the naming pattern is: ``${scope}-${citus.group}-${type}``.
+
+All Kubernetes objects are discovered by Patroni using the `label selector`__,
+therefore all Pods with Patroni&Citus and Endpoints/ConfigMaps must have
+similar labels, and Patroni must be configured to use them using Kubernetes
+:ref:`settings <kubernetes_settings>` or :ref:`environment variables
+<kubernetes_environment>`.
+
+__ https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#label-selectors
+
+A couple of examples of Patroni configuration using Pods environment variables:
+
+1. for the coordinator cluster
+
+.. code:: YAML
+
+        apiVersion: v1
+        kind: Pod
+        metadata:
+          labels:
+            application: patroni
+            citus-group: "0"
+            citus-type: coordinator
+            cluster-name: citusdemo
+          name: citusdemo-0-0
+          namespace: default
+        spec:
+          containers:
+          - env:
+            - name: PATRONI_SCOPE
+              value: citusdemo
+            - name: PATRONI_NAME
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: metadata.name
+            - name: PATRONI_KUBERNETES_POD_IP
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: status.podIP
+            - name: PATRONI_KUBERNETES_NAMESPACE
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: metadata.namespace
+            - name: PATRONI_KUBERNETES_LABELS
+              value: '{application: patroni}'
+            - name: PATRONI_CITUS_DATABASE
+              value: citus
+            - name: PATRONI_CITUS_GROUP
+              value: "0"
+
+2. for the worker cluster from the group 2
+
+.. code:: YAML
+
+        apiVersion: v1
+        kind: Pod
+        metadata:
+          labels:
+            application: patroni
+            citus-group: "2"
+            citus-type: worker
+            cluster-name: citusdemo
+          name: citusdemo-2-0
+          namespace: default
+        spec:
+          containers:
+          - env:
+            - name: PATRONI_SCOPE
+              value: citusdemo
+            - name: PATRONI_NAME
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: metadata.name
+            - name: PATRONI_KUBERNETES_POD_IP
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: status.podIP
+            - name: PATRONI_KUBERNETES_NAMESPACE
+              valueFrom:
+                fieldRef:
+                  apiVersion: v1
+                  fieldPath: metadata.namespace
+            - name: PATRONI_KUBERNETES_LABELS
+              value: '{application: patroni}'
+            - name: PATRONI_CITUS_DATABASE
+              value: citus
+            - name: PATRONI_CITUS_GROUP
+              value: "2"
+
+As you may noticed, both examples have ``citus-group`` label set. This label
+allows Patroni to identify object as belonging to a certain Citus group. In
+addition to that, there is also ``PATRONI_CITUS_GROUP`` environment variable,
+which has the same value as the ``citus-group`` label. When Patroni creates
+new Kubernetes objects ConfigMaps or Endpoints, it automatically puts the
+``citus-group: ${env.PATRONI_CITUS_GROUP}`` label on them:
+
+.. code:: YAML
+
+        apiVersion: v1
+        kind: ConfigMap
+        metadata:
+          name: citusdemo-0-leader  # Is generated as ${env.PATRONI_SCOPE}-${env.PATRONI_CITUS_GROUP}-leader
+          labels:
+            application: patroni    # Is set from the ${env.PATRONI_KUBERNETES_LABELS}
+            cluster-name: citusdemo # Is automatically set from the ${env.PATRONI_SCOPE}
+            citus-group: '0'        # Is automatically set from the ${env.PATRONI_CITUS_GROUP}
+
+You can find a complete example of Patroni deployment on Kubernetes with Citus
+support in the `kubernetes`__ folder of the Patroni repository.
+
+__ https://github.com/zalando/patroni/tree/master/kubernetes
+
+There are two important files for you:
+
+1. Dockerfile.citus
+2. citus_k8s.yaml
+
+Citus upgrades and PostgreSQL major upgrades
+--------------------------------------------
+
+First, please read about upgrading Citus version in the `documentation`__.
+There is one minor change in the process. When executing upgrade, you have to
+use ``patronictl restart`` instead of ``systemctl restart`` to restart
+PostgreSQL.
+
+__ https://docs.citusdata.com/en/latest/admin_guide/upgrading_citus.html
+
+The PostgreSQL major upgrade with Citus is a bit more complex. You will have to
+combine techniques used in the Citus documentation about major upgrades and
+Patroni documentation about :ref:`PostgreSQL major upgrade<major_upgrade>`.
+Please keep in mind that Citus cluster consists of many Patroni clusters
+(coordinator and workers) and they all have to be upgraded independently.

docs/conf.py

@@ -18,9 +18,11 @@
 # documentation root, use os.path.abspath to make it absolute, like shown here.
 #
 import os
-# import sys
-# sys.path.insert(0, os.path.abspath('.'))
 
+import sys
+sys.path.insert(0, os.path.abspath('..'))
+
+from patroni.version import __version__
 
 # -- General configuration ------------------------------------------------
 
@@ -51,7 +53,7 @@ master_doc = 'index'
 
 # General information about the project.
 project = 'Patroni'
-copyright = '2016, Zalando SE'
+copyright = '2015 Compose, Zalando SE'
 author = 'Zalando SE'
 
 # The version info for the project you're documenting, acts as replacement for
@@ -59,9 +61,9 @@ author = 'Zalando SE'
 # built documents.
 #
 # The short X.Y version.
-version = '1.2'
+version = __version__[:__version__.rfind('.')]
 # The full version, including alpha/beta/rc tags.
-release = '1.2.2'
+release = __version__
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.
@@ -192,4 +194,7 @@ intersphinx_mapping = {'https://docs.python.org/': None}
 # A possibility to have an own stylesheet, to add new rules or override existing ones
 # For the latter case, the CSS specificity of the rules should be higher than the default ones
 def setup(app):
-   app.add_stylesheet("custom.css")
+    if hasattr(app, 'add_css_file'):
+        app.add_css_file('custom.css')
+    else:
+        app.add_stylesheet('custom.css')

docs/dcs_failsafe_mode.rst

@@ -0,0 +1,63 @@
+.. _dcs_failsafe_mode:
+
+DCS Failsafe Mode
+=================
+
+The problem
+-----------
+
+Patroni is heavily relying on Distributed Configuration Store (DCS) to solve the task of leader elections and detect network partitioning. That is, the node is allowed to run Postgres as the primary only if it can update the leader lock in DCS. In case the update of the leader lock fails, Postgres is immediately demoted and started as read-only. Depending on which DCS is used, the chances of hitting the "problem" differ. For example, with Etcd which is only used for Patroni, chances are close to zero, while with K8s API (backed by Etcd) it could be observed more frequently.
+
+
+Reasons for the current implementation
+---------------------------------------
+
+The leader lock update failure could be caused by two main reasons:
+
+1. Network partitioning
+2. DCS being down
+
+In general, it is impossible to distinguish between these two from a single node, and therefore Patroni assumes the worst case - network partitioning. In the case of a partitioned network, other nodes of the Patroni cluster may successfully grab the leader lock and promote Postgres to primary. In order to avoid a split-brain, the old primary is demoted before the leader lock expires.
+
+
+DCS Failsafe Mode
+-----------------
+
+We introduce a new special option, the ``failsafe_mode``. It could be enabled only via global configuration stored in the DCS ``/config`` key. If the failsafe mode is enabled and the leader lock update in DCS failed due to reasons different from the version/value/index mismatch, Postgres may continue to run as a primary if it can access all known members of the cluster via Patroni REST API.
+
+
+Low-level implementation details
+--------------------------------
+
+- We introduce a new, permanent key in DCS, named ``/failsafe``.
+- The ``/failsafe`` key contains all known members of the given Patroni cluster at a given time.
+- The current leader maintains the ``/failsafe`` key.
+- The member is allowed to participate in the leader race and become the new leader only if it is present in the ``/failsafe`` key.
+- If the cluster consists of a single node the ``/failsafe`` key will contain a single member.
+- In the case of DCS "outage" the existing primary connects to all members presented in the ``/failsafe`` key via the ``POST /failsafe`` REST API and may continue to run as the primary if all replicas acknowledge it.
+- If one of the members doesn't respond, the primary is demoted.
+- Replicas are using incoming ``POST /failsafe`` REST API requests as an indicator that the primary is still alive. This information is cached for ``ttl`` seconds.
+
+
+F.A.Q.
+------
+
+- Why MUST the current primary see ALL other members? Can’t we rely on quorum here?
+
+  This is a great question! The problem is that the view on the quorum might be different from the perspective of DCS and Patroni. While DCS nodes must be evenly distributed across availability zones, there is no such rule for Patroni, and more importantly, there is no mechanism for introducing and enforcing such a rule. If the majority of Patroni nodes ends up in the losing part of the partitioned network (including primary) while minority nodes are in the winning part, the primary must be demoted. Only checking ALL other members allows detecting such a situation.
+
+- What if node/pod gets terminated while DCS is down?
+
+  If DCS isn’t accessible, the check “are ALL other cluster members accessible?” is executed every cycle of the heartbeat loop (every ``loop_wait`` seconds). If pod/node is terminated, the check will fail and Postgres will be demoted to a read-only and will not recover until DCS is restored.
+
+- What if all members of the Patroni cluster are lost while DCS is down?
+
+  Patroni could be configured to create the new replica from the backup even when the cluster doesn't have a leader. But, if the new member isn't present in the ``/failsafe`` key, it will not be able to grab the leader lock and promote.
+  
+- What will happen if the primary lost access to DCS while replicas didn't?
+
+  The primary will execute the failsafe code and contact all known replicas. These replicas will use this information as an indicator that the primary is alive and will not start the leader race even if the leader lock in DCS has expired.
+
+- How to enable the Failsafe Mode?
+
+  Before enabling the ``failsafe_mode`` please make sure that Patroni version on all members is up-to-date. After that, you can use either the ``PATCH /config`` :ref:`REST API <rest_api>` or ``patronictl edit-config -s failsafe_mode=true``

docs/dynamic_configuration.rst

@@ -14,13 +14,15 @@ Patroni configuration is stored in the DCS (Distributed Configuration Store). Th
 
 - Local :ref:`configuration <settings>` (patroni.yml).
 	These options are defined in the configuration file and take precedence over dynamic configuration.
-	patroni.yml could be changed and reload in runtime (without restart of Patroni) by sending SIGHUP to the Patroni process or by performing ``POST /reload`` REST-API request.
+	patroni.yml could be changed and reloaded in runtime (without restart of Patroni) by sending SIGHUP to the Patroni process, performing ``POST /reload`` REST-API request or executing ``patronictl reload``.
 
-- Environment :ref:`configuration <environment>` .
+- Environment :ref:`configuration <environment>`.
 	It is possible to set/override some of the "Local" configuration parameters with environment variables.
-	Environment configuration is very useful when you are running in a dynamic environment and you don't know some of the parameters in advance (for example it's not possible to know you external IP address when you are running inside ``docker``).
+	Environment configuration is very useful when you are running in a dynamic environment and you don't know some of the parameters in advance (for example it's not possible to know your external IP address when you are running inside ``docker``).
 
-Some of the PostgreSQL parameters must hold the same values on the master and the replicas. For those, values set either in the local patroni configuration files or via the environment variables take no effect. To alter or set their values one must change the shared configuration in the DCS. Below is the actual list of such parameters together with the default values:
+The local configuration can be either a single YAML file or a directory. When it is a directory, all YAML files in that directory are loaded one by one in sorted order. In case a key is defined in multiple files, the occurrence in the last file takes precedence.
+
+Some of the PostgreSQL parameters must hold the same values on the primary and the replicas. For those, values set either in the local patroni configuration files or via the environment variables take no effect. To alter or set their values one must change the shared configuration in the DCS. Below is the actual list of such parameters together with the default values:
 
 - max_connections: 100
 - max_locks_per_transaction: 64
@@ -30,11 +32,12 @@ Some of the PostgreSQL parameters must hold the same values on the master and th
 - wal_log_hints: on
 - track_commit_timestamp: off
 
-For the parameters below, PostgreSQL does not require equal values among the master and all the replicas. However, considering the possibility of a replica to become the master at any time, it doesn't really make sense to set them differently; therefore, Patroni restricts setting their values to the Dynamic configuration
+For the parameters below, PostgreSQL does not require equal values among the primary and all the replicas. However, considering the possibility of a replica to become the primary at any time, it doesn't really make sense to set them differently; therefore, Patroni restricts setting their values to the Dynamic configuration
 
 - max_wal_senders: 5
 - max_replication_slots: 5
 - wal_keep_segments: 8
+- wal_keep_size: 128MB
 
 These parameters are validated to ensure they are sane, or meet a minimum value.
 
@@ -42,7 +45,7 @@ There are some other Postgres parameters controlled by Patroni:
 
 - listen_addresses - is set either from ``postgresql.listen`` or from ``PATRONI_POSTGRESQL_LISTEN`` environment variable
 - port - is set either from ``postgresql.listen`` or from ``PATRONI_POSTGRESQL_LISTEN`` environment variable
-- cluster_name - is set either from ``scope`` or from ``PATRRONI_SCOPE`` environment variable
+- cluster_name - is set either from ``scope`` or from ``PATRONI_SCOPE`` environment variable
 - hot_standby: on
 
 To be on the safe side parameters from the above lists are not written into ``postgresql.conf``, but passed as a list of arguments to the ``pg_ctl start`` which gives them the highest precedence, even above `ALTER SYSTEM <https://www.postgresql.org/docs/current/static/sql-altersystem.html>`__
@@ -76,153 +79,11 @@ Also, the following Patroni configuration options can be changed only dynamicall
 - loop_wait: 10
 - retry_timeouts: 10
 - maximum_lag_on_failover: 1048576
+- max_timelines_history: 0
+- check_timeline: false
 - postgresql.use_slots: true
 
 Upon changing these options, Patroni will read the relevant section of the configuration stored in DCS and change its
 run-time values.
 
-Patroni nodes are dumping the state of the DCS options to disk upon for every change of the configuration into the file ``patroni.dynamic.json`` located in the Postgres data directory. Only the master is allowed to restore these options from the on-disk dump if these are completely absent from the DCS or if they are invalid.
-
-REST API
-========
-
-We provide a REST API endpoint for working with dynamic configuration.
-
-GET /config
------------
-Get current version of dynamic configuration.
-
-.. code-block:: bash
-
-	$ curl -s localhost:8008/config | jq .
-	{
-	  "ttl": 30,
-	  "loop_wait": 10,
-	  "retry_timeout": 10,
-	  "maximum_lag_on_failover": 1048576,
-	  "postgresql": {
-	    "use_slots": true,
-	    "use_pg_rewind": true,
-	    "parameters": {
-	      "hot_standby": "on",
-	      "wal_log_hints": "on",
-	      "wal_keep_segments": 8,
-	      "wal_level": "hot_standby",
-	      "max_wal_senders": 5,
-	      "max_replication_slots": 5,
-	      "max_connections": "100"
-	    }
-	  }
-	}
-
-PATCH /config
--------------
-Change existing configuration.
-
-.. code-block:: bash
-
-	$ curl -s -XPATCH -d \
-		'{"loop_wait":5,"ttl":20,"postgresql":{"parameters":{"max_connections":"101"}}}' \
-		http://localhost:8008/config | jq .
-	{
-	  "ttl": 20,
-	  "loop_wait": 5,
-	  "maximum_lag_on_failover": 1048576,
-	  "retry_timeout": 10,
-	  "postgresql": {
-	    "use_slots": true,
-	    "use_pg_rewind": true,
-	    "parameters": {
-	      "hot_standby": "on",
-	      "wal_log_hints": "on",
-	      "wal_keep_segments": 8,
-	      "wal_level": "hot_standby",
-	      "max_wal_senders": 5,
-	      "max_replication_slots": 5,
-	      "max_connections": "101"
-	    }
-	  }
-	}
-
-The above REST API call patches the existing configuration and returns the new configuration.
-
-Let's check that the node processed this configuration. First of all it should start printing log lines every 5 seconds (loop_wait=5). The change of "max_connections" requires a restart, so the "restart_pending" flag should be exposed:
-
-.. code-block:: bash
-
-	$ curl -s http://localhost:8008/patroni | jq .
-	{
-	  "pending_restart": true,
-	  "database_system_identifier": "6287881213849985952",
-	  "postmaster_start_time": "2016-06-13 13:13:05.211 CEST",
-	  "xlog": {
-	    "location": 2197818976
-	  },
-	  "patroni": {
-	    "scope": "batman",
-	    "version": "1.0"
-	  },
-	  "state": "running",
-	  "role": "master",
-	  "server_version": 90503
-	}
-
-Removing parameters:
-
-If you want to remove (reset) some setting just patch it with ``null``:
-
-.. code-block:: bash
-
-	$ curl -s -XPATCH -d \
-		'{"postgresql":{"parameters":{"max_connections":null}}}' \
-		http://localhost:8008/config | jq .
-	{
-	  "ttl": 20,
-	  "loop_wait": 5,
-	  "retry_timeout": 10,
-	  "maximum_lag_on_failover": 1048576,
-	  "postgresql": {
-	    "use_slots": true,
-	    "use_pg_rewind": true,
-	    "parameters": {
-	      "hot_standby": "on",
-	      "unix_socket_directories": ".",
-	      "wal_keep_segments": 8,
-	      "wal_level": "hot_standby",
-	      "wal_log_hints": "on",
-	      "max_wal_senders": 5,
-	      "max_replication_slots": 5
-	    }
-	  }
-	}
-
-Above call removes ``postgresql.parameters.max_connections`` from the dynamic configuration.
-
-PUT /config
------------
-
-It's also possible to perform the full rewrite of an existing dynamic configuration unconditionally:
-
-.. code-block:: bash
-
-	$ curl -s -XPUT -d \
-		'{"maximum_lag_on_failover":1048576,"retry_timeout":10,"postgresql":{"use_slots":true,"use_pg_rewind":true,"parameters":{"hot_standby":"on","wal_log_hints":"on","wal_keep_segments":8,"wal_level":"hot_standby","unix_socket_directories":".","max_wal_senders":5}},"loop_wait":3,"ttl":20}' \
-		http://localhost:8008/config | jq .
-	{
-	  "ttl": 20,
-	  "maximum_lag_on_failover": 1048576,
-	  "retry_timeout": 10,
-	  "postgresql": {
-	    "use_slots": true,
-	    "parameters": {
-	      "hot_standby": "on",
-	      "unix_socket_directories": ".",
-	      "wal_keep_segments": 8,
-	      "wal_level": "hot_standby",
-	      "wal_log_hints": "on",
-	      "max_wal_senders": 5
-	    },
-	    "use_pg_rewind": true
-	  },
-	  "loop_wait": 3
-	}
+Patroni nodes are dumping the state of the DCS options to disk upon for every change of the configuration into the file ``patroni.dynamic.json`` located in the Postgres data directory. Only the leader is allowed to restore these options from the on-disk dump if these are completely absent from the DCS or if they are invalid.

docs/existing_data.rst

@@ -0,0 +1,53 @@
+.. _existing_data:
+
+Convert a Standalone to a Patroni Cluster
+=========================================
+
+This section describes the process for converting a standalone PostgreSQL instance into a Patroni cluster.
+
+To deploy a Patroni cluster without using a pre-existing PostgreSQL instance, see :ref:`Running and Configuring <running_configuring>` instead.
+
+Procedure
+---------
+
+A Patroni cluster can be started with a data directory from a single-node PostgreSQL database. This is achieved by following closely these steps:
+
+1. Manually start PostgreSQL daemon
+2. Create Patroni superuser and replication users as defined in the :ref:`authentication <postgresql_settings>` section of the Patroni configuration. If this user is created in SQL, the following queries achieve this:
+
+.. code-block:: sql
+
+   CREATE USER $PATRONI_SUPERUSER_USERNAME WITH SUPERUSER ENCRYPTED PASSWORD '$PATRONI_SUPERUSER_PASSWORD';
+   CREATE USER $PATRONI_REPLICATION_USERNAME WITH REPLICATION ENCRYPTED PASSWORD '$PATRONI_REPLICATION_PASSWORD';
+
+3. Start Patroni (e.g. ``patroni /etc/patroni/patroni.yml``). It automatically detects that PostgreSQL daemon is already running but its configuration might be out-of-date.
+4. Ask Patroni to restart the node with ``patronictl restart cluster-name node-name``. This step is only required if PostgreSQL configuration is out-of-date.
+
+.. _major_upgrade:
+
+Major Upgrade of PostgreSQL Version
+===================================
+
+The only possible way to do a major upgrade currently is:
+
+1. Stop Patroni
+2. Upgrade PostgreSQL binaries and perform `pg_upgrade <https://www.postgresql.org/docs/current/pgupgrade.html>`_ on the primary node
+3. Update patroni.yml
+4. Remove the initialize key from DCS or wipe complete cluster state from DCS. The second one could be achieved by running ``patronictl remove <cluster-name>``. It is necessary because pg_upgrade runs initdb which actually creates a new database with a new PostgreSQL system identifier.
+5. If you wiped the cluster state in the previous step, you may wish to copy patroni.dynamic.json from old data dir to the new one.  It will help you to retain some PostgreSQL parameters you had set before.
+6. Start Patroni on the primary node.
+7. Upgrade PostgreSQL binaries, update patroni.yml and wipe the data_dir on standby nodes.
+8. Start Patroni on the standby nodes and wait for the replication to complete.
+
+Running pg_upgrade on standby nodes is not supported by PostgreSQL. If you know what you are doing, you can try the rsync procedure described in https://www.postgresql.org/docs/current/pgupgrade.html instead of wiping data_dir on standby nodes. The safest way is however to let Patroni replicate the data for you.
+
+FAQ
+---
+
+- During Patroni startup, Patroni complains that it cannot bind to the PostgreSQL port.
+
+  You need to verify ``listen_addresses`` and ``port`` in ``postgresql.conf`` and ``postgresql.listen`` in ``patroni.yml``. Don't forget that ``pg_hba.conf`` should allow such access.
+
+- After asking Patroni to restart the node, PostgreSQL displays the error message ``could not open configuration file "/etc/postgresql/10/main/pg_hba.conf": No such file or directory``
+
+  It can mean various things depending on how you manage PostgreSQL configuration. If you specified `postgresql.config_dir`, Patroni generates the ``pg_hba.conf`` based on the settings in the :ref:`bootstrap <bootstrap_settings>` section only when it bootstraps a new cluster. In this scenario the ``PGDATA`` was not empty, therefore no bootstrap happened. This file must exist beforehand.

docs/index.rst

@@ -6,11 +6,15 @@
 Introduction
 ============
 
-Patroni is a template for you to create your own customized, high-availability solution using Python and - for maximum accessibility - a distributed configuration store like `ZooKeeper <https://zookeeper.apache.org/>`__, `etcd <https://github.com/coreos/etcd>`__ or `Consul <https://github.com/hashicorp/consul>`__. Database engineers, DBAs, DevOps engineers, and SREs who are looking to quickly deploy HA PostgreSQL in the datacenter-or anywhere else-will hopefully find it useful.
+Patroni is a template for you to create your own customized, high-availability solution using Python and - for maximum accessibility - a distributed configuration store like `ZooKeeper <https://zookeeper.apache.org/>`__, `etcd <https://github.com/coreos/etcd>`__, `Consul <https://github.com/hashicorp/consul>`__ or `Kubernetes <https://kubernetes.io>`__. Database engineers, DBAs, DevOps engineers, and SREs who are looking to quickly deploy HA PostgreSQL in the datacenter-or anywhere else-will hopefully find it useful.
 
 We call Patroni a "template" because it is far from being a one-size-fits-all or plug-and-play replication system. It will have its own caveats. Use wisely. There are many ways to run high availability with PostgreSQL; for a list, see the `PostgreSQL Documentation <https://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling>`__.
 
-**Note to Kubernetes users**: We're currently developing Patroni to be as useful as possible for teams running Kubernetes on top of Google Compute Engine; Patroni can be the HA solution for Postgres in such an environment. To this end, we've created a `Helm Chart <https://github.com/kubernetes/charts/tree/master/incubator/patroni>`__ that enables you to deploy a five-node Patroni cluster using a Kubernetes PetSet.
+Currently supported PostgreSQL versions: 9.3 to 15.
+
+**Note to Citus users**: Starting from 3.0 Patroni nicely integrates with the `Citus <https://github.com/citusdata/citus>`__ database extension to Postgres. Please check the :ref:`Citus support page <citus>` in the Patroni documentation for more info about how to use Patroni high availability together with a Citus distributed cluster.
+
+**Note to Kubernetes users**: Patroni can run natively on top of Kubernetes. Take a look at the :ref:`Kubernetes <kubernetes>` chapter of the Patroni documentation.
 
 
 .. toctree::
@@ -18,11 +22,19 @@ We call Patroni a "template" because it is far from being a one-size-fits-all or
    :caption: Contents:
 
    README
+   citus
    dynamic_configuration
+   dcs_failsafe_mode
+   rest_api
+   existing_data
    ENVIRONMENT
    SETTINGS
+   security
+   replica_bootstrap
    replication_modes
    pause
+   kubernetes
+   watchdog
    releases
    CONTRIBUTING
 
@@ -32,5 +44,3 @@ Indices and tables
 * :ref:`genindex`
 * :ref:`modindex`
 * :ref:`search`
-
-

docs/kubernetes.rst

@@ -0,0 +1,50 @@
+.. _kubernetes:
+
+Using Patroni with Kubernetes
+=============================
+
+Patroni can use Kubernetes objects in order to store the state of the cluster and manage the leader key. That makes it
+capable of operating Postgres in Kubernetes environment without any consistency store, namely, one doesn't
+need to run an extra Etcd deployment. There are two different type of Kubernetes objects Patroni can use to store the
+leader and the configuration keys, they are configured with the `kubernetes.use_endpoints` or `PATRONI_KUBERNETES_USE_ENDPOINTS`
+environment variable.
+
+Use Endpoints
+-------------
+
+Despite the fact that this is the recommended mode, it is turned off by default for compatibility reasons. When it is on, Patroni stores
+the cluster configuration and the leader key in the `metadata: annotations` fields of the respective `Endpoints` it creates.
+Changing the leader is safer than when using `ConfigMaps`, since both the annotations, containing the leader information, and the actual addresses
+pointing to the running leader pod are updated simultaneously in one go.
+
+Use ConfigMaps
+--------------
+
+In this mode, Patroni will create ConfigMaps instead of Endpoints and store keys inside meta-data of those ConfigMaps.
+Changing the leader takes at least two updates, one to the leader ConfigMap and another to the respective Endpoint.
+
+To direct the traffic to the Postgres leader you need to configure the Kubernetes Postgres service to use the label selector with the `role_label` (configured in patroni configuration).
+
+Note that in some cases, for instance, when running on OpenShift, there is no alternative to using ConfigMaps.
+
+Configuration
+-------------
+
+Patroni Kubernetes :ref:`settings <kubernetes_settings>` and :ref:`environment variables <kubernetes_environment>` are described in the general chapters of the documentation.
+
+Examples
+--------
+
+- The `kubernetes <https://github.com/zalando/patroni/tree/master/kubernetes>`__ folder of the Patroni repository contains
+  examples of the Docker image, and the Kubernetes manifest to test Patroni Kubernetes setup.
+  Note that in the current state it will not be able to use PersistentVolumes because of permission issues.
+
+- You can find the full-featured Docker image that can use Persistent Volumes in the
+  `Spilo Project <https://github.com/zalando/spilo>`_.
+
+- There is also a `Helm chart <https://github.com/kubernetes/charts/tree/master/incubator/patroni>`_
+  to deploy the Spilo image configured with Patroni running using Kubernetes.
+
+- In order to run your database clusters at scale using Patroni and Spilo, take a look at the
+  `postgres-operator <https://github.com/zalando-incubator/postgres-operator>`_ project. It implements the operator pattern
+  to manage Spilo clusters.

docs/pause.rst

@@ -6,7 +6,7 @@ Pause/Resume mode for the cluster
 The goal
 --------
 
-Under certain circumstances Patroni needs to temporary step down from managing the cluster, while still retaining the cluster state in DCS. Possible use cases are uncommon activities on the cluster, such as major version upgrades or corruption recovery. During those activities nodes are often started and stopped for the reason unknown to Patroni, some nodes can be even temporary promoted, violating the assumption of running only one master. Therefore, Patroni needs to be able to "detach" from the running cluster, implementing an equivalent of the maintenance mode in Pacemaker.
+Under certain circumstances Patroni needs to temporarily step down from managing the cluster, while still retaining the cluster state in DCS. Possible use cases are uncommon activities on the cluster, such as major version upgrades or corruption recovery. During those activities nodes are often started and stopped for reasons unknown to Patroni, some nodes can be even temporarily promoted, violating the assumption of running only one primary. Therefore, Patroni needs to be able to "detach" from the running cluster, implementing an equivalent of the maintenance mode in Pacemaker.
 
 
 
@@ -17,15 +17,17 @@ When Patroni runs in a paused mode, it does not change the state of PostgreSQL,
 
 - For each node, the member key in DCS is updated with the current information about the cluster. This causes Patroni to run read-only queries on a member node if the member is running.
 
-- For the Postgres master with the leader lock Patroni updates the lock. If the node with the leader lock stops being the master (i.e. is demoted manually), Patroni will release the lock instead of promoting the node back.
+- For the Postgres primary with the leader lock Patroni updates the lock. If the node with the leader lock stops being the primary (i.e. is demoted manually), Patroni will release the lock instead of promoting the node back.
 
-- Manual unscheduled restart, reinitialize and manual failover are allowed. Manual failover is only allowed if the node to failover to is specified. In the paused mode, manual failover does not require a running master node.
+- Manual unscheduled restart, reinitialize and manual failover are allowed. Manual failover is only allowed if the node to failover to is specified. In the paused mode, manual failover does not require a running primary node.
 
-- If 'parallel' masters are detected by Patroni, it emits a warning, but does not demote the masters without the leader lock.
+- If 'parallel' primaries are detected by Patroni, it emits a warning, but does not demote the primary without the leader lock.
 
-- If there is no leader lock in the cluster, the running master acquires the lock. If there is more than one master node, then the first master to acquire the lock wins. If there are no masters altogether, Patroni does not try to promote any replicas. There is an exception in this rule: if there is no leader lock because the old master has demoted itself due to the manual promotion, then only the candidate node mentioned in the promotion request may take the leader lock. When the new leader lock is granted (i.e. after promoting a replica manually), Patroni makes sure the replicas that were streaming from the previous leader will switch to the new one.
+- If there is no leader lock in the cluster, the running primary acquires the lock. If there is more than one primary node, then the first primary to acquire the lock wins. If there are no primary altogether, Patroni does not try to promote any replicas. There is an exception in this rule: if there is no leader lock because the old primary has demoted itself due to the manual promotion, then only the candidate node mentioned in the promotion request may take the leader lock. When the new leader lock is granted (i.e. after promoting a replica manually), Patroni makes sure the replicas that were streaming from the previous leader will switch to the new one.
 
-- When Postgres is stopped, Patroni does not try to start it. When Patroni is stopped, it does not to stop Postgres instance it is managing.
+- When Postgres is stopped, Patroni does not try to start it. When Patroni is stopped, it does not try to stop the Postgres instance it is managing.
+
+- Patroni will not try to remove replication slots that don't represent the other cluster member or are not listed in the configuration of the permanent slots.
 
 User guide
 ----------

docs/replica_bootstrap.rst

@@ -0,0 +1,227 @@
+Replica imaging and bootstrap
+=============================
+
+Patroni allows customizing creation of a new replica. It also supports defining what happens when the new empty cluster
+is being bootstrapped. The distinction between two is well defined: Patroni creates replicas only if the ``initialize``
+key is present in DCS for the cluster. If there is no ``initialize`` key - Patroni calls bootstrap exclusively on the
+first node that takes the initialize key lock.
+
+.. _custom_bootstrap:
+
+Bootstrap
+---------
+
+PostgreSQL provides ``initdb`` command to initialize a new cluster and Patroni calls it by default. In certain cases,
+particularly when creating a new cluster as a copy of an existing one, it is necessary to replace a built-in method with
+custom actions. Patroni supports executing user-defined scripts to bootstrap new clusters, supplying some required
+arguments to them, i.e. the name of the cluster and the path to the data directory. This is configured in the
+``bootstrap`` section of the Patroni configuration. For example:
+
+.. code:: YAML
+
+    bootstrap:
+        method: <custom_bootstrap_method_name>
+        <custom_bootstrap_method_name>:
+            command: <path_to_custom_bootstrap_script> [param1 [, ...]]
+            keep_existing_recovery_conf: False
+            no_params: False
+            recovery_conf:
+                recovery_target_action: promote
+                recovery_target_timeline: latest
+                restore_command: <method_specific_restore_command>
+
+
+Each bootstrap method must define at least a ``name`` and a ``command``. A special ``initdb`` method is available to trigger
+the default behavior, in which case ``method`` parameter can be omitted altogether. The ``command`` can be specified using either
+an absolute path, or the one relative to the ``patroni`` command location. In addition to the fixed parameters defined
+in the configuration files, Patroni supplies two cluster-specific ones:
+
+--scope
+    Name of the cluster to be bootstrapped
+--datadir
+    Path to the data directory of the cluster instance to be bootstrapped
+
+Passing these two additional flags can be disabled by setting a special ``no_params`` parameter to ``True``.
+
+If the bootstrap script returns 0, Patroni tries to configure and start the PostgreSQL instance produced by it. If any
+of the intermediate steps fail, or the script returns a non-zero value, Patroni assumes that the bootstrap has failed,
+cleans up after itself and releases the initialize lock to give another node the opportunity to bootstrap.
+
+If a ``recovery_conf`` block is defined in the same section as the custom bootstrap method, Patroni will generate a
+``recovery.conf`` before starting the newly bootstrapped instance. Typically, such recovery.conf should contain at least
+one of the ``recovery_target_*`` parameters, together with the ``recovery_target_timeline`` set to ``promote``.
+
+If ``keep_existing_recovery_conf`` is defined and set to ``True``, Patroni will not remove the existing ``recovery.conf`` file if it exists.
+This is useful when bootstrapping from a backup with tools like pgBackRest that generate the appropriate ``recovery.conf`` for you.
+
+ .. note:: Bootstrap methods are neither chained, nor fallen-back to the default one in case the primary one fails
+
+
+.. _custom_replica_creation:
+
+Building replicas
+-----------------
+
+Patroni uses tried and proven ``pg_basebackup`` in order to create new replicas. One downside of it is that it requires
+a running leader node. Another one is the lack of 'on-the-fly' compression for the backup data and no built-in cleanup
+for outdated backup files. Some people prefer other backup solutions, such as ``WAL-E``, ``pgBackRest``, ``Barman`` and
+others, or simply roll their own scripts. In order to accommodate all those use-cases Patroni supports running custom
+scripts to clone a new replica. Those are configured in the ``postgresql`` configuration block:
+
+.. code:: YAML
+
+    postgresql:
+        create_replica_methods:
+            - <method name>
+        <method name>:
+            command: <command name>
+            keep_data: True
+            no_params: True
+            no_leader: 1
+
+example: wal_e
+
+.. code:: YAML
+
+    postgresql:
+        create_replica_methods:
+            - wal_e
+            - basebackup
+        wal_e:
+            command: patroni_wale_restore
+            no_leader: 1
+            envdir: {{WALE_ENV_DIR}}
+            use_iam: 1
+        basebackup:
+            max-rate: '100M'
+
+example: pgbackrest
+
+.. code:: YAML
+
+    postgresql:
+        create_replica_methods:
+            - pgbackrest
+            - basebackup
+        pgbackrest:
+            command: /usr/bin/pgbackrest --stanza=<scope> --delta restore
+            keep_data: True
+            no_params: True
+        basebackup:
+            max-rate: '100M'
+
+
+The ``create_replica_methods`` defines available replica creation methods and the order of executing them. Patroni will
+stop on the first one that returns 0. Each method should define a separate section in the configuration file, listing the command
+to execute and any custom parameters that should be passed to that command. All parameters will be passed in a
+``--name=value`` format. Besides user-defined parameters, Patroni supplies a couple of cluster-specific ones:
+
+--scope
+    Which cluster this replica belongs to
+--datadir
+    Path to the data directory of the replica
+--role
+    Always 'replica'
+--connstring
+    Connection string to connect to the cluster member to clone from (primary or other replica). The user in the
+    connection string can execute SQL and replication protocol commands.
+
+A special ``no_leader`` parameter, if defined, allows Patroni to call the replica creation method even if there is no
+running leader or replicas. In that case, an empty string will be passed in a connection string. This is useful for
+restoring the formerly running cluster from the binary backup.
+
+A special ``keep_data`` parameter, if defined, will instruct Patroni to  not clean PGDATA folder before calling restore.
+
+A special ``no_params`` parameter, if defined, restricts passing parameters to custom command.
+
+A ``basebackup`` method is a special case: it will be used if
+``create_replica_methods`` is empty, although it is possible
+to list it explicitly among the ``create_replica_methods`` methods. This method initializes a new replica with the
+``pg_basebackup``, the base backup is taken from the leader unless there are replicas with ``clonefrom`` tag, in which case one
+of such replicas will be used as the origin for pg_basebackup. It works without any configuration; however, it is
+possible to specify a ``basebackup`` configuration section. Same rules as with the other method configuration apply,
+namely, only long (with --) options should be specified there. Not all parameters make sense, if you override a connection
+string or provide an option to created tar-ed or compressed base backups, patroni won't be able to make a replica out
+of it. There is no validation performed on the names or values of the parameters passed to the ``basebackup`` section.
+Also note that in case symlinks are used for the WAL folder it is up to the user to specify the correct ``--waldir``
+path as an option, so that after replica buildup or re-initialization the symlink would persist. This option is supported
+only since v10 though.
+
+You can specify basebackup parameters as either a map (key-value pairs) or a list of elements, where each element
+could be either a key-value pair or a single key (for options that does not receive any values, for instance, ``--verbose``).
+Consider those 2 examples:
+
+.. code:: YAML
+
+    postgresql:
+        basebackup:
+            max-rate: '100M'
+            checkpoint: 'fast'
+
+and
+
+.. code:: YAML
+
+    postgresql:
+        basebackup:
+            - verbose
+            - max-rate: '100M'
+            - waldir: /pg-wal-mount/external-waldir
+
+If all replica creation methods fail, Patroni will try again all methods in order during the next event loop cycle.
+
+.. _standby_cluster:
+
+Standby cluster
+---------------
+
+Another available option is to run a "standby cluster", that contains only of
+standby nodes replicating from some remote node. This type of clusters has:
+
+* "standby leader", that behaves pretty much like a regular cluster leader,
+  except it replicates from a remote node.
+
+* cascade replicas, that are replicating from standby leader.
+
+Standby leader holds and updates a leader lock in DCS. If the leader lock
+expires, cascade replicas will perform an election to choose another leader
+from the standbys.
+
+There is no further relationship between the standby cluster and the primary
+cluster it replicates from, in particular, they must not share the same DCS
+scope if they use the same DCS. They do not know anything else from each other
+apart from replication information. Also, the standby cluster is not being
+displayed in ``patronictl list`` or ``patronictl topology`` output on the
+primary cluster.
+
+For the sake of flexibility, you can specify methods of creating a replica and
+recovery WAL records when a cluster is in the "standby mode" by providing
+`create_replica_methods` key in `standby_cluster` section. It is distinct from
+creating replicas, when cluster is detached and functions as a normal cluster,
+which is controlled by `create_replica_methods` in `postgresql` section. Both
+"standby" and "normal" `create_replica_methods` reference  keys in `postgresql`
+section.
+
+To configure such cluster you need to specify the section ``standby_cluster``
+in a patroni configuration:
+
+.. code:: YAML
+
+    bootstrap:
+        dcs:
+            standby_cluster:
+                host: 1.2.3.4
+                port: 5432
+                primary_slot_name: patroni
+                create_replica_methods:
+                - basebackup
+
+Note, that these options will be applied only once during cluster bootstrap,
+and the only way to change them afterwards is through DCS.
+
+Patroni expects to find `postgresql.conf` or `postgresql.conf.backup` in PGDATA
+of the remote primary and will not start if it does not find it after a
+basebackup. If the remote primary keeps its `postgresql.conf` elsewhere, it is
+your responsibility to copy it to PGDATA.
+
+If you use replication slots on the standby cluster, you must also create the corresponding replication slot on the primary cluster.  It will not be done automatically by the standby cluster implementation.  You can use Patroni's permanent replication slots feature on the primary cluster to maintain a replication slot with the same name as ``primary_slot_name``, or its default value if ``primary_slot_name`` is not provided.

docs/replication_modes.rst

@@ -9,9 +9,11 @@ Patroni uses PostgreSQL streaming replication. For more information about stream
 Asynchronous mode durability
 ----------------------------
 
-In asynchronous mode the cluster is allowed to lose some committed transactions to ensure availability. When master server fails or becomes unavailable for any other reason Patroni will automatically promote a sufficiently healthy standby to master. Any transactions that have not been replicated to that standby remain in a "forked timeline" on the master, and are effectively unrecoverable [1]_.
+In asynchronous mode the cluster is allowed to lose some committed transactions to ensure availability. When the primary server fails or becomes unavailable for any other reason Patroni will automatically promote a sufficiently healthy standby to primary. Any transactions that have not been replicated to that standby remain in a "forked timeline" on the primary, and are effectively unrecoverable [1]_.
 
-The amount of transactions that can be lost is controlled via ``maximum_lag_on_failover`` parameter. Because master transaction log position is not sampled in real time, in reality the amount of lost data on failover is worst case bounded by  ``maximum_lag_on_failover`` bytes of transaction log plus the amount that is written in the last ``ttl`` seconds (``loop_wait``/2 seconds in the average case). However typical steady state replication delay is well under a second.
+The amount of transactions that can be lost is controlled via ``maximum_lag_on_failover`` parameter. Because the primary transaction log position is not sampled in real time, in reality the amount of lost data on failover is worst case bounded by  ``maximum_lag_on_failover`` bytes of transaction log plus the amount that is written in the last ``ttl`` seconds (``loop_wait``/2 seconds in the average case). However typical steady state replication delay is well under a second.
+
+By default, when running leader elections, Patroni does not take into account the current timeline of replicas, what in some cases could be undesirable behavior. You can prevent the node not having the same timeline as a former primary become the new leader by changing the value of ``check_timeline`` parameter to ``true``.
 
 PostgreSQL synchronous replication
 ----------------------------------
@@ -20,7 +22,7 @@ You can use Postgres's `synchronous replication <http://www.postgresql.org/docs/
 
 In hosted datacenter environments (like AWS, Rackspace, or any network you do not control), synchronous replication significantly increases the variability of write performance. If followers become inaccessible from the leader, the leader effectively becomes read-only.
 
-To enable a simple synchronous replication test, add the follow lines to the ``parameters`` section of your YAML configuration files:
+To enable a simple synchronous replication test, add the following lines to the ``parameters`` section of your YAML configuration files:
 
 .. code:: YAML
 
@@ -29,28 +31,42 @@ To enable a simple synchronous replication test, add the follow lines to the ``p
 
 When using PostgreSQL synchronous replication, use at least three Postgres data nodes to ensure write availability if one host fails.
 
-Using PostgreSQL synchronous replication does not guarantee zero lost transactions under all circumstances. When master and standby that is currently acting as synchronous fail simultaneously a third node that might not contain all transactions will be promoted.
+Using PostgreSQL synchronous replication does not guarantee zero lost transactions under all circumstances. When the primary and the secondary that is currently acting as a synchronous replica fail simultaneously a third node that might not contain all transactions will be promoted.
 
 .. _synchronous_mode:
 
 Synchronous mode
 ----------------
 
-For use cases where losing committed transactions is not permissible you can turn on Patronis ``synchronous_mode``. When ``synchronous_mode`` is turned on Patroni will not promote a standby unless it is certain that the standby contains all transactions that may have returned a successful commit status to client [2]_. This means that the system may be unavailable for writes even though some servers are available. System administrators can still use manual failover commmands to promote a standby even if it results in transaction loss.
+For use cases where losing committed transactions is not permissible you can turn on Patroni's ``synchronous_mode``. When ``synchronous_mode`` is turned on Patroni will not promote a standby unless it is certain that the standby contains all transactions that may have returned a successful commit status to client [2]_. This means that the system may be unavailable for writes even though some servers are available. System administrators can still use manual failover commands to promote a standby even if it results in transaction loss.
 
-Turning on ``synchronous_mode`` does not guarantee multi node durability of commits under all circumstances. When no suitable standby is available, master server will still accept writes, but does not guarantee their replication. When the master fails in this mode no standby will be promote.  When the host that used to be master comes back it will get promoted automatically, unless system administrator performed a manual failover. This behavior makes synchronous mode usable with 2 node clusters.
+Turning on ``synchronous_mode`` does not guarantee multi node durability of commits under all circumstances. When no suitable standby is available, primary server will still accept writes, but does not guarantee their replication. When the primary fails in this mode no standby will be promoted. When the host that used to be the primary comes back it will get promoted automatically, unless system administrator performed a manual failover. This behavior makes synchronous mode usable with 2 node clusters.
 
-When ``synchronous_mode`` is on and a standby crashes, commits will block until next iteration of Patroni runs and switches master to standalone mode (worst case delay for writes ``ttl`` seconds, average case ``loop_wait``/2 seconds). Manually shutting down or restarting a standby will not cause a commit service interruption. Standby will signal the master to release itself from synchronous standby duties before PostgreSQL shutdown is initiated.
+When ``synchronous_mode`` is on and a standby crashes, commits will block until next iteration of Patroni runs and switches the primary to standalone mode (worst case delay for writes ``ttl`` seconds, average case ``loop_wait``/2 seconds). Manually shutting down or restarting a standby will not cause a commit service interruption. Standby will signal the primary to release itself from synchronous standby duties before PostgreSQL shutdown is initiated.
+
+When it is absolutely necessary to guarantee that each write is stored durably
+on at least two nodes, enable ``synchronous_mode_strict`` in addition to the
+``synchronous_mode``. This parameter prevents Patroni from switching off the
+synchronous replication on the primary when no synchronous standby candidates
+are available. As a downside, the primary is not be available for writes
+(unless the Postgres transaction explicitly turns of ``synchronous_mode``),
+blocking all client write requests until at least one synchronous replica comes
+up.
 
 You can ensure that a standby never becomes the synchronous standby by setting ``nosync`` tag to true. This is recommended to set for standbys that are behind slow network connections and would cause performance degradation when becoming a synchronous standby.
 
-Synchronous mode can be switched on and off via Patroni REST interface. See `dynamic configuration <https://github.com/zalando/patroni/blob/master/docs/dynamic_configuration.rst>`__ for instructions.
+Synchronous mode can be switched on and off via Patroni REST interface. See :ref:`dynamic configuration <dynamic_configuration>` for instructions.
 
+Note: Because of the way synchronous replication is implemented in PostgreSQL it is still possible to lose transactions even when using ``synchronous_mode_strict``. If the PostgreSQL backend is cancelled while waiting to acknowledge replication (as a result of packet cancellation due to client timeout or backend failure) transaction changes become visible for other backends. Such changes are not yet replicated and may be lost in case of standby promotion.
+
+Synchronous Replication Factor
+------------------------------
+The parameter ``synchronous_node_count`` is used by Patroni to manage number of synchronous standby databases. It is set to 1 by default. It has no effect when ``synchronous_mode`` is set to off. When enabled, Patroni manages precise number of synchronous standby databases based on parameter ``synchronous_node_count`` and adjusts the state in DCS & synchronous_standby_names as members join and leave.
 
 Synchronous mode implementation
 -------------------------------
 
-When in synchronous mode Patroni maintains synchronization state in the DCS, containing the latest master and current synchronous standby. This state is updated with strict ordering constraints to ensure the following invariants:
+When in synchronous mode Patroni maintains synchronization state in the DCS, containing the latest primary and current synchronous standby databases. This state is updated with strict ordering constraints to ensure the following invariants:
 
 - A node must be marked as the latest leader whenever it can accept write transactions. Patroni crashing or PostgreSQL not shutting down can cause violations of this invariant.
 
@@ -58,11 +74,11 @@ When in synchronous mode Patroni maintains synchronization state in the DCS, con
 
 - A node that is not the leader or current synchronous standby is not allowed to promote itself automatically.
 
-Patroni will only ever assign one standby to ``synchronous_standby_names`` because with multiple candidates it is not possible to know which node was acting as synchronous during the failure.
+Patroni will only assign one or more synchronous standby nodes based on ``synchronous_node_count`` parameter to ``synchronous_standby_names``.
 
-On each HA loop iteration Patroni re-evaluates synchronous standby choice. If the current synchronous standby is connected and has not requested its synchronous status to be removed it remains picked. Otherwise the cluster member avaiable for sync that is furthest ahead in replication is picked.
+On each HA loop iteration Patroni re-evaluates synchronous standby nodes choice. If the current list of synchronous standby nodes are connected and has not requested its synchronous status to be removed it remains picked. Otherwise the cluster member available for sync that is furthest ahead in replication is picked.
 
 
-.. [1] The data is still there, but recovering it requires a manual recovery effort by data recovery specialists. When Patroni is allowed to rewind with ``use_pg_rewind`` the forked timeline will be automatically erased to rejoin the failed master with the cluster.
+.. [1] The data is still there, but recovering it requires a manual recovery effort by data recovery specialists. When Patroni is allowed to rewind with ``use_pg_rewind`` the forked timeline will be automatically erased to rejoin the failed primary with the cluster.
 
-.. [2] Clients can change the behavior per transaction using PostgreSQL's ``synchronous_commit`` setting. Transactions with ``synchronous_commit`` values of ``off`` and ``local`` may be lost on fail over, but will not be blocked by replication delays.
+.. [2] Clients can change the behavior per transaction using PostgreSQL's ``synchronous_commit`` setting. Transactions with ``synchronous_commit`` values of ``off`` and ``local`` may be lost on fail over, but will not be blocked by replication delays.

docs/rest_api.rst

@@ -0,0 +1,392 @@
+.. _rest_api:
+
+Patroni REST API
+================
+
+Patroni has a rich REST API, which is used by Patroni itself during the leader race, by the ``patronictl`` tool in order to perform failovers/switchovers/reinitialize/restarts/reloads, by HAProxy or any other kind of load balancer to perform HTTP health checks, and of course could also be used for monitoring. Below you will find the list of Patroni REST API endpoints.
+
+Health check endpoints
+----------------------
+For all health check ``GET`` requests Patroni returns a JSON document with the status of the node, along with the HTTP status code. If you don't want or don't need the JSON document, you might consider using the ``HEAD`` or ``OPTIONS`` method instead of ``GET``.
+
+- The following requests to Patroni REST API will return HTTP status code **200** only when the Patroni node is running as the primary with leader lock:
+
+  - ``GET /``
+  - ``GET /primary``
+  - ``GET /read-write``
+
+- ``GET /standby-leader``: returns HTTP status code **200** only when the Patroni node is running as the leader in a :ref:`standby cluster <standby_cluster>`.
+
+- ``GET /leader``: returns HTTP status code **200** when the Patroni node has the leader lock. The major difference from the two previous endpoints is that it doesn't take into account whether PostgreSQL is running as the ``primary`` or the ``standby_leader``.
+
+- ``GET /replica``: replica health check endpoint. It returns HTTP status code **200** only when the Patroni node is in the state ``running``, the role is ``replica`` and ``noloadbalance`` tag is not set.
+
+- ``GET /replica?lag=<max-lag>``: replica check endpoint. In addition to checks from ``replica``, it also checks replication latency and returns status code **200** only when it is below specified value. The key cluster.last_leader_operation from DCS is used for Leader wal position and compute latency on replica for performance reasons. max-lag can be specified in bytes (integer) or in human readable values, for e.g. 16kB, 64MB, 1GB.
+
+  - ``GET /replica?lag=1048576``
+  - ``GET /replica?lag=1024kB``
+  - ``GET /replica?lag=10MB``
+  - ``GET /replica?lag=1GB``
+
+- ``GET /replica?tag_key1=value1&tag_key2=value2``: replica check endpoint. In addition, It will also check for user defined tags ``key1`` and ``key2`` and their respective values in the **tags** section of the yaml configuration management. If the tag isn't defined for an instance, or if the value in the yaml configuration doesn't match the querying value, it will return HTTP Status Code 503.
+
+ In the following requests, since we are checking for the leader or standby-leader status, Patroni doesn't apply any of the user defined tags and they will be ignored.
+  - ``GET /?tag_key1=value1&tag_key2=value2``
+  - ``GET /leader?tag_key1=value1&tag_key2=value2``
+  - ``GET /primary?tag_key1=value1&tag_key2=value2``
+  - ``GET /read-write?tag_key1=value1&tag_key2=value2``
+  - ``GET /standby_leader?tag_key1=value1&tag_key2=value2``
+  - ``GET /standby-leader?tag_key1=value1&tag_key2=value2``
+
+- ``GET /read-only``: like the above endpoint, but also includes the primary.
+
+- ``GET /synchronous`` or ``GET /sync``: returns HTTP status code **200** only when the Patroni node is running as a synchronous standby.
+
+- ``GET /read-only-sync``: like the above endpoint, but also includes the primary.
+
+- ``GET /asynchronous`` or ``GET /async``: returns HTTP status code **200** only when the Patroni node is running as an asynchronous standby.
+
+
+- ``GET /asynchronous?lag=<max-lag>`` or ``GET /async?lag=<max-lag>``: asynchronous standby check endpoint. In addition to checks from ``asynchronous`` or ``async``, it also checks replication latency and returns status code **200** only when it is below specified value. The key cluster.last_leader_operation from DCS is used for Leader wal position and compute latency on replica for performance reasons. max-lag can be specified in bytes (integer) or in human readable values, for e.g. 16kB, 64MB, 1GB.
+
+  - ``GET /async?lag=1048576``
+  - ``GET /async?lag=1024kB``
+  - ``GET /async?lag=10MB``
+  - ``GET /async?lag=1GB``
+
+- ``GET /health``: returns HTTP status code **200** only when PostgreSQL is up and running.
+
+- ``GET /liveness``: returns HTTP status code **200** if Patroni heartbeat loop is properly running and **503** if the last run was more than ``ttl`` seconds ago on the primary or ``2*ttl`` on the replica. Could be used for ``livenessProbe``.
+
+- ``GET /readiness``: returns HTTP status code **200** when the Patroni node is running as the leader or when PostgreSQL is up and running. The endpoint could be used for ``readinessProbe`` when it is not possible to use Kubernetes endpoints for leader elections (OpenShift).
+
+Both, ``readiness`` and ``liveness`` endpoints are very light-weight and not executing any SQL. Probes should be configured in such a way that they start failing about time when the leader key is expiring. With the default value of ``ttl``, which is ``30s`` example probes would look like:
+
+.. code-block:: yaml
+
+    readinessProbe:
+      httpGet:
+        scheme: HTTP
+        path: /readiness
+        port: 8008
+      initialDelaySeconds: 3
+      periodSeconds: 10
+      timeoutSeconds: 5
+      successThreshold: 1
+      failureThreshold: 3
+    livenessProbe:
+      httpGet:
+        scheme: HTTP
+        path: /liveness
+        port: 8008
+      initialDelaySeconds: 3
+      periodSeconds: 10
+      timeoutSeconds: 5
+      successThreshold: 1
+      failureThreshold: 3
+
+
+Monitoring endpoint
+-------------------
+
+The ``GET /patroni`` is used by Patroni during the leader race. It also could be used by your monitoring system. The JSON document produced by this endpoint has the same structure as the JSON produced by the health check endpoints.
+
+.. code-block:: bash
+
+    $ curl -s http://localhost:8008/patroni | jq .
+    {
+      "state": "running",
+      "postmaster_start_time": "2019-09-24 09:22:32.555 CEST",
+      "role": "master",
+      "server_version": 110005,
+      "cluster_unlocked": false,
+      "xlog": {
+        "location": 25624640
+      },
+      "timeline": 3,
+      "database_system_identifier": "6739877027151648096",
+      "patroni": {
+        "version": "1.6.0",
+        "scope": "batman"
+      }
+    }
+
+Cluster status endpoints
+------------------------
+
+- The ``GET /cluster`` endpoint generates a JSON document describing the current cluster topology and state:
+
+.. code-block:: bash
+
+    $ curl -s http://localhost:8008/cluster | jq .
+    {
+      "members": [
+        {
+          "name": "postgresql0",
+          "host": "127.0.0.1",
+          "port": 5432,
+          "role": "leader",
+          "state": "running",
+          "api_url": "http://127.0.0.1:8008/patroni",
+          "timeline": 5,
+          "tags": {
+            "clonefrom": true
+          }
+        },
+        {
+          "name": "postgresql1",
+          "host": "127.0.0.1",
+          "port": 5433,
+          "role": "replica",
+          "state": "running",
+          "api_url": "http://127.0.0.1:8009/patroni",
+          "timeline": 5,
+          "tags": {
+            "clonefrom": true
+          },
+          "lag": 0
+        }
+      ],
+      "scheduled_switchover": {
+        "at": "2019-09-24T10:36:00+02:00",
+        "from": "postgresql0"
+      }
+    }
+
+
+- The ``GET /history`` endpoint provides a view on the history of cluster switchovers/failovers. The format is very similar to the content of history files in the ``pg_wal`` directory. The only difference is the timestamp field showing when the new timeline was created.
+
+.. code-block:: bash
+
+    $ curl -s http://localhost:8008/history | jq .
+    [
+      [
+        1,
+        25623960,
+        "no recovery target specified",
+        "2019-09-23T16:57:57+02:00"
+      ],
+      [
+        2,
+        25624344,
+        "no recovery target specified",
+        "2019-09-24T09:22:33+02:00"
+      ],
+      [
+        3,
+        25624752,
+        "no recovery target specified",
+        "2019-09-24T09:26:15+02:00"
+      ],
+      [
+        4,
+        50331856,
+        "no recovery target specified",
+        "2019-09-24T09:35:52+02:00"
+      ]
+    ]
+
+
+Config endpoint
+---------------
+
+``GET /config``: Get the current version of the dynamic configuration:
+
+.. code-block:: bash
+
+	$ curl -s localhost:8008/config | jq .
+	{
+	  "ttl": 30,
+	  "loop_wait": 10,
+	  "retry_timeout": 10,
+	  "maximum_lag_on_failover": 1048576,
+	  "postgresql": {
+	    "use_slots": true,
+	    "use_pg_rewind": true,
+	    "parameters": {
+	      "hot_standby": "on",
+	      "wal_log_hints": "on",
+	      "wal_level": "hot_standby",
+	      "max_wal_senders": 5,
+	      "max_replication_slots": 5,
+	      "max_connections": "100"
+	    }
+	  }
+	}
+
+
+``PATCH /config``: Change the existing configuration.
+
+.. code-block:: bash
+
+	$ curl -s -XPATCH -d \
+		'{"loop_wait":5,"ttl":20,"postgresql":{"parameters":{"max_connections":"101"}}}' \
+		http://localhost:8008/config | jq .
+	{
+	  "ttl": 20,
+	  "loop_wait": 5,
+	  "maximum_lag_on_failover": 1048576,
+	  "retry_timeout": 10,
+	  "postgresql": {
+	    "use_slots": true,
+	    "use_pg_rewind": true,
+	    "parameters": {
+	      "hot_standby": "on",
+	      "wal_log_hints": "on",
+	      "wal_level": "hot_standby",
+	      "max_wal_senders": 5,
+	      "max_replication_slots": 5,
+	      "max_connections": "101"
+	    }
+	  }
+	}
+
+The above REST API call patches the existing configuration and returns the new configuration.
+
+Let's check that the node processed this configuration. First of all it should start printing log lines every 5 seconds (loop_wait=5). The change of "max_connections" requires a restart, so the "pending_restart" flag should be exposed:
+
+.. code-block:: bash
+
+	$ curl -s http://localhost:8008/patroni | jq .
+	{
+	  "pending_restart": true,
+	  "database_system_identifier": "6287881213849985952",
+	  "postmaster_start_time": "2016-06-13 13:13:05.211 CEST",
+	  "xlog": {
+	    "location": 2197818976
+	  },
+	  "patroni": {
+	    "scope": "batman",
+	    "version": "1.0"
+	  },
+	  "state": "running",
+	  "role": "master",
+	  "server_version": 90503
+	}
+
+Removing parameters:
+
+If you want to remove (reset) some setting just patch it with ``null``:
+
+.. code-block:: bash
+
+	$ curl -s -XPATCH -d \
+		'{"postgresql":{"parameters":{"max_connections":null}}}' \
+		http://localhost:8008/config | jq .
+	{
+	  "ttl": 20,
+	  "loop_wait": 5,
+	  "retry_timeout": 10,
+	  "maximum_lag_on_failover": 1048576,
+	  "postgresql": {
+	    "use_slots": true,
+	    "use_pg_rewind": true,
+	    "parameters": {
+	      "hot_standby": "on",
+	      "unix_socket_directories": ".",
+	      "wal_level": "hot_standby",
+	      "wal_log_hints": "on",
+	      "max_wal_senders": 5,
+	      "max_replication_slots": 5
+	    }
+	  }
+	}
+
+The above call removes ``postgresql.parameters.max_connections`` from the dynamic configuration.
+
+``PUT /config``: It's also possible to perform the full rewrite of an existing dynamic configuration unconditionally:
+
+.. code-block:: bash
+
+	$ curl -s -XPUT -d \
+		'{"maximum_lag_on_failover":1048576,"retry_timeout":10,"postgresql":{"use_slots":true,"use_pg_rewind":true,"parameters":{"hot_standby":"on","wal_log_hints":"on","wal_level":"hot_standby","unix_socket_directories":".","max_wal_senders":5}},"loop_wait":3,"ttl":20}' \
+		http://localhost:8008/config | jq .
+	{
+	  "ttl": 20,
+	  "maximum_lag_on_failover": 1048576,
+	  "retry_timeout": 10,
+	  "postgresql": {
+	    "use_slots": true,
+	    "parameters": {
+	      "hot_standby": "on",
+	      "unix_socket_directories": ".",
+	      "wal_level": "hot_standby",
+	      "wal_log_hints": "on",
+	      "max_wal_senders": 5
+	    },
+	    "use_pg_rewind": true
+	  },
+	  "loop_wait": 3
+	}
+
+
+Switchover and failover endpoints
+---------------------------------
+
+``POST /switchover`` or ``POST /failover``. These endpoints are very similar to each other. There are a couple of minor differences though:
+
+1. The failover endpoint allows to perform a manual failover when there are no healthy nodes, but at the same time it will not allow you to schedule a switchover.
+
+2. The switchover endpoint is the opposite. It works only when the cluster is healthy (there is a leader) and allows to schedule a switchover at a given time.
+
+
+In the JSON body of the ``POST`` request you must specify at least the ``leader`` or ``candidate`` fields and optionally the ``scheduled_at`` field if you want to schedule a switchover at a specific time.
+
+
+Example: perform a failover to the specific node:
+
+.. code-block:: bash
+
+    $ curl -s http://localhost:8009/failover -XPOST -d '{"candidate":"postgresql1"}'
+    Successfully failed over to "postgresql1"
+
+
+Example: schedule a switchover from the leader to any other healthy replica in the cluster at a specific time:
+
+.. code-block:: bash
+
+    $ curl -s http://localhost:8008/switchover -XPOST -d \
+	    '{"leader":"postgresql0","scheduled_at":"2019-09-24T12:00+00"}'
+    Switchover scheduled
+
+
+Depending on the situation the request might finish with a different HTTP status code and body. The status code **200** is returned when the switchover or failover successfully completed. If the switchover was successfully scheduled, Patroni will return HTTP status code **202**. In case something went wrong, the error status code (one of **400**, **412** or **503**) will be returned with some details in the response body. For more information please check the source code of ``patroni/api.py:do_POST_failover()`` method.
+
+- ``DELETE /switchover``: delete the scheduled switchover
+
+The ``POST /switchover`` and ``POST failover`` endpoints are used by ``patronictl switchover`` and ``patronictl failover``, respectively.
+The ``DELETE /switchover`` is used by ``patronictl flush <cluster-name> switchover``.
+
+
+Restart endpoint
+----------------
+
+- ``POST /restart``: You can restart Postgres on the specific node by performing the ``POST /restart`` call. In the JSON body of ``POST`` request it is possible to optionally specify some restart conditions:
+
+  - **restart_pending**: boolean, if set to ``true`` Patroni will restart PostgreSQL only when restart is pending in order to apply some changes in the PostgreSQL config.
+  - **role**: perform restart only if the current role of the node matches with the role from the POST request.
+  - **postgres_version**: perform restart only if the current version of postgres is smaller than specified in the POST request.
+  - **timeout**: how long we should wait before PostgreSQL starts accepting connections. Overrides ``primary_start_timeout``.
+  - **schedule**: timestamp with time zone, schedule the restart somewhere in the future.
+
+- ``DELETE /restart``: delete the scheduled restart
+
+``POST /restart`` and ``DELETE /restart`` endpoints are used by ``patronictl restart`` and ``patronictl flush <cluster-name> restart`` respectively.
+
+
+Reload endpoint
+---------------
+
+The ``POST /reload`` call will order Patroni to re-read and apply the configuration file. This is the equivalent of sending the ``SIGHUP`` signal to the Patroni process. In case you changed some of the Postgres parameters which require a restart (like **shared_buffers**), you still have to explicitly do the restart of Postgres by either calling the ``POST /restart`` endpoint or with the help of ``patronictl restart``.
+
+The reload endpoint is used by ``patronictl reload``.
+
+
+Reinitialize endpoint
+---------------------
+
+``POST /reinitialize``: reinitialize the PostgreSQL data directory on the specified node. It is allowed to be executed only on replicas. Once called, it will remove the data directory and start ``pg_basebackup`` or some alternative :ref:`replica creation method <custom_replica_creation>`.
+
+The call might fail if Patroni is in a loop trying to recover (restart) a failed Postgres. In order to overcome this problem one can specify ``{"force":true}`` in the request body.
+
+The reinitialize endpoint is used by ``patronictl reinit``.

docs/security.rst

@@ -0,0 +1,37 @@
+.. _security:
+
+=======================
+Security Considerations
+=======================
+
+A Patroni cluster has two interfaces to be protected from unauthorized access: the distributed configuration storage (DCS) and the Patroni REST API.
+
+Protecting DCS
+==============
+
+Patroni and patronictl both store and retrieve data to/from the DCS.
+
+Despite DCS doesn't contain any sensitive information, it allows changing some of Patroni/Postgres configuration. Therefore the very first thing that should be protected is DCS itself.
+
+The details of protection depend on the type of DCS used. The authentication and encryption parameters (tokens/basic-auth/client certificates) for the supported types of DCS are covered in :ref:`SETTINGS <bootstrap_settings>`
+
+The general recommendation is to enable TLS for all DCS communication.
+
+Protecting the REST API
+=======================
+
+Protecting the REST API is a more complicated task.
+
+The Patroni REST API is used by Patroni itself during the leader race, by the ``patronictl`` tool in order to perform failovers/switchovers/reinitialize/restarts/reloads, by HAProxy or any other kind of load balancer to perform HTTP health checks, and of course could also be used for monitoring.
+
+From the point of view of security, REST API contains safe (``GET`` requests, only retrieve information) and unsafe (``PUT``, ``POST``, ``PATCH`` and ``DELETE`` requests, change the state of nodes) endpoints.
+
+The unsafe endpoints can be protected with HTTP basic-auth by setting the ``restapi.authentication.username`` and ``restapi.authentication.password`` parameters. There is no way to protect the safe endpoints without enabling TLS.
+
+When TLS for the REST API is enabled and a PKI is established, mutual authentication of the API server and API client is possible for all endpoints.
+
+The ``restapi`` section parameters enable TLS client authentication to the server. Depending on the value of the ``verify_client`` parameter, the API server requires a successful client certificate verification for both safe and unsafe API calls (``verify_client: required``), or only for unsafe API calls (``verify_client: optional``), or for no API calls (``verify_client: none``).
+
+The ``ctl`` section parameters enable TLS server authentication to the client (the ``patronictl`` tool which uses the same config as patroni). Set ``insecure: true`` to disable the server certificate verification by the client. See :ref:`SETTINGS <patronictl_settings>` for a detailed description of the TLS client parameters.
+
+Protecting the PostgreSQL database proper from unauthorized access is beyond the scope of this document and is covered in https://www.postgresql.org/docs/current/client-authentication.html

docs/watchdog.rst

@@ -0,0 +1,39 @@
+.. _watchdog:
+
+Watchdog support
+================
+
+Having multiple PostgreSQL servers running as primary can result in transactions lost due to diverging timelines. This situation is also called a split-brain problem. To avoid split-brain Patroni needs to ensure PostgreSQL will not accept any transaction commits after leader key expires in the DCS. Under normal circumstances Patroni will try to achieve this by stopping PostgreSQL when leader lock update fails for any reason. However, this may fail to happen due to various reasons:
+
+- Patroni has crashed due to a bug, out-of-memory condition or by being accidentally killed by a system administrator.
+
+- Shutting down PostgreSQL is too slow.
+
+- Patroni does not get to run due to high load on the system, the VM being paused by the hypervisor, or other infrastructure issues.
+
+To guarantee correct behavior under these conditions Patroni supports watchdog devices. Watchdog devices are software or hardware mechanisms that will reset the whole system when they do not get a keepalive heartbeat within a specified timeframe. This adds an additional layer of fail safe in case usual Patroni split-brain protection mechanisms fail.
+
+Patroni will try to activate the watchdog before promoting PostgreSQL to primary. If watchdog activation fails and watchdog mode is ``required`` then the node will refuse to become leader. When deciding to participate in leader election Patroni will also check that watchdog configuration will allow it to become leader at all. After demoting PostgreSQL (for example due to a manual failover) Patroni will disable the watchdog again. Watchdog will also be disabled while Patroni is in paused state.
+
+By default Patroni will set up the watchdog to expire 5 seconds before TTL expires. With the default setup of ``loop_wait=10`` and ``ttl=30`` this gives HA loop at least 15 seconds (``ttl`` - ``safety_margin`` - ``loop_wait``) to complete before the system gets forcefully reset. By default accessing DCS is configured to time out after 10 seconds. This means that when DCS is unavailable, for example due to network issues, Patroni and PostgreSQL will have at least 5 seconds (``ttl`` - ``safety_margin`` - ``loop_wait`` - ``retry_timeout``) to come to a state where all client connections are terminated.
+
+Safety margin is the amount of time that Patroni reserves for time between leader key update and watchdog keepalive. Patroni will try to send a keepalive immediately after confirmation of leader key update. If Patroni process is suspended for extended amount of time at exactly the right moment the keepalive may be delayed for more than the safety margin without triggering the watchdog. This results in a window of time where watchdog will not trigger before leader key expiration, invalidating the guarantee. To be absolutely sure that watchdog will trigger under all circumstances set up the watchdog to expire after half of TTL by setting ``safety_margin`` to -1 to set watchdog timeout to ``ttl // 2``. If you need this guarantee you probably should increase ``ttl`` and/or reduce ``loop_wait`` and ``retry_timeout``.
+
+Currently watchdogs are only supported using Linux watchdog device interface.
+
+Setting up software watchdog on Linux
+-------------------------------------
+
+Default Patroni configuration will try to use ``/dev/watchdog`` on Linux if it is accessible to Patroni. For most use cases using software watchdog built into the Linux kernel is secure enough.
+
+To enable software watchdog issue the following commands as root before starting Patroni:
+
+.. code-block:: bash
+
+    modprobe softdog
+    # Replace postgres with the user you will be running patroni under
+    chown postgres /dev/watchdog
+
+For testing it may be helpful to disable rebooting by adding ``soft_noboot=1`` to the modprobe command line. In this case the watchdog will just log a line in kernel ring buffer, visible via `dmesg`.
+
+Patroni will log information about the watchdog when it is successfully enabled.

extras/README.md

@@ -1,11 +1,11 @@
 ### confd
 
-`confd` directory contains haproxy template files for the [confd](https://github.com/kelseyhightower/confd) -- lightweight configuration management tool
+`confd` directory contains haproxy and pgbouncer template files for the [confd](https://github.com/kelseyhightower/confd) -- lightweight configuration management tool
 You need to copy content of `confd` directory into /etcd/confd and run confd service:
 ```bash
-$ confd -prefix=/service/$PATRONI_SCOPE -backend etcd -node $PATRONI_ETCD_HOST -interval=10
+$ confd -prefix=/service/$PATRONI_SCOPE -backend etcd -node $PATRONI_ETCD_URL -interval=10
 ```
-It will periodically update haproxy.cfg with the actual list of Patroni nodes from `etcd` and "reload" haproxy when it is necessary.
+It will periodically update haproxy.cfg and pgbouncer.ini with the actual list of Patroni nodes from `etcd` and "reload" haproxy and pgbouncer.ini when it is necessary.
 
 
 ### startup-scripts

extras/confd/conf.d/haproxy.toml

@@ -9,5 +9,5 @@ check_cmd = "/usr/sbin/haproxy -c -f {{ .src }}"
 reload_cmd = "haproxy -f /etc/haproxy/haproxy.cfg -p /var/run/haproxy.pid -D -sf $(cat /var/run/haproxy.pid)"
 
 keys = [
-    "/members/",
+    "/",
 ]

extras/confd/conf.d/pgbouncer.toml

@@ -0,0 +1,12 @@
+[template]
+prefix = "/service/batman"
+owner = "postgres"
+mode = "0644"
+src = "pgbouncer.tmpl"
+dest = "/etc/pgbouncer/pgbouncer.ini"
+
+reload_cmd = "systemctl reload pgbouncer"
+
+keys = [
+    "/members/","/leader"
+]

extras/confd/templates/haproxy-citus.tmpl

@@ -0,0 +1,32 @@
+global
+	maxconn 100
+
+defaults
+	log	global
+	mode	tcp
+	retries 2
+	timeout client 30m
+	timeout connect 4s
+	timeout server 30m
+	timeout check 5s
+
+listen stats
+	mode http
+	bind *:7000
+	stats enable
+	stats uri /
+
+listen coordinator
+	bind *:5000
+	option httpchk HEAD /primary
+	http-check expect status 200
+	default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
+{{range gets "/0/members/*"}}	server {{base .Key}} {{$data := json .Value}}{{base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}} maxconn 100 check check-ssl port {{index (split (index (split $data.api_url "/") 2) ":") 1}} verify required ca-file /etc/ssl/certs/ssl-cert-snakeoil.pem crt /etc/ssl/private/ssl-cert-snakeoil.crt
+{{end}}
+listen workers
+	bind *:5001
+	option httpchk HEAD /primary
+	http-check expect status 200
+	default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
+{{range gets "/*/members/*"}}{{$group := index (split .Key "/") 1}}{{if ne $group "0"}}	server {{base .Key}} {{$data := json .Value}}{{base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}} maxconn 100 check check-ssl port {{index (split (index (split $data.api_url "/") 2) ":") 1}} verify required ca-file /etc/ssl/certs/ssl-cert-snakeoil.pem crt /etc/ssl/private/ssl-cert-snakeoil.crt
+{{end}}{{end}}

extras/confd/templates/haproxy.tmpl

@@ -10,19 +10,23 @@ defaults
 	timeout server 30m
 	timeout check 5s
 
-frontend master_postgresql
+listen stats
+	mode http
+	bind *:7000
+	stats enable
+	stats uri /
+
+listen primary
 	bind *:5000
-	default_backend backend_master
-
-frontend replicas_postgresql
+	option httpchk HEAD /primary
+	http-check expect status 200
+	default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
+{{range gets "/members/*"}}	server {{base .Key}} {{$data := json .Value}}{{base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}} maxconn 100 check port {{index (split (index (split $data.api_url "/") 2) ":") 1}}
+{{end}}
+listen replicas
 	bind *:5001
-	default_backend backend_replicas
-
-backend backend_master
-	option httpchk OPTIONS /master
-{{range gets "/members/*"}}	server {{base .Key}} {{$data := json .Value}}{{base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}} maxconn 100 check port {{index (split (index (split $data.api_url "/") 2) ":") 1}}
-{{end}}
-backend backend_replicas
-	option httpchk OPTIONS /replica
+	option httpchk HEAD /replica
+	http-check expect status 200
+	default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
 {{range gets "/members/*"}}	server {{base .Key}} {{$data := json .Value}}{{base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}} maxconn 100 check port {{index (split (index (split $data.api_url "/") 2) ":") 1}}
 {{end}}

extras/confd/templates/pgbouncer.tmpl

@@ -0,0 +1,17 @@
+[databases]
+{{with get "/leader"}}{{$leader := .Value}}{{$leadkey := printf "/members/%s" $leader}}{{with get $leadkey}}{{$data := json .Value}}{{$hostport := base (replace (index (split $data.conn_url "/") 2) "@" "/" -1)}}{{ $host := base (index (split $hostport ":") 0)}}{{ $port := base (index (split $hostport ":") 1)}}* = host={{ $host }} port={{ $port }} pool_size=10{{end}}{{end}}
+
+[pgbouncer]
+logfile = /var/log/postgresql/pgbouncer.log
+pidfile = /var/run/postgresql/pgbouncer.pid
+listen_addr = *
+listen_port = 6432
+unix_socket_dir = /var/run/postgresql
+auth_type = trust
+auth_file = /etc/pgbouncer/userlist.txt
+auth_hba_file = /etc/pgbouncer/pg_hba.txt
+admin_users = pgbouncer
+stats_users = pgbouncer
+pool_mode = session
+max_client_conn = 100
+default_pool_size = 20

extras/startup-scripts/README.md

@@ -1,6 +1,6 @@
 # startup scripts for Patroni
 
-This directory contains sample startup scripts for various OSes 
+This directory contains sample startup scripts for various OSes
 and management tools for Patroni.
 
 Scripts supplied:
@@ -10,4 +10,13 @@ Scripts supplied:
 Upstart job for Ubuntu 12.04 or 14.04.  Requires Upstart > 1.4. Intended for systems where Patroni has been installed on a base system, rather than in Docker.
 
 ### patroni.service
-Systemd service file, to be copied to /etc/systemd/system/patroni.service, tested on Centos 7.1 with Patroni installed from pip. 
+Systemd service file, to be copied to /etc/systemd/system/patroni.service, tested on Centos 7.1 with Patroni installed from pip.
+
+### patroni
+Init.d service file for Debian-like distributions. Copy it to /etc/init.d/, make executable:
+```chmod 755 /etc/init.d/patroni``` and run with ```service patroni start```, or make it starting on boot with ```update-rc.d patroni defaults```. Also you might edit some configuration variables in it:
+PATRONI for patroni.py location
+CONF for configuration file
+LOGFILE for log (script creates it if does not exist)
+
+Note. If you have several versions of Postgres installed, please add to POSTGRES_VERSION the release number which you wish to run. Script uses this value to append PATH environment with correct path to Postgres bin.

extras/startup-scripts/patroni

@@ -0,0 +1,144 @@
+#!/bin/sh
+#
+### BEGIN INIT INFO
+# Provides:          patroni
+# Required-Start:    $remote_fs $syslog
+# Required-Stop:     $remote_fs $syslog
+# Default-Start:     2 3 4 5
+# Default-Stop:      0 1 6
+# Short-Description: Patroni init script
+# Description:       Runners to orchestrate a high-availability PostgreSQL
+### END INIT INFO
+
+### BEGIN USER CONFIGURATION
+
+CONF="/etc/patroni/postgres.yml"
+LOGFILE="/var/log/patroni.log"
+USER="postgres"
+GROUP="postgres"
+
+NAME=patroni
+PATRONI="/opt/patroni/$NAME.py"
+PIDFILE="/var/run/$NAME.pid"
+
+# Set this parameter, if you have several Postgres versions installed
+# POSTGRES_VERSION="9.4"
+POSTGRES_VERSION=""
+
+### END USER CONFIGURATION
+
+. /lib/lsb/init-functions
+
+# Loading this library for get_versions() function
+if test ! -e /usr/share/postgresql-common/init.d-functions; then
+  log_failure_msg "Probably postgresql-common does not installed."
+  exit 1
+else
+  . /usr/share/postgresql-common/init.d-functions
+fi
+
+# Is there Patroni executable?
+if test ! -e $PATRONI; then
+  log_failure_msg "Patroni executable $PATRONI does not exist."
+  exit 1
+fi
+
+# Is there Patroni configuration file?
+if test ! -e $CONF; then
+  log_failure_msg "Patroni configuration file $CONF does not exist."
+  exit 1
+fi
+
+# Create logfile if doesn't exist
+if test ! -e $LOGFILE; then
+  log_action_msg "Creating logfile for Patroni..."
+  touch $LOGFILE
+  chown $USER:$GROUP $LOGFILE
+fi
+
+prepare_pgpath() {
+  if [ "$POSTGRES_VERSION" != "" ]; then
+    if [ -x /usr/lib/postgresql/$POSTGRES_VERSION/bin/pg_ctl ]; then
+      PGPATH="/usr/lib/postgresql/$POSTGRES_VERSION/bin"
+    else
+      log_failure_msg "Postgres version incorrect, check POSTGRES_VERSION variable."
+      exit 0
+    fi
+  else
+    get_versions
+    if echo $versions | grep -q -e "\s"; then
+      log_warning_msg "You have several Postgres versions installed. Please, use POSTGRES_VERSION to define correct environment."
+    else
+      versions=`echo $versions | sed -e 's/^[ \t]*//'`
+      PGPATH="/usr/lib/postgresql/$versions/bin"
+    fi
+  fi
+}
+
+get_pid() {
+  if test -e $PIDFILE; then
+    PID=`cat $PIDFILE`
+    CHILDPID=`ps --ppid $PID -o %p --no-headers`
+  else
+    log_failure_msg "Could not find PID file. Patroni probably down."
+    exit 1
+  fi
+}
+
+
+case "$1" in
+  start)
+    prepare_pgpath
+    PGPATH=$PATH:$PGPATH
+    log_success_msg "Starting Patroni\n"
+    exec start-stop-daemon --start --quiet \
+        --background \
+        --pidfile $PIDFILE --make-pidfile \
+        --chuid $USER:$GROUP \
+        --chdir `eval echo ~$USER` \
+        --exec $PATRONI \
+        --startas /bin/sh -- \
+        -c "/usr/bin/env PATH=$PGPATH /usr/bin/python $PATRONI $CONF >> $LOGFILE 2>&1"
+    ;;
+
+  stop)
+    log_success_msg "Stopping Patroni"
+    get_pid
+    start-stop-daemon --stop --pid $CHILDPID
+    start-stop-daemon --stop --pidfile $PIDFILE --remove-pidfile --quiet
+    ;;
+
+  reload)
+    log_success_msg "Reloading Patroni configuration"
+    get_pid
+    kill -HUP $CHILDPID
+    ;;
+
+  status)
+    get_pid
+    if start-stop-daemon -T --pid $CHILDPID; then
+      log_success_msg "Patroni is running\n"
+      exit 0
+    else
+      log_warning_msg "Patroni in not running\n"
+    fi
+    ;;
+
+  restart)
+    $0 stop
+    $0 start
+    ;;
+
+  *)
+    echo "Usage: /etc/init.d/$NAME {start|stop|restart|reload|status}"
+    exit 1
+    ;;
+esac
+
+if [ $? -eq 0 ]; then
+        echo .
+        exit 0
+else
+        echo " failed"
+        exit 1
+fi

extras/startup-scripts/patroni.service

@@ -11,20 +11,35 @@ Type=simple
 User=postgres
 Group=postgres
 
+# Read in configuration file if it exists, otherwise proceed
+EnvironmentFile=-/etc/patroni_env.conf
+
+# The default is the user's home directory, and if you want to change it, you must provide an absolute path.
+# WorkingDirectory=/home/sameuser
+
 # Where to send early-startup messages from the server
 # This is normally controlled by the global default set by systemd
-# StandardOutput=syslog
+#StandardOutput=syslog
 
+# Pre-commands to start watchdog device
+# Uncomment if watchdog is part of your patroni setup
+#ExecStartPre=-/usr/bin/sudo /sbin/modprobe softdog
+#ExecStartPre=-/usr/bin/sudo /bin/chown postgres /dev/watchdog
+
+# Start the patroni process
 ExecStart=/bin/patroni /etc/patroni.yml
 
-# only kill the patroni process, not it's children, so it will gracefully stop postgres
+# Send HUP to reload from patroni.yml
+ExecReload=/bin/kill -s HUP $MAINPID
+
+# Only kill the patroni process, not it's children, so it will gracefully stop postgres
 KillMode=process
 
 # Give a reasonable amount of time for the server to start up/shut down
-TimeoutSec=10
+TimeoutSec=30
 
-# Do not restart the service if it crashes, we want to manually inspect database on failure
-Restart=no
+# Restart the service if it crashed
+Restart=on-failure
 
 [Install]
 WantedBy=multi-user.target

features/archive-restore.py

@@ -0,0 +1,21 @@
+#!/usr/bin/env python
+import os
+import argparse
+import shutil
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--dirname", required=True)
+    parser.add_argument("--pathname", required=True)
+    parser.add_argument("--filename", required=True)
+    parser.add_argument("--mode", required=True, choices=("archive", "restore"))
+    args, _ = parser.parse_known_args()
+
+    full_filename = os.path.join(args.dirname, args.filename)
+    if args.mode == "archive":
+        if not os.path.isdir(args.dirname):
+            os.makedirs(args.dirname)
+        if not os.path.exists(full_filename):
+            shutil.copy(args.pathname, full_filename)
+    else:
+        shutil.copy(full_filename, args.pathname)

features/backup_create.py

@@ -0,0 +1,14 @@
+#!/usr/bin/env python
+import argparse
+import subprocess
+import sys
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--datadir", required=True)
+    parser.add_argument("--dbname", required=True)
+    parser.add_argument("--walmethod", required=True, choices=("fetch", "stream", "none"))
+    args, _ = parser.parse_known_args()
+
+    walmethod = ["-X", args.walmethod] if args.walmethod != "none" else []
+    sys.exit(subprocess.call(["pg_basebackup", "-D", args.datadir, "-c", "fast", "-d", args.dbname] + walmethod))

features/backup_restore.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+import argparse
+import shutil
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--datadir", required=True)
+    parser.add_argument("--sourcedir", required=True)
+    args, _ = parser.parse_known_args()
+
+    shutil.copytree(args.sourcedir, args.datadir)

features/basic_replication.feature

@@ -4,7 +4,8 @@ Feature: basic replication
   Scenario: check replication of a single table
     Given I start postgres0
     Then postgres0 is a leader after 10 seconds
-    When I issue a PATCH request to http://127.0.0.1:8008/config with {"ttl": 20, "loop_wait": 2, "synchronous_mode": true}
+    And there is a non empty initialize key in DCS after 15 seconds
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"ttl": 20, "synchronous_mode": true}
     Then I receive a response code 200
     When I start postgres1
     And I configure and start postgres2 with a tag replicatefrom postgres0
@@ -14,27 +15,71 @@ Feature: basic replication
     Then table foo is present on postgres2 after 20 seconds
 
   Scenario: check restart of sync replica
-    Given I run patronictl.py restart batman postgres2 --force
-    And "sync" key in DCS has sync_standby=postgres1 after 2 seconds
-    And I run patronictl.py restart batman postgres1 --force
-    Then I receive a response returncode 0
-    And "sync" key in DCS has sync_standby=postgres2 after 10 seconds
+    Given I shut down postgres2
+    Then "sync" key in DCS has sync_standby=postgres1 after 5 seconds
+    When I start postgres2
+    And I shut down postgres1
+    Then "sync" key in DCS has sync_standby=postgres2 after 10 seconds
+    When I start postgres1
+    Then "members/postgres1" key in DCS has state=running after 10 seconds
+    And Status code on GET http://127.0.0.1:8010/sync is 200 after 3 seconds
+    And Status code on GET http://127.0.0.1:8009/async is 200 after 3 seconds
+
+  Scenario: check stuck sync replica
+    Given I issue a PATCH request to http://127.0.0.1:8008/config with {"pause": true, "maximum_lag_on_syncnode": 15000000, "postgresql": {"parameters": {"synchronous_commit": "remote_apply"}}}
+    Then I receive a response code 200
+    And I create table on postgres0
+    And table mytest is present on postgres1 after 2 seconds
+    And table mytest is present on postgres2 after 2 seconds
+    When I pause wal replay on postgres2
+    And I load data on postgres0
+    Then "sync" key in DCS has sync_standby=postgres1 after 15 seconds
+    And I resume wal replay on postgres2
+    And Status code on GET http://127.0.0.1:8009/sync is 200 after 3 seconds
+    And Status code on GET http://127.0.0.1:8010/async is 200 after 3 seconds
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"pause": null, "maximum_lag_on_syncnode": -1, "postgresql": {"parameters": {"synchronous_commit": "on"}}}
+    Then I receive a response code 200
+    And I drop table on postgres0
+
+  Scenario: check multi sync replication
+    Given I issue a PATCH request to http://127.0.0.1:8008/config with {"synchronous_node_count": 2}
+    Then I receive a response code 200
+    Then "sync" key in DCS has sync_standby=postgres1,postgres2 after 10 seconds
+    And Status code on GET http://127.0.0.1:8010/sync is 200 after 3 seconds
+    And Status code on GET http://127.0.0.1:8009/sync is 200 after 3 seconds
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"synchronous_node_count": 1}
+    Then I receive a response code 200
+    And I shut down postgres1
+    Then "sync" key in DCS has sync_standby=postgres2 after 10 seconds
+    When I start postgres1
+    Then "members/postgres1" key in DCS has state=running after 10 seconds
+    And Status code on GET http://127.0.0.1:8010/sync is 200 after 3 seconds
+    And Status code on GET http://127.0.0.1:8009/async is 200 after 3 seconds
 
   Scenario: check the basic failover in synchronous mode
-    When I kill postgres0
-    Then postgres2 role is the primary after 22 seconds
-    When I issue a PATCH request to http://127.0.0.1:8009/config with {"synchronous_mode": null, "master_start_timeout": 0}
+    Given I run patronictl.py pause batman
+    Then I receive a response returncode 0
+    When I sleep for 2 seconds
+    And I shut down postgres0
+    And I run patronictl.py resume batman
+    Then I receive a response returncode 0
+    And postgres2 role is the primary after 24 seconds
+    And Response on GET http://127.0.0.1:8010/history contains recovery after 10 seconds
+    And there is a postgres2_cb.log with "on_role_change master batman" in postgres2 data directory
+    When I issue a PATCH request to http://127.0.0.1:8010/config with {"synchronous_mode": null, "master_start_timeout": 0}
     Then I receive a response code 200
     When I add the table bar to postgres2
     Then table bar is present on postgres1 after 20 seconds
+    And Response on GET http://127.0.0.1:8010/config contains master_start_timeout after 10 seconds
 
   Scenario: check immediate failover when master_start_timeout=0
     Given I kill postmaster on postgres2
     Then postgres1 is a leader after 10 seconds
     And postgres1 role is the primary after 10 seconds
 
-  Scenario: check rejoin of the former master with pg_rewind
-    Given I start postgres0
+  Scenario: check rejoin of the former primary with pg_rewind
+    Given I add the table splitbrain to postgres0
+    And I start postgres0
     Then postgres0 role is the secondary after 20 seconds
     When I add the table buz to postgres1
     Then table buz is present on postgres0 after 20 seconds

features/callback2.py

@@ -0,0 +1,5 @@
+#!/usr/bin/env python
+
+import sys
+with open("data/{0}/{0}_cb.log".format(sys.argv[1]), "a+") as log:
+    log.write(" ".join(sys.argv[-3:]) + "\n")

features/citus.feature

@@ -0,0 +1,72 @@
+Feature: citus
+  We should check that coordinator discovers and registers workers and clients don't have errors when worker cluster switches over
+
+  Scenario: check that worker cluster is registered in the coordinator
+    Given I start postgres0 in citus group 0
+    And I start postgres2 in citus group 1
+    Then postgres0 is a leader in a group 0 after 10 seconds
+    And postgres2 is a leader in a group 1 after 10 seconds
+    When I start postgres1 in citus group 0
+    And I start postgres3 in citus group 1
+    Then replication works from postgres0 to postgres1 after 15 seconds
+    Then replication works from postgres2 to postgres3 after 15 seconds
+    And postgres0 is registered in the coordinator postgres0 as the worker in group 0
+    And postgres2 is registered in the coordinator postgres0 as the worker in group 1
+
+  Scenario: coordinator failover updates pg_dist_node
+    Given I run patronictl.py failover batman --group 0 --candidate postgres1 --force
+    Then postgres1 role is the primary after 10 seconds
+    And replication works from postgres1 to postgres0 after 15 seconds
+    And "sync" key in a group 0 in DCS has sync_standby=postgres0 after 15 seconds
+    And postgres1 is registered in the coordinator postgres1 as the worker in group 0
+    When I run patronictl.py failover batman --group 0 --candidate postgres0 --force
+    Then postgres0 role is the primary after 10 seconds
+    And replication works from postgres0 to postgres1 after 15 seconds
+    And "sync" key in a group 0 in DCS has sync_standby=postgres1 after 15 seconds
+    And postgres0 is registered in the coordinator postgres0 as the worker in group 0
+
+  Scenario: worker switchover doesn't break client queries on the coordinator
+    Given I create a distributed table on postgres0
+    And I start a thread inserting data on postgres0
+    When I run patronictl.py switchover batman --group 1 --force
+    Then I receive a response returncode 0
+    And postgres3 role is the primary after 10 seconds
+    And replication works from postgres3 to postgres2 after 15 seconds
+    And "sync" key in a group 1 in DCS has sync_standby=postgres2 after 15 seconds
+    And postgres3 is registered in the coordinator postgres0 as the worker in group 1
+    And a thread is still alive
+    When I run patronictl.py switchover batman --group 1 --force
+    Then I receive a response returncode 0
+    And postgres2 role is the primary after 10 seconds
+    And replication works from postgres2 to postgres3 after 15 seconds
+    And "sync" key in a group 1 in DCS has sync_standby=postgres3 after 15 seconds
+    And postgres2 is registered in the coordinator postgres0 as the worker in group 1
+    And a thread is still alive
+    When I stop a thread
+    Then a distributed table on postgres0 has expected rows
+
+  Scenario: worker primary restart doesn't break client queries on the coordinator
+    Given I cleanup a distributed table on postgres0
+    And I start a thread inserting data on postgres0
+    When I run patronictl.py restart batman postgres2 --group 1 --force
+    Then I receive a response returncode 0
+    And postgres2 role is the primary after 10 seconds
+    And replication works from postgres2 to postgres3 after 15 seconds
+    And postgres2 is registered in the coordinator postgres0 as the worker in group 1
+    And a thread is still alive
+    When I stop a thread
+    Then a distributed table on postgres0 has expected rows
+
+  Scenario: check that in-flight transaction is rolled back after timeout when other workers need to change pg_dist_node
+    Given I start postgres4 in citus group 2
+    Then postgres4 is a leader in a group 2 after 10 seconds
+    And "members/postgres4" key in a group 2 in DCS has role=master after 3 seconds
+    When I run patronictl.py edit-config batman --group 2 -s ttl=20 --force
+    Then I receive a response returncode 0
+    And I receive a response output "+ttl: 20"
+    When I sleep for 2 seconds
+    Then postgres4 is registered in the coordinator postgres0 as the worker in group 2
+    When I shut down postgres4
+    Then There is a transaction in progress on postgres0 changing pg_dist_node
+    When I run patronictl.py restart batman postgres2 --group 1 --force
+    Then a transaction finishes in 20 seconds

features/custom_bootstrap.feature

@@ -0,0 +1,17 @@
+Feature: custom bootstrap
+    We should check that patroni can bootstrap a new cluster from a backup
+
+Scenario: clone existing cluster using pg_basebackup
+    Given I start postgres0
+    Then postgres0 is a leader after 10 seconds
+    When I add the table foo to postgres0
+    And I start postgres1 in a cluster batman1 as a clone of postgres0
+    Then postgres1 is a leader of batman1 after 10 seconds
+    Then table foo is present on postgres1 after 10 seconds
+
+Scenario: make a backup and do a restore into a new cluster
+    Given I add the table bar to postgres1
+    And I do a backup of postgres1
+    When I start postgres2 in a cluster batman2 from backup
+    Then postgres2 is a leader of batman2 after 30 seconds
+    And table bar is present on postgres2 after 10 seconds

features/dcs_failsafe_mode.feature

@@ -0,0 +1,85 @@
+Feature: dcs failsafe mode
+  We should check the basic dcs failsafe mode functioning
+
+  Scenario: check failsafe mode can be successfully enabled
+    Given I start postgres0
+    And postgres0 is a leader after 10 seconds
+    And I sleep for 3 seconds
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"loop_wait": 2, "ttl": 20, "retry_timeout": 5, "failsafe_mode": true}
+    Then I receive a response code 200
+    And Response on GET http://127.0.0.1:8008/failsafe contains postgres0 after 10 seconds
+    When I issue a GET request to http://127.0.0.1:8008/failsafe
+    Then I receive a response code 200
+    And I receive a response postgres0 http://127.0.0.1:8008/patroni
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"postgresql": {"parameters": {"wal_level": "logical"}}}
+    Then I receive a response code 200
+    When I issue a PATCH request to http://127.0.0.1:8008/config with {"slots": {"dcs_slot_0": {"type": "logical", "database": "postgres", "plugin": "test_decoding"}}}
+    Then I receive a response code 200
+
+  @dcs-failsafe
+  Scenario: check one-node cluster is functioning while DCS is down
+    Given DCS is down
+    Then Response on GET http://127.0.0.1:8008/primary contains failsafe_mode_is_active after 12 seconds
+    And postgres0 role is the primary after 10 seconds
+
+  @dcs-failsafe
+  Scenario: check new replica isn't promoted when leader is down and DCS is up
+    Given DCS is up
+    When I do a backup of postgres0
+    And I shut down postgres0
+    When I start postgres1 in a cluster batman from backup with no_leader
+    And I sleep for 2 seconds
+    Then postgres1 role is the replica after 12 seconds
+
+  Scenario: check leader and replica are both in /failsafe key after leader is back
+    Given I start postgres0
+    And I start postgres1
+    Then "members/postgres0" key in DCS has state=running after 10 seconds
+    And "members/postgres1" key in DCS has state=running after 2 seconds
+    And Response on GET http://127.0.0.1:8009/failsafe contains postgres1 after 10 seconds
+    When I issue a GET request to http://127.0.0.1:8009/failsafe
+    Then I receive a response code 200
+    And I receive a response postgres0 http://127.0.0.1:8008/patroni
+    And I receive a response postgres1 http://127.0.0.1:8009/patroni
+
+  @dcs-failsafe
+  @slot-advance
+  Scenario: check leader and replica are functioning while DCS is down
+    Given logical slot dcs_slot_0 is in sync between postgres0 and postgres1 after 10 seconds
+    And DCS is down
+    Then Response on GET http://127.0.0.1:8008/primary contains failsafe_mode_is_active after 12 seconds
+    Then postgres0 role is the primary after 10 seconds
+    And postgres1 role is the replica after 2 seconds
+    And replication works from postgres0 to postgres1 after 10 seconds
+    And I get all changes from logical slot dcs_slot_0 on postgres0
+    And logical slot dcs_slot_0 is in sync between postgres0 and postgres1 after 20 seconds
+
+  @dcs-failsafe
+  Scenario: check primary is demoted when one replica is shut down and DCS is down
+    Given DCS is down
+    And I kill postgres1
+    And I kill postmaster on postgres1
+    And I sleep for 2 seconds
+    Then postgres0 role is the replica after 12 seconds
+
+  @dcs-failsafe
+  Scenario: check known replica is promoted when leader is down and DCS is up
+    Given I shut down postgres0
+    And DCS is up
+    When I start postgres1
+    Then "members/postgres1" key in DCS has state=running after 10 seconds
+    And postgres1 role is the primary after 25 seconds
+
+  @dcs-failsafe
+  Scenario: check three-node cluster is functioning while DCS is down
+    Given I start postgres0
+    And I start postgres2
+    Then "members/postgres2" key in DCS has state=running after 10 seconds
+    And "members/postgres0" key in DCS has state=running after 20 seconds
+    And Response on GET http://127.0.0.1:8008/failsafe contains postgres2 after 10 seconds
+    And replication works from postgres1 to postgres0 after 10 seconds
+    Given DCS is down
+    Then Response on GET http://127.0.0.1:8008/primary contains failsafe_mode_is_active after 12 seconds
+    Then postgres1 role is the primary after 10 seconds
+    And postgres0 role is the replica after 2 seconds
+    And postgres2 role is the replica after 2 seconds

features/ignored_slots.feature

@@ -0,0 +1,61 @@
+Feature: ignored slots
+  Scenario: check ignored slots aren't removed on failover/switchover
+    Given I start postgres1
+    Then postgres1 is a leader after 10 seconds
+    And there is a non empty initialize key in DCS after 15 seconds
+    When I issue a PATCH request to http://127.0.0.1:8009/config with {"ignore_slots": [{"name": "unmanaged_slot_0", "database": "postgres", "plugin": "test_decoding", "type": "logical"}, {"name": "unmanaged_slot_1", "database": "postgres", "plugin": "test_decoding"}, {"name": "unmanaged_slot_2", "database": "postgres"}, {"name": "unmanaged_slot_3"}], "postgresql": {"parameters": {"wal_level": "logical"}}}
+    Then I receive a response code 200
+    And Response on GET http://127.0.0.1:8009/config contains ignore_slots after 10 seconds
+    # Make sure the wal_level has been changed.
+    When I shut down postgres1
+    And I start postgres1
+    Then postgres1 is a leader after 10 seconds
+    And "members/postgres1" key in DCS has role=master after 3 seconds
+    # Make sure Patroni has finished telling Postgres it should be accepting writes.
+    And postgres1 role is the primary after 20 seconds
+    # 1. Create our test logical replication slot.
+    # Test that ny subset of attributes in the ignore slots matcher is enough to match a slot
+    # by using 3 different slots.
+    When I create a logical replication slot unmanaged_slot_0 on postgres1 with the test_decoding plugin
+    And I create a logical replication slot unmanaged_slot_1 on postgres1 with the test_decoding plugin
+    And I create a logical replication slot unmanaged_slot_2 on postgres1 with the test_decoding plugin
+    And I create a logical replication slot unmanaged_slot_3 on postgres1 with the test_decoding plugin
+    And I create a logical replication slot dummy_slot on postgres1 with the test_decoding plugin
+    # It seems like it'd be obvious that these slots exist since we just created them,
+    # but Patroni can actually end up dropping them almost immediately, so it's helpful
+    # to verify they exist before we begin testing whether they persist through failover
+    # cycles.
+    Then postgres1 has a logical replication slot named unmanaged_slot_0 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_1 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_2 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_3 with the test_decoding plugin
+
+    When I start postgres0
+    Then "members/postgres0" key in DCS has role=replica after 3 seconds
+    And postgres0 role is the secondary after 20 seconds
+    # Verify that the replica has advanced beyond the point in the WAL
+    # where we created the replication slot so that on the next failover
+    # cycle we don't accidentally rewind to before the slot creation.
+    And replication works from postgres1 to postgres0 after 20 seconds
+    When I shut down postgres1
+    Then "members/postgres0" key in DCS has role=master after 3 seconds
+
+    # 2. After a failover the server (now a replica) still has the slot.
+    When I start postgres1
+    Then postgres1 role is the secondary after 20 seconds
+    And "members/postgres1" key in DCS has role=replica after 3 seconds
+    # give Patroni time to sync replication slots
+    And I sleep for 2 seconds
+    And postgres1 has a logical replication slot named unmanaged_slot_0 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_1 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_2 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_3 with the test_decoding plugin
+    And postgres1 does not have a logical replication slot named dummy_slot
+
+    # 3. After a failover the server (now a primary) still has the slot.
+    When I shut down postgres0
+    Then "members/postgres1" key in DCS has role=master after 3 seconds
+    And postgres1 has a logical replication slot named unmanaged_slot_0 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_1 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_2 with the test_decoding plugin
+    And postgres1 has a logical replication slot named unmanaged_slot_3 with the test_decoding plugin

features/patroni_api.feature

@@ -8,48 +8,70 @@ Scenario: check API requests on a stand-alone server
 	Then I receive a response code 200
 	And I receive a response state running
 	And I receive a response role master
+	When I issue a GET request to http://127.0.0.1:8008/standby_leader
+	Then I receive a response code 503
+	When I issue a GET request to http://127.0.0.1:8008/health
+	Then I receive a response code 200
 	When I issue a GET request to http://127.0.0.1:8008/replica
 	Then I receive a response code 503
-	When I run patronictl.py reinit batman postgres0 --force
-	Then I receive a response returncode 0
-	And I receive a response output "Failed: reinitialize for member postgres0, status code=503, (I am the leader, can not reinitialize)"
-	When I run patronictl.py failover batman --master postgres0 --force
+	When I issue a POST request to http://127.0.0.1:8008/reinitialize with {"force": true}
+	Then I receive a response code 503
+	And I receive a response text I am the leader, can not reinitialize
+	When I run patronictl.py switchover batman --master postgres0 --force
 	Then I receive a response returncode 1
-	And I receive a response output "Error: No candidates found to failover to"
-	When I issue a POST request to http://127.0.0.1:8008/failover with {"leader": "postgres0"}
-	Then I receive a response code 500
-	And I receive a response text failover is not possible: cluster does not have members except leader
+	And I receive a response output "Error: No candidates found to switchover to"
+	When I issue a POST request to http://127.0.0.1:8008/switchover with {"leader": "postgres0"}
+	Then I receive a response code 412
+	And I receive a response text switchover is not possible: cluster does not have members except leader
 	When I issue an empty POST request to http://127.0.0.1:8008/failover
 	Then I receive a response code 400
 	When I issue a POST request to http://127.0.0.1:8008/failover with {"foo": "bar"}
 	Then I receive a response code 400
-	And I receive a response text "No values given for required parameters leader and candidate"
+	And I receive a response text "Failover could be performed only to a specific candidate"
 
 Scenario: check local configuration reload
-	Given I issue an empty POST request to http://127.0.0.1:8008/reload
-	Then I receive a response code 200
-	And I receive a response text nothing changed
-	When I add tag new_tag new_value to postgres0 config
+	Given I add tag new_tag new_value to postgres0 config
 	And I issue an empty POST request to http://127.0.0.1:8008/reload
 	Then I receive a response code 202
 
 Scenario: check dynamic configuration change via DCS
-	Given I issue a PATCH request to http://127.0.0.1:8008/config with {"ttl": 10, "loop_wait": 2, "postgresql": {"parameters": {"max_connections": 101}}}
-	Then I receive a response code 200
-	And I receive a response loop_wait 2
+	Given I run patronictl.py edit-config -s 'ttl=10' -p 'max_connections=101' --force batman
+	Then I receive a response returncode 0
+	And I receive a response output "+ttl: 10"
 	And Response on GET http://127.0.0.1:8008/patroni contains pending_restart after 11 seconds
 	When I issue a GET request to http://127.0.0.1:8008/config
 	Then I receive a response code 200
-	And I receive a response loop_wait 2
+	And I receive a response ttl 10
 	When I issue a GET request to http://127.0.0.1:8008/patroni
 	Then I receive a response code 200
-	And I receive a response tags {'tag': 'new_value'}
+	And I receive a response tags {'new_tag': 'new_value'}
+	And I sleep for 4 seconds
+
+Scenario: check the scheduled restart
+	Given I issue a PATCH request to http://127.0.0.1:8008/config with {"postgresql": {"parameters": {"superuser_reserved_connections": "6"}}}
+	Then I receive a response code 200
+		And Response on GET http://127.0.0.1:8008/patroni contains pending_restart after 5 seconds
+	Given I issue a scheduled restart at http://127.0.0.1:8008 in 5 seconds with {"role": "replica"}
+	Then I receive a response code 202
+		And I sleep for 8 seconds
+		And Response on GET http://127.0.0.1:8008/patroni contains pending_restart after 10 seconds
+	Given I issue a scheduled restart at http://127.0.0.1:8008 in 5 seconds with {"restart_pending": "True"}
+	Then I receive a response code 202
+		And Response on GET http://127.0.0.1:8008/patroni does not contain pending_restart after 10 seconds
+		And postgres0 role is the primary after 10 seconds
 
 Scenario: check API requests for the primary-replica pair in the pause mode
-	Given I run patronictl.py pause batman
-	Then I receive a response returncode 0
-	When I start postgres1
+	Given I start postgres1
 	Then replication works from postgres0 to postgres1 after 20 seconds
+	When I run patronictl.py pause batman
+	Then I receive a response returncode 0
+	When I kill postmaster on postgres1
+	And I issue a GET request to http://127.0.0.1:8009/replica
+	Then I receive a response code 503
+	When I run patronictl.py restart batman postgres1 --force
+	Then I receive a response returncode 0
+	Then replication works from postgres0 to postgres1 after 20 seconds
+	And I sleep for 2 seconds
 	When I issue a GET request to http://127.0.0.1:8009/replica
 	Then I receive a response code 200
 	And I receive a response state running
@@ -64,36 +86,41 @@ Scenario: check API requests for the primary-replica pair in the pause mode
 	When I sleep for 10 seconds
 	Then postgres1 role is the secondary after 15 seconds
 
-Scenario: check the failover via the API in the pause mode
-	Given I run patronictl.py failover batman --master postgres0 --candidate postgres1 --force
-	Then I receive a response returncode 0
+Scenario: check the switchover via the API in the pause mode
+	Given I issue a POST request to http://127.0.0.1:8008/switchover with {"leader": "postgres0", "candidate": "postgres1"}
+	Then I receive a response code 200
 	And postgres1 is a leader after 5 seconds
 	And postgres1 role is the primary after 10 seconds
 	And postgres0 role is the secondary after 10 seconds
 	And replication works from postgres1 to postgres0 after 20 seconds
+	And "members/postgres0" key in DCS has state=running after 10 seconds
+	When I issue a GET request to http://127.0.0.1:8008/primary
+	Then I receive a response code 503
+	When I issue a GET request to http://127.0.0.1:8008/replica
+	Then I receive a response code 200
+	When I issue a GET request to http://127.0.0.1:8009/primary
+	Then I receive a response code 200
+	When I issue a GET request to http://127.0.0.1:8009/replica
+	Then I receive a response code 503
 
-Scenario: check the scheduled failover
-	Given I issue a scheduled failover from postgres1 to postgres0 in 3 seconds
+Scenario: check the scheduled switchover
+	Given I issue a scheduled switchover from postgres1 to postgres0 in 10 seconds
 	Then I receive a response returncode 1
-	And I receive a response output "Can't schedule failover in the paused state"
+	And I receive a response output "Can't schedule switchover in the paused state"
 	When I run patronictl.py resume batman
 	Then I receive a response returncode 0
-	Given I issue a scheduled failover from postgres1 to postgres0 in 3 seconds
+	Given I issue a scheduled switchover from postgres1 to postgres0 in 10 seconds
 	Then I receive a response returncode 0
 	And postgres0 is a leader after 20 seconds
 	And postgres0 role is the primary after 10 seconds
 	And postgres1 role is the secondary after 10 seconds
 	And replication works from postgres0 to postgres1 after 25 seconds
-
-Scenario: check the scheduled restart
-	Given I issue a PATCH request to http://127.0.0.1:8008/config with {"postgresql": {"parameters": {"superuser_reserved_connections": "6"}}}
+	And "members/postgres1" key in DCS has state=running after 10 seconds
+	When I issue a GET request to http://127.0.0.1:8008/primary
+	Then I receive a response code 200
+	When I issue a GET request to http://127.0.0.1:8008/replica
+	Then I receive a response code 503
+	When I issue a GET request to http://127.0.0.1:8009/primary
+	Then I receive a response code 503
+	When I issue a GET request to http://127.0.0.1:8009/replica
 	Then I receive a response code 200
-		And Response on GET http://127.0.0.1:8008/patroni contains pending_restart after 5 seconds
-	Given I issue a scheduled restart at http://127.0.0.1:8008 in 3 seconds with {"role": "replica"}
-	Then I receive a response code 202
-		And I sleep for 4 seconds
-		And Response on GET http://127.0.0.1:8008/patroni contains pending_restart after 10 seconds
-	Given I issue a scheduled restart at http://127.0.0.1:8008 in 3 seconds with {"restart_pending": "True"}
-	Then I receive a response code 202
-		And Response on GET http://127.0.0.1:8008/patroni does not contain pending_restart after 10 seconds
-

features/standby_cluster.feature

@@ -0,0 +1,60 @@
+Feature: standby cluster
+  Scenario: prepare the cluster with logical slots
+    Given I start postgres1
+    Then postgres1 is a leader after 10 seconds
+    And there is a non empty initialize key in DCS after 15 seconds
+    When I issue a PATCH request to http://127.0.0.1:8009/config with {"slots": {"pm_1": {"type": "physical"}}, "postgresql": {"parameters": {"wal_level": "logical"}}}
+    Then I receive a response code 200
+    And Response on GET http://127.0.0.1:8009/config contains slots after 10 seconds
+    And I sleep for 3 seconds
+    When I issue a PATCH request to http://127.0.0.1:8009/config with {"slots": {"test_logical": {"type": "logical", "database": "postgres", "plugin": "test_decoding"}}}
+    Then I receive a response code 200
+    And I do a backup of postgres1
+    When I start postgres0
+    Then "members/postgres0" key in DCS has state=running after 10 seconds
+    And replication works from postgres1 to postgres0 after 15 seconds
+
+  @slot-advance
+  Scenario: check permanent logical slots are synced to the replica
+    Given I run patronictl.py restart batman postgres1 --force
+    Then Logical slot test_logical is in sync between postgres0 and postgres1 after 10 seconds
+    When I add the table replicate_me to postgres1
+    And I get all changes from logical slot test_logical on postgres1
+    Then Logical slot test_logical is in sync between postgres0 and postgres1 after 10 seconds
+
+  Scenario: Detach exiting node from the cluster
+    When I shut down postgres1
+    Then postgres0 is a leader after 10 seconds
+    And "members/postgres0" key in DCS has role=master after 3 seconds
+    When I issue a GET request to http://127.0.0.1:8008/
+    Then I receive a response code 200
+
+  Scenario: check replication of a single table in a standby cluster
+    Given I start postgres1 in a standby cluster batman1 as a clone of postgres0
+    Then postgres1 is a leader of batman1 after 10 seconds
+    When I add the table foo to postgres0
+    Then table foo is present on postgres1 after 20 seconds
+    And I sleep for 3 seconds
+    When I issue a GET request to http://127.0.0.1:8009/primary
+    Then I receive a response code 503
+    When I issue a GET request to http://127.0.0.1:8009/standby_leader
+    Then I receive a response code 200
+    And I receive a response role standby_leader
+    And there is a postgres1_cb.log with "on_role_change standby_leader batman1" in postgres1 data directory
+    When I start postgres2 in a cluster batman1
+    Then postgres2 role is the replica after 24 seconds
+    And table foo is present on postgres2 after 20 seconds
+    And postgres1 does not have a logical replication slot named test_logical
+
+  Scenario: check failover
+    When I kill postgres1
+    And I kill postmaster on postgres1
+    Then postgres2 is replicating from postgres0 after 32 seconds
+    When I issue a GET request to http://127.0.0.1:8010/primary
+    Then I receive a response code 503
+    And I sleep for 3 seconds
+    When I issue a GET request to http://127.0.0.1:8010/standby_leader
+    Then I receive a response code 200
+    And I receive a response role standby_leader
+    And replication works from postgres0 to postgres2 after 15 seconds
+    And there is a postgres2_cb.log with "on_start replica batman1\non_role_change standby_leader batman1" in postgres2 data directory

features/steps/basic_replication.py

@@ -1,4 +1,4 @@
-import psycopg2 as pg
+import patroni.psycopg as pg
 
 from behave import step, then
 from time import sleep, time
@@ -11,7 +11,7 @@ def start_patroni(context, name):
 
 @step('I shut down {name:w}')
 def stop_patroni(context, name):
-    return context.pctl.stop(name)
+    return context.pctl.stop(name, timeout=60)
 
 
 @step('I kill {name:w}')
@@ -28,16 +28,47 @@ def stop_postgres(context, name):
 def add_table(context, table_name, pg_name):
     # parse the configuration file and get the port
     try:
-        context.pctl.query(pg_name, "CREATE TABLE {0}()".format(table_name))
+        context.pctl.query(pg_name, "CREATE TABLE public.{0}()".format(table_name))
     except pg.Error as e:
         assert False, "Error creating table {0} on {1}: {2}".format(table_name, pg_name, e)
 
 
+@step('I {action:w} wal replay on {pg_name:w}')
+def toggle_wal_replay(context, action, pg_name):
+    # pause or resume the wal replay process
+    try:
+        version = context.pctl.query(pg_name, "SHOW server_version_num").fetchone()[0]
+        wal_name = 'xlog' if int(version)/10000 < 10 else 'wal'
+        context.pctl.query(pg_name, "SELECT pg_{0}_replay_{1}()".format(wal_name, action))
+    except pg.Error as e:
+        assert False, "Error during {0} wal recovery on {1}: {2}".format(action, pg_name, e)
+
+
+@step('I {action:w} table on {pg_name:w}')
+def crdr_mytest(context, action, pg_name):
+    try:
+        if (action == "create"):
+            context.pctl.query(pg_name, "create table if not exists public.mytest(id numeric)")
+        else:
+            context.pctl.query(pg_name, "drop table if exists public.mytest")
+    except pg.Error as e:
+        assert False, "Error {0} table mytest on {1}: {2}".format(action, pg_name, e)
+
+
+@step('I load data on {pg_name:w}')
+def initiate_load(context, pg_name):
+    # perform dummy load
+    try:
+        context.pctl.query(pg_name, "insert into public.mytest select r::numeric from generate_series(1, 350000) r")
+    except pg.Error as e:
+        assert False, "Error loading test data on {0}: {1}".format(pg_name, e)
+
+
 @then('Table {table_name:w} is present on {pg_name:w} after {max_replication_delay:d} seconds')
 def table_is_present_on(context, table_name, pg_name, max_replication_delay):
     max_replication_delay *= context.timeout_multiplier
     for _ in range(int(max_replication_delay)):
-        if context.pctl.query(pg_name, "SELECT 1 FROM {0}".format(table_name), fail_ok=True) is not None:
+        if context.pctl.query(pg_name, "SELECT 1 FROM public.{0}".format(table_name), fail_ok=True) is not None:
             break
         sleep(1)
     else:
@@ -52,10 +83,10 @@ def check_role(context, pg_name, pg_role, max_promotion_timeout):
         "{0} role didn't change to {1} after {2} seconds".format(pg_name, pg_role, max_promotion_timeout)
 
 
-@step('replication works from {master:w} to {replica:w} after {time_limit:d} seconds')
-@then('replication works from {master:w} to {replica:w} after {time_limit:d} seconds')
-def replication_works(context, master, replica, time_limit):
+@step('replication works from {primary:w} to {replica:w} after {time_limit:d} seconds')
+@then('replication works from {primary:w} to {replica:w} after {time_limit:d} seconds')
+def replication_works(context, primary, replica, time_limit):
     context.execute_steps(u"""
         When I add the table test_{0} to {1}
         Then table test_{0} is present on {2} after {3} seconds
-    """.format(int(time()), master, replica, time_limit))
+    """.format(int(time()), primary, replica, time_limit))

features/steps/cascading_replication.py

@@ -6,13 +6,13 @@ from behave import step, then
 
 @step('I configure and start {name:w} with a tag {tag_name:w} {tag_value:w}')
 def start_patroni_with_a_name_value_tag(context, name, tag_name, tag_value):
-    return context.pctl.start(name, tags={tag_name: tag_value})
+    return context.pctl.start(name, custom_config={'tags': {tag_name: tag_value}})
 
 
-@then('There is a label with "{content:w}" in {name:w} data directory')
-def check_label(context, content, name):
-    label = context.pctl.read_label(name)
-    assert label == content, "{0} is not equal to {1}".format(label, content)
+@then('There is a {label} with "{content}" in {name:w} data directory')
+def check_label(context, label, content, name):
+    value = (context.pctl.read_label(name, label) or '').replace('\n', '\\n')
+    assert content in value, "\"{0}\" in {1} doesn't contain {2}".format(value, label, content)
 
 
 @step('I create label with "{content:w}" in {name:w} data directory')
@@ -20,16 +20,33 @@ def write_label(context, content, name):
     context.pctl.write_label(name, content)
 
 
-@step('"{name}" key in DCS has {key:w}={value:w} after {time_limit:d} seconds')
+@step('"{name}" key in DCS has {key:w}={value} after {time_limit:d} seconds')
 def check_member(context, name, key, value, time_limit):
     time_limit *= context.timeout_multiplier
     max_time = time.time() + int(time_limit)
+    dcs_value = None
     while time.time() < max_time:
         try:
             response = json.loads(context.dcs_ctl.query(name))
-            if response.get(key) == value:
+            dcs_value = response.get(key)
+            if dcs_value == value:
                 return
         except Exception:
             pass
         time.sleep(1)
-    assert False, "{0} does not have {1}={2} in dcs after {3} seconds".format(name, key, value, time_limit)
+    assert False, "{0} does not have {1}={2} (found {3}) in dcs after {4} seconds".format(name, key, value,
+                                                                                          dcs_value, time_limit)
+
+
+@step('there is a non empty {key:w} key in DCS after {time_limit:d} seconds')
+def check_initialize(context, key, time_limit):
+    time_limit *= context.timeout_multiplier
+    max_time = time.time() + int(time_limit)
+    while time.time() < max_time:
+        try:
+            if context.dcs_ctl.query(key):
+                return
+        except Exception:
+            pass
+        time.sleep(1)
+    assert False, "There is no {0} in dcs after {1} seconds".format(key, time_limit)

features/steps/citus.py

@@ -0,0 +1,117 @@
+import json
+import time
+
+from behave import step, then
+from dateutil import tz
+from datetime import datetime
+from functools import partial
+from threading import Thread, Event
+
+tzutc = tz.tzutc()
+
+
+@step('{name:w} is a leader in a group {group:d} after {time_limit:d} seconds')
+@then('{name:w} is a leader in a group {group:d} after {time_limit:d} seconds')
+def is_a_group_leader(context, name, group, time_limit):
+    time_limit *= context.timeout_multiplier
+    max_time = time.time() + int(time_limit)
+    while (context.dcs_ctl.query("leader", group=group) != name):
+        time.sleep(1)
+        assert time.time() < max_time, "{0} is not a leader in dcs after {1} seconds".format(name, time_limit)
+
+
+@step('"{name}" key in a group {group:d} in DCS has {key:w}={value} after {time_limit:d} seconds')
+def check_group_member(context, name, group, key, value, time_limit):
+    time_limit *= context.timeout_multiplier
+    max_time = time.time() + int(time_limit)
+    dcs_value = None
+    response = None
+    while time.time() < max_time:
+        try:
+            response = json.loads(context.dcs_ctl.query(name, group=group))
+            dcs_value = response.get(key)
+            if dcs_value == value:
+                return
+        except Exception:
+            pass
+        time.sleep(1)
+    assert False, ("{0} in a group {1} does not have {2}={3} (found {4}) in dcs" +
+                   " after {5} seconds").format(name, group, key, value, response, time_limit)
+
+
+@step('I start {name:w} in citus group {group:d}')
+def start_citus(context, name, group):
+    return context.pctl.start(name, custom_config={"citus": {"database": "postgres", "group": int(group)}})
+
+
+@step('{name1:w} is registered in the coordinator {name2:w} as the worker in group {group:d}')
+def check_registration(context, name1, name2, group):
+    worker_port = int(context.pctl.query(name1, "SHOW port").fetchone()[0])
+    r = context.pctl.query(name2, "SELECT nodeport FROM pg_catalog.pg_dist_node WHERE groupid = {0}".format(group))
+    assert worker_port == r.fetchone()[0],\
+        "Worker {0} is not registered in pg_dist_node on the coordinator {1}".format(name1, name2)
+
+
+@step('I create a distributed table on {name:w}')
+def create_distributed_table(context, name):
+    context.pctl.query(name, 'CREATE TABLE public.d(id int not null)')
+    context.pctl.query(name, "SELECT create_distributed_table('public.d', 'id')")
+
+
+@step('I cleanup a distributed table on {name:w}')
+def cleanup_distributed_table(context, name):
+    context.pctl.query(name, 'TRUNCATE public.d')
+
+
+def insert_thread(query_func, context):
+    while True:
+        if context.thread_stop_event.is_set():
+            break
+
+        context.insert_counter += 1
+        query_func('INSERT INTO public.d VALUES({0})'.format(context.insert_counter))
+
+        context.thread_stop_event.wait(0.01)
+
+
+@step('I start a thread inserting data on {name:w}')
+def start_insert_thread(context, name):
+    context.thread_stop_event = Event()
+    context.insert_counter = 0
+    query_func = partial(context.pctl.query, name)
+    thread_func = partial(insert_thread, query_func, context)
+    context.thread = Thread(target=thread_func)
+    context.thread.daemon = True
+    context.thread.start()
+
+
+@then('a thread is still alive')
+def thread_is_alive(context):
+    assert context.thread.is_alive(), "Thread is not alive"
+
+
+@step("I stop a thread")
+def stop_insert_thread(context):
+    context.thread_stop_event.set()
+    context.thread.join(1*context.timeout_multiplier)
+    assert not context.thread.is_alive(), "Thread is still alive"
+
+
+@step("a distributed table on {name:w} has expected rows")
+def count_rows(context, name):
+    rows = context.pctl.query(name, "SELECT COUNT(*) FROM public.d").fetchone()[0]
+    assert rows == context.insert_counter, "Distributed table doesn't have expected amount of rows"
+
+
+@step("There is a transaction in progress on {name:w} changing pg_dist_node")
+def check_transaction(context, name):
+    cur = context.pctl.query(name, "SELECT xact_start FROM pg_stat_activity WHERE pid <> pg_backend_pid()"
+                                   " AND state = 'idle in transaction' AND query ~ 'citus_update_node'")
+    assert cur.rowcount == 1, "There is no idle in transaction updating pg_dist_node"
+    context.xact_start = cur.fetchone()[0]
+
+
+@step("a transaction finishes in {timeout:d} seconds")
+def check_transaction_timeout(context, timeout):
+    assert (datetime.now(tzutc) - context.xact_start).seconds > timeout,\
+            "a transaction finished earlier than in {0} seconds".format(timeout)

features/steps/custom_bootstrap.py

@@ -0,0 +1,27 @@
+import time
+
+from behave import step, then
+
+
+@step('I start {name:w} in a cluster {cluster_name:w} as a clone of {name2:w}')
+def start_cluster_clone(context, name, cluster_name, name2):
+    context.pctl.clone(name2, cluster_name, name)
+
+
+@step('I start {name:w} in a cluster {cluster_name:w} from backup')
+def start_cluster_from_backup(context, name, cluster_name):
+    context.pctl.bootstrap_from_backup(name, cluster_name)
+
+
+@then('{name:w} is a leader of {cluster_name:w} after {time_limit:d} seconds')
+def is_a_leader(context, name, cluster_name, time_limit):
+    time_limit *= context.timeout_multiplier
+    max_time = time.time() + int(time_limit)
+    while (context.dcs_ctl.query("leader", scope=cluster_name) != name):
+        time.sleep(1)
+        assert time.time() < max_time, "{0} is not a leader in dcs after {1} seconds".format(name, time_limit)
+
+
+@step('I do a backup of {name:w}')
+def do_backup(context, name):
+    context.pctl.backup(name)

features/steps/dcs_failsafe_mode.py

@@ -0,0 +1,16 @@
+from behave import step
+
+
+@step('DCS is down')
+def start_dcs_outage(context):
+    context.dcs_ctl.start_outage()
+
+
+@step('DCS is up')
+def stop_dcs_outage(context):
+    context.dcs_ctl.stop_outage()
+
+
+@step('I start {name:w} in a cluster {cluster_name:w} from backup with no_leader')
+def start_cluster_from_backup_no_leader(context, name, cluster_name):
+    context.pctl.bootstrap_from_backup_no_leader(name, cluster_name)

features/steps/patroni_api.py

@@ -1,9 +1,8 @@
 import json
-import os
 import parse
-import requests
 import shlex
 import subprocess
+import sys
 import time
 import yaml
 
@@ -43,9 +42,9 @@ def sleep_for_n_seconds(context, value):
 
 
 def _set_response(context, response):
-    context.status_code = response.status_code
-    data = response.content.decode('utf-8')
-    ct = response.headers.get('content-type', '')
+    context.status_code = response.status
+    data = response.data.decode('utf-8')
+    ct = response.getheader('content-type', '')
     if ct.startswith('application/json') or\
             ct.startswith('text/yaml') or\
             ct.startswith('text/x-yaml') or\
@@ -61,13 +60,7 @@ def _set_response(context, response):
 
 @step('I issue a GET request to {url:url}')
 def do_get(context, url):
-    try:
-        r = requests.get(url)
-    except requests.exceptions.RequestException:
-        context.status_code = None
-        context.response = None
-    else:
-        _set_response(context, r)
+    do_request(context, 'GET', url, None)
 
 
 @step('I issue an empty POST request to {url:url}')
@@ -77,27 +70,24 @@ def do_post_empty(context, url):
 
 @step('I issue a {request_method:w} request to {url:url} with {data}')
 def do_request(context, request_method, url, data):
-    data = data and json.loads(data) or {}
+    if context.certfile:
+        url = url.replace('http://', 'https://')
+    data = data and json.loads(data)
     try:
-        if request_method == 'PATCH':
-            r = requests.patch(url, json=data)
-        else:
-            r = requests.post(url, json=data)
-    except requests.exceptions.RequestException:
-        context.status_code = None
-        context.response = None
+        r = context.request_executor.request(request_method, url, data)
+        if request_method == 'PATCH' and r.status == 409:
+            r = context.request_executor.request(request_method, url, data)
+    except Exception:
+        context.status_code = context.response = None
     else:
         _set_response(context, r)
 
 
 @step('I run {cmd}')
 def do_run(context, cmd):
-    cmd = ['coverage', 'run', '--source=patroni', '-p'] + shlex.split(cmd)
+    cmd = [sys.executable, '-m', 'coverage', 'run', '--source=patroni', '-p'] + shlex.split(cmd)
     try:
-        # XXX: Dirty hack! We need to take name/passwd from the config!
-        env = os.environ.copy()
-        env.update({'PATRONI_RESTAPI_USERNAME': 'username', 'PATRONI_RESTAPI_PASSWORD': 'password'})
-        response = subprocess.check_output(cmd, stderr=subprocess.STDOUT, env=env)
+        response = subprocess.check_output(cmd, stderr=subprocess.STDOUT)
         context.status_code = 0
     except subprocess.CalledProcessError as e:
         response = e.output
@@ -111,20 +101,23 @@ def check_response(context, component, data):
         assert context.status_code == int(data),\
             "status code {0} != {1}, response: {2}".format(context.status_code, data, context.response)
     elif component == 'returncode':
-        assert context.status_code == int(data), "return code {0} != {1}".format(context.status_code, data)
+        assert context.status_code == int(data), "return code {0} != {1}, {2}".format(context.status_code,
+                                                                                      data, context.response)
     elif component == 'text':
         assert context.response == data.strip('"'), "response {0} does not contain {1}".format(context.response, data)
     elif component == 'output':
         assert data.strip('"') in context.response, "response {0} does not contain {1}".format(context.response, data)
     else:
         assert component in context.response, "{0} is not part of the response".format(component)
+        if context.certfile:
+            data = data.replace('http://', 'https://')
         assert str(context.response[component]) == str(data), "{0} does not contain {1}".format(component, data)
 
 
-@step('I issue a scheduled failover from {from_host:w} to {to_host:w} in {in_seconds:d} seconds')
-def scheduled_failover(context, from_host, to_host, in_seconds):
+@step('I issue a scheduled switchover from {from_host:w} to {to_host:w} in {in_seconds:d} seconds')
+def scheduled_switchover(context, from_host, to_host, in_seconds):
     context.execute_steps(u"""
-        Given I run patronictl.py failover batman --master {0} --candidate {1} --scheduled "{2}" --force
+        Given I run patronictl.py switchover batman --master {0} --candidate {1} --scheduled "{2}" --force
     """.format(from_host, to_host, datetime.now(tzutc) + timedelta(seconds=int(in_seconds))))
 
 
@@ -140,12 +133,28 @@ def add_tag_to_config(context, tag, value, pg_name):
     context.pctl.add_tag_to_config(pg_name, tag, value)
 
 
-@then('Response on GET {url} contains {value} after {timeout:d} seconds')
-def check_http_response(context, url, value, timeout, negate=False):
+@then('Status code on GET {url:url} is {code:d} after {timeout:d} seconds')
+def check_http_code(context, url, code, timeout):
+    if context.certfile:
+        url = url.replace('http://', 'https://')
     timeout *= context.timeout_multiplier
     for _ in range(int(timeout)):
-        r = requests.get(url)
-        if (value in r.content.decode('utf-8')) != negate:
+        r = context.request_executor.request('GET', url)
+        if int(code) == int(r.status):
+            break
+        time.sleep(1)
+    else:
+        assert False, "HTTP Status Code is not {0} after {1} seconds".format(code, timeout)
+
+
+@then('Response on GET {url:url} contains {value} after {timeout:d} seconds')
+def check_http_response(context, url, value, timeout, negate=False):
+    if context.certfile:
+        url = url.replace('http://', 'https://')
+    timeout *= context.timeout_multiplier
+    for _ in range(int(timeout)):
+        r = context.request_executor.request('GET', url)
+        if (value in r.data.decode('utf-8')) != negate:
             break
         time.sleep(1)
     else:

features/steps/slots.py

@@ -0,0 +1,60 @@
+import time
+
+from behave import step, then
+import patroni.psycopg as pg
+
+
+@step('I create a logical replication slot {slot_name} on {pg_name:w} with the {plugin:w} plugin')
+def create_logical_replication_slot(context, slot_name, pg_name, plugin):
+    try:
+        output = context.pctl.query(pg_name, ("SELECT pg_create_logical_replication_slot('{0}', '{1}'),"
+                                              " current_database()").format(slot_name, plugin))
+        print(output.fetchone())
+    except pg.Error as e:
+        print(e)
+        assert False, "Error creating slot {0} on {1} with plugin {2}".format(slot_name, pg_name, plugin)
+
+
+@then('{pg_name:w} has a logical replication slot named {slot_name} with the {plugin:w} plugin')
+def has_logical_replication_slot(context, pg_name, slot_name, plugin):
+    try:
+        row = context.pctl.query(pg_name, ("SELECT slot_type, plugin FROM pg_replication_slots"
+                                           " WHERE slot_name = '{0}'").format(slot_name)).fetchone()
+        assert row, "Couldn't find replication slot named {0}".format(slot_name)
+        assert row[0] == "logical", "Found replication slot named {0} but wasn't a logical slot".format(slot_name)
+        assert row[1] == plugin, ("Found replication slot named {0} but was using plugin "
+                                  "{1} rather than {2}").format(slot_name, row[1], plugin)
+    except pg.Error:
+        assert False, "Error looking for slot {0} on {1} with plugin {2}".format(slot_name, pg_name, plugin)
+
+
+@then('{pg_name:w} does not have a logical replication slot named {slot_name}')
+def does_not_have_logical_replication_slot(context, pg_name, slot_name):
+    try:
+        row = context.pctl.query(pg_name, ("SELECT 1 FROM pg_replication_slots"
+                                           " WHERE slot_name = '{0}'").format(slot_name)).fetchone()
+        assert not row, "Found unexpected replication slot named {0}".format(slot_name)
+    except pg.Error:
+        assert False, "Error looking for slot {0} on {1}".format(slot_name, pg_name)
+
+
+@step('Logical slot {slot_name:w} is in sync between {pg_name1:w} and {pg_name2:w} after {time_limit:d} seconds')
+def logical_slots_in_sync(context, slot_name, pg_name1, pg_name2, time_limit):
+    time_limit *= context.timeout_multiplier
+    max_time = time.time() + int(time_limit)
+    while time.time() < max_time:
+        try:
+            query = "SELECT confirmed_flush_lsn FROM pg_replication_slots WHERE slot_name = '{0}'".format(slot_name)
+            slot1 = context.pctl.query(pg_name1, query).fetchone()
+            slot2 = context.pctl.query(pg_name2, query).fetchone()
+            if slot1[0] == slot2[0]:
+                return
+        except Exception:
+            pass
+        time.sleep(1)
+    assert False, "Logical slot {0} is not in sync between {1} and {2}".format(slot_name, pg_name1, pg_name2)
+
+
+@step('I get all changes from logical slot {slot_name:w} on {pg_name:w}')
+def logical_slot_get_changes(context, slot_name, pg_name):
+    context.pctl.query(pg_name, "SELECT * FROM pg_logical_slot_get_changes('{0}', NULL, NULL)".format(slot_name))

features/steps/standby_cluster.py

@@ -0,0 +1,69 @@
+import os
+import time
+
+from behave import step
+
+
+def callbacks(context, name):
+    return {c: '{0} features/callback2.py {1}'.format(context.pctl.PYTHON, name)
+            for c in ('on_start', 'on_stop', 'on_restart', 'on_role_change')}
+
+
+@step('I start {name:w} in a cluster {cluster_name:w}')
+def start_patroni(context, name, cluster_name):
+    return context.pctl.start(name, custom_config={
+        "scope": cluster_name,
+        "postgresql": {
+            "callbacks": callbacks(context, name),
+            "backup_restore": {
+                "command": (context.pctl.PYTHON + " features/backup_restore.py --sourcedir=" +
+                            os.path.join(context.pctl.patroni_path, 'data', 'basebackup').replace('\\', '/'))}
+        }
+    })
+
+
+@step('I start {name:w} in a standby cluster {cluster_name:w} as a clone of {name2:w}')
+def start_patroni_standby_cluster(context, name, cluster_name, name2):
+    # we need to remove patroni.dynamic.json in order to "bootstrap" standby cluster with existing PGDATA
+    os.unlink(os.path.join(context.pctl._processes[name]._data_dir, 'patroni.dynamic.json'))
+    port = context.pctl._processes[name2]._connkwargs.get('port')
+    context.pctl._processes[name].update_config({
+        "scope": cluster_name,
+        "bootstrap": {
+            "dcs": {
+                "ttl": 20,
+                "loop_wait": 2,
+                "retry_timeout": 5,
+                "standby_cluster": {
+                    "host": "localhost",
+                    "port": port,
+                    "primary_slot_name": "pm_1",
+                    "create_replica_methods": ["backup_restore", "basebackup"]
+                },
+                "postgresql": {"parameters": {"wal_level": "logical"}}
+            }
+        },
+        "postgresql": {
+            "callbacks": callbacks(context, name)
+        }
+    })
+    return context.pctl.start(name)
+
+
+@step('{pg_name1:w} is replicating from {pg_name2:w} after {timeout:d} seconds')
+def check_replication_status(context, pg_name1, pg_name2, timeout):
+    bound_time = time.time() + timeout * context.timeout_multiplier
+
+    while time.time() < bound_time:
+        cur = context.pctl.query(
+            pg_name2,
+            "SELECT * FROM pg_catalog.pg_stat_replication WHERE application_name = '{0}'".format(pg_name1),
+            fail_ok=True
+        )
+
+        if cur and len(cur.fetchall()) != 0:
+            break
+
+        time.sleep(1)
+    else:
+        assert False, "{0} is not replicating from {1} after {2} seconds".format(pg_name1, pg_name2, timeout)

features/steps/watchdog.py

@@ -0,0 +1,54 @@
+from behave import step, then
+import time
+
+
+def polling_loop(timeout, interval=1):
+    """Returns an iterator that returns values until timeout has passed. Timeout is measured from start of iteration."""
+    start_time = time.time()
+    iteration = 0
+    end_time = start_time + timeout
+    while time.time() < end_time:
+        yield iteration
+        iteration += 1
+        time.sleep(interval)
+
+
+@step('I start {name:w} with watchdog')
+def start_patroni_with_watchdog(context, name):
+    return context.pctl.start(name, custom_config={'watchdog': True, 'bootstrap': {'dcs': {'ttl': 20}}})
+
+
+@step('{name:w} watchdog has been pinged after {timeout:d} seconds')
+def watchdog_was_pinged(context, name, timeout):
+    for _ in polling_loop(timeout):
+        if context.pctl.get_watchdog(name).was_pinged:
+            return True
+    return False
+
+
+@then('{name:w} watchdog has been closed')
+def watchdog_was_closed(context, name):
+    assert context.pctl.get_watchdog(name).was_closed
+
+
+@step('{name:w} watchdog has a {timeout:d} second timeout')
+def watchdog_has_timeout(context, name, timeout):
+    assert context.pctl.get_watchdog(name).timeout == timeout
+
+
+@step('I reset {name:w} watchdog state')
+def watchdog_reset_pinged(context, name):
+    context.pctl.get_watchdog(name).reset()
+
+
+@then('{name:w} watchdog is triggered after {timeout:d} seconds')
+def watchdog_was_triggered(context, name, timeout):
+    for _ in polling_loop(timeout):
+        if context.pctl.get_watchdog(name).was_triggered:
+            return True
+    assert False
+
+
+@step('{name:w} hangs for {timeout:d} seconds')
+def patroni_hang(context, name, timeout):
+    return context.pctl.patroni_hang(name, timeout)

features/watchdog.feature

@@ -0,0 +1,39 @@
+Feature: watchdog
+  Verify that watchdog gets pinged and triggered under appropriate circumstances.
+
+  Scenario: watchdog is opened and pinged
+    Given I start postgres0 with watchdog
+    Then postgres0 is a leader after 10 seconds
+    And postgres0 role is the primary after 10 seconds
+    And postgres0 watchdog has been pinged after 10 seconds
+    And postgres0 watchdog has a 15 second timeout
+
+  Scenario: watchdog is reconfigured after global ttl changed
+    Given I run patronictl.py edit-config batman -s ttl=30 --force
+    Then I receive a response returncode 0
+    And I receive a response output "+ttl: 30"
+    When I sleep for 4 seconds
+    Then postgres0 watchdog has a 25 second timeout
+
+  Scenario: watchdog is disabled during pause
+    Given I run patronictl.py pause batman
+    Then I receive a response returncode 0
+    When I sleep for 2 seconds
+    Then postgres0 watchdog has been closed
+
+  Scenario: watchdog is opened and pinged after resume
+    Given I reset postgres0 watchdog state
+    And I run patronictl.py resume batman
+    Then I receive a response returncode 0
+    And postgres0 watchdog has been pinged after 10 seconds
+
+  Scenario: watchdog is disabled when shutting down
+    Given I shut down postgres0
+    Then postgres0 watchdog has been closed
+
+  Scenario: watchdog is triggered if patroni stops responding
+    Given I reset postgres0 watchdog state
+    And I start postgres0 with watchdog
+    Then postgres0 role is the primary after 10 seconds
+    When postgres0 hangs for 30 seconds
+    Then postgres0 watchdog is triggered after 30 seconds

haproxy.cfg

@@ -1,21 +1,25 @@
 global
-	maxconn 100
+    maxconn 100
 
 defaults
-	log	global
-	mode	tcp
-	retries 2
-	timeout client 30m
-	timeout connect 4s
-	timeout server 30m
-	timeout check 5s
+    log global
+    mode tcp
+    retries 2
+    timeout client 30m
+    timeout connect 4s
+    timeout server 30m
+    timeout check 5s
 
-frontend ft_postgresql
-	bind *:5000
-	default_backend bk_db
-
-backend bk_db
-	option httpchk
+listen stats
+    mode http
+    bind *:7000
+    stats enable
+    stats uri /
 
+listen batman
+    bind *:5000
+    option httpchk
+    http-check expect status 200
+    default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
     server postgresql_127.0.0.1_5432 127.0.0.1:5432 maxconn 100 check port 8008
     server postgresql_127.0.0.1_5433 127.0.0.1:5433 maxconn 100 check port 8009

kubernetes/Dockerfile

@@ -0,0 +1,33 @@
+FROM postgres:15
+LABEL maintainer="Alexander Kukushkin <[email protected]>"
+
+RUN export DEBIAN_FRONTEND=noninteractive \
+    && echo 'APT::Install-Recommends "0";\nAPT::Install-Suggests "0";' > /etc/apt/apt.conf.d/01norecommend \
+    && apt-get update -y \
+    && apt-cache depends patroni | sed -n -e 's/.* Depends: \(python3-.\+\)$/\1/p' \
+            | grep -Ev '^python3-(sphinx|etcd|consul|kazoo|kubernetes)' \
+            | xargs apt-get install -y vim-tiny curl jq locales git python3-pip python3-wheel \
+    ## Make sure we have a en_US.UTF-8 locale available
+    && localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8 \
+    && pip3 install setuptools \
+    && pip3 install 'git+https://github.com/zalando/patroni.git#egg=patroni[kubernetes]' \
+    && PGHOME=/home/postgres \
+    && mkdir -p $PGHOME \
+    && chown postgres $PGHOME \
+    && sed -i "s|/var/lib/postgresql.*|$PGHOME:/bin/bash|" /etc/passwd \
+    # Set permissions for OpenShift
+    && chmod 775 $PGHOME \
+    && chmod 664 /etc/passwd \
+    # Clean up
+    && apt-get remove -y git python3-pip python3-wheel \
+    && apt-get autoremove -y \
+    && apt-get clean -y \
+    && rm -rf /var/lib/apt/lists/* /root/.cache
+
+COPY entrypoint.sh /
+
+EXPOSE 5432 8008
+ENV LC_ALL=en_US.UTF-8 LANG=en_US.UTF-8 EDITOR=/usr/bin/editor
+USER postgres
+WORKDIR /home/postgres
+CMD ["/bin/bash", "/entrypoint.sh"]

kubernetes/Dockerfile.citus

@@ -0,0 +1,42 @@
+FROM postgres:15
+LABEL maintainer="Alexander Kukushkin <[email protected]>"
+
+RUN export DEBIAN_FRONTEND=noninteractive \
+    && echo 'APT::Install-Recommends "0";\nAPT::Install-Suggests "0";' > /etc/apt/apt.conf.d/01norecommend \
+    && apt-get update -y \
+    && apt-get upgrade -y \
+    && apt-cache depends patroni | sed -n -e 's/.* Depends: \(python3-.\+\)$/\1/p' \
+            | grep -Ev '^python3-(sphinx|etcd|consul|kazoo|kubernetes)' \
+            | xargs apt-get install -y busybox vim-tiny curl jq less locales git python3-pip python3-wheel \
+    ## Make sure we have a en_US.UTF-8 locale available
+    && localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8 \
+    && curl https://install.citusdata.com/community/deb.sh | bash \
+    && apt-get -y install postgresql-15-citus-11.2 \
+    && pip3 install setuptools \
+    && pip3 install 'git+https://github.com/zalando/patroni.git#egg=patroni[kubernetes]' \
+    && PGHOME=/home/postgres \
+    && mkdir -p $PGHOME \
+    && chown postgres $PGHOME \
+    && sed -i "s|/var/lib/postgresql.*|$PGHOME:/bin/bash|" /etc/passwd \
+    && /bin/busybox --install -s \
+    # Set permissions for OpenShift
+    && chmod 775 $PGHOME \
+    && chmod 664 /etc/passwd \
+    # Clean up
+    && apt-get remove -y git python3-pip python3-wheel \
+    && apt-get autoremove -y \
+    && apt-get clean -y \
+    && rm -rf /var/lib/apt/lists/* /root/.cache
+
+ADD entrypoint.sh /
+ENV PGSSLMODE=verify-ca PGSSLKEY=/etc/ssl/private/ssl-cert-snakeoil.key PGSSLCERT=/etc/ssl/certs/ssl-cert-snakeoil.pem PGSSLROOTCERT=/etc/ssl/certs/ssl-cert-snakeoil.pem
+
+RUN sed -i 's/^postgresql:/&\n  basebackup:\n    checkpoint: fast/' /entrypoint.sh \
+    && sed -i "s|^    postgresql:|&\n      pg_hba:\n      - local all all trust\n      - hostssl replication all all md5 clientcert=$PGSSLMODE\n      - hostssl all all all md5 clientcert=$PGSSLMODE\n      parameters:\n        max_connections: 100\n        shared_buffers: 16MB\n        ssl: 'on'\n        ssl_ca_file: $PGSSLROOTCERT\n        ssl_cert_file: $PGSSLCERT\n        ssl_key_file: $PGSSLKEY\n        citus.node_conninfo: 'sslrootcert=$PGSSLROOTCERT sslkey=$PGSSLKEY sslcert=$PGSSLCERT sslmode=$PGSSLMODE'|"  /entrypoint.sh \
+    && sed -i "s#^    \(superuser\|replication\):#&\n      sslmode: $PGSSLMODE\n      sslkey: $PGSSLKEY\n      sslcert: $PGSSLCERT\n      sslrootcert: $PGSSLROOTCERT#" /entrypoint.sh
+
+EXPOSE 5432 8008
+ENV LC_ALL=en_US.UTF-8 LANG=en_US.UTF-8 EDITOR=/usr/bin/editor
+USER postgres
+WORKDIR /home/postgres
+CMD ["/bin/bash", "/entrypoint.sh"]

kubernetes/README.md

@@ -0,0 +1,154 @@
+# Kubernetes deployment examples
+Below you will find examples of Patroni deployments using [kind](https://kind.sigs.k8s.io/).
+
+# Patroni on K8s
+The Patroni cluster deployment with a StatefulSet consisting of three Pods.
+
+Example session:
+
+    $ kind create cluster
+    Creating cluster "kind" ...
+     ✓ Ensuring node image (kindest/node:v1.25.3) 🖼
+     ✓ Preparing nodes 📦
+     ✓ Writing configuration 📜
+     ✓ Starting control-plane 🕹️
+     ✓ Installing CNI 🔌
+     ✓ Installing StorageClass 💾
+    Set kubectl context to "kind-kind"
+    You can now use your cluster with:
+
+    kubectl cluster-info --context kind-kind
+
+    Thanks for using kind! 😊
+
+    $ docker build -t patroni .
+    Sending build context to Docker daemon  138.8kB
+    Step 1/9 : FROM postgres:15
+    ...
+    Successfully built e9bfe69c5d2b
+    Successfully tagged patroni:latest
+
+    $ kind load docker-image patroni
+    Image: "" with ID "sha256:e9bfe69c5d2b319dec0cf564fb895484537664775e18f37f9b707914cc5537e6" not yet present on node "kind-control-plane", loading...
+
+    $ kubectl apply -f patroni_k8s.yaml
+    service/patronidemo-config created
+    statefulset.apps/patronidemo created
+    endpoints/patronidemo created
+    service/patronidemo created
+    service/patronidemo-repl created
+    secret/patronidemo created
+    serviceaccount/patronidemo created
+    role.rbac.authorization.k8s.io/patronidemo created
+    rolebinding.rbac.authorization.k8s.io/patronidemo created
+    clusterrole.rbac.authorization.k8s.io/patroni-k8s-ep-access created
+    clusterrolebinding.rbac.authorization.k8s.io/patroni-k8s-ep-access created
+
+    $ kubectl get pods -L role
+    NAME            READY   STATUS    RESTARTS   AGE   ROLE
+    patronidemo-0   1/1     Running   0          34s   master
+    patronidemo-1   1/1     Running   0          30s   replica
+    patronidemo-2   1/1     Running   0          26s   replica
+
+    $ kubectl exec -ti patronidemo-0 -- bash
+    postgres@patronidemo-0:~$ patronictl list
+    + Cluster: patronidemo (7186662553319358497) ----+----+-----------+
+    | Member        | Host       | Role    | State   | TL | Lag in MB |
+    +---------------+------------+---------+---------+----+-----------+
+    | patronidemo-0 | 10.244.0.5 | Leader  | running |  1 |           |
+    | patronidemo-1 | 10.244.0.6 | Replica | running |  1 |         0 |
+    | patronidemo-2 | 10.244.0.7 | Replica | running |  1 |         0 |
+    +---------------+------------+---------+---------+----+-----------+
+
+# Citus on K8s
+The Citus cluster with the StatefulSets, one coordinator with three Pods and two workers with two pods each.
+
+Example session:
+
+    $ kind create cluster
+    Creating cluster "kind" ...
+     ✓ Ensuring node image (kindest/node:v1.25.3) 🖼
+     ✓ Preparing nodes 📦
+     ✓ Writing configuration 📜
+     ✓ Starting control-plane 🕹️
+     ✓ Installing CNI 🔌
+     ✓ Installing StorageClass 💾
+    Set kubectl context to "kind-kind"
+    You can now use your cluster with:
+
+    kubectl cluster-info --context kind-kind
+
+    Thanks for using kind! 😊
+
+    demo@localhost:~/git/patroni/kubernetes$ docker build -f Dockerfile.citus -t patroni-citus-k8s .
+    Sending build context to Docker daemon  138.8kB
+    Step 1/11 : FROM postgres:15
+    ...
+    Successfully built 8cd73e325028
+    Successfully tagged patroni-citus-k8s:latest
+
+    $ kind load docker-image patroni-citus-k8s
+    Image: "" with ID "sha256:8cd73e325028d7147672494965e53453f5540400928caac0305015eb2c7027c7" not yet present on node "kind-control-plane", loading...
+
+    $ kubectl apply -f citus_k8s.yaml
+    service/citusdemo-0-config created
+    service/citusdemo-1-config created
+    service/citusdemo-2-config created
+    statefulset.apps/citusdemo-0 created
+    statefulset.apps/citusdemo-1 created
+    statefulset.apps/citusdemo-2 created
+    endpoints/citusdemo-0 created
+    service/citusdemo-0 created
+    endpoints/citusdemo-1 created
+    service/citusdemo-1 created
+    endpoints/citusdemo-2 created
+    service/citusdemo-2 created
+    service/citusdemo-workers created
+    secret/citusdemo created
+    serviceaccount/citusdemo created
+    role.rbac.authorization.k8s.io/citusdemo created
+    rolebinding.rbac.authorization.k8s.io/citusdemo created
+    clusterrole.rbac.authorization.k8s.io/patroni-k8s-ep-access created
+    clusterrolebinding.rbac.authorization.k8s.io/patroni-k8s-ep-access created
+
+    $ kubectl get sts
+    NAME          READY   AGE
+    citusdemo-0   1/3     6s  # coodinator (group=0)
+    citusdemo-1   1/2     6s  # worker (group=1)
+    citusdemo-2   1/2     6s  # worker (group=2)
+
+    $ kubectl get pods -l cluster-name=citusdemo -L role
+    NAME            READY   STATUS    RESTARTS   AGE    ROLE
+    citusdemo-0-0   1/1     Running   0          105s   master
+    citusdemo-0-1   1/1     Running   0          101s   replica
+    citusdemo-0-2   1/1     Running   0          96s    replica
+    citusdemo-1-0   1/1     Running   0          105s   master
+    citusdemo-1-1   1/1     Running   0          101s   replica
+    citusdemo-2-0   1/1     Running   0          105s   master
+    citusdemo-2-1   1/1     Running   0          101s   replica
+
+    $ kubectl exec -ti citusdemo-0-0 -- bash
+    postgres@citusdemo-0-0:~$ patronictl list
+    + Citus cluster: citusdemo -----------+--------------+---------+----+-----------+
+    | Group | Member        | Host        | Role         | State   | TL | Lag in MB |
+    +-------+---------------+-------------+--------------+---------+----+-----------+
+    |     0 | citusdemo-0-0 | 10.244.0.10 | Leader       | running |  1 |           |
+    |     0 | citusdemo-0-1 | 10.244.0.12 | Replica      | running |  1 |         0 |
+    |     0 | citusdemo-0-2 | 10.244.0.14 | Sync Standby | running |  1 |         0 |
+    |     1 | citusdemo-1-0 | 10.244.0.8  | Leader       | running |  1 |           |
+    |     1 | citusdemo-1-1 | 10.244.0.11 | Sync Standby | running |  1 |         0 |
+    |     2 | citusdemo-2-0 | 10.244.0.9  | Leader       | running |  1 |           |
+    |     2 | citusdemo-2-1 | 10.244.0.13 | Sync Standby | running |  1 |         0 |
+    +-------+---------------+-------------+--------------+---------+----+-----------+
+
+    postgres@citusdemo-0-0:~$ psql citus
+    psql (15.1 (Debian 15.1-1.pgdg110+1))
+    Type "help" for help.
+
+    citus=# table pg_dist_node;
+     nodeid | groupid |  nodename   | nodeport | noderack | hasmetadata | isactive | noderole | nodecluster | metadatasynced | shouldhaveshards
+    --------+---------+-------------+----------+----------+-------------+----------+----------+-------------+----------------+------------------
+          1 |       0 | 10.244.0.10 |     5432 | default  | t           | t        | primary  | default     | t              | f
+          2 |       1 | 10.244.0.8  |     5432 | default  | t           | t        | primary  | default     | t              | t
+          3 |       2 | 10.244.0.9  |     5432 | default  | t           | t        | primary  | default     | t              | t
+    (3 rows)

kubernetes/citus_k8s.yaml

@@ -0,0 +1,590 @@
+# headless services to avoid deletion of citusdemo-*-config endpoints
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-0-config
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '0'
+spec:
+  clusterIP: None
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-1-config
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '1'
+spec:
+  clusterIP: None
+
+---
+
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-2-config
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '2'
+spec:
+  clusterIP: None
+
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: &cluster_name citusdemo-0
+  labels: &labels
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '0'
+    citus-type: coordinator
+spec:
+  replicas: 3
+  serviceName: *cluster_name
+  selector:
+    matchLabels:
+      <<: *labels
+  template:
+    metadata:
+      labels:
+        <<: *labels
+    spec:
+      serviceAccountName: citusdemo
+      containers:
+      - name: *cluster_name
+        image: patroni-citus-k8s  # docker build -f Dockerfile.citus -t patroni-citus-k8s .
+        imagePullPolicy: IfNotPresent
+        readinessProbe:
+          httpGet:
+            scheme: HTTP
+            path: /readiness
+            port: 8008
+          initialDelaySeconds: 3
+          periodSeconds: 10
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 3
+        ports:
+        - containerPort: 8008
+          protocol: TCP
+        - containerPort: 5432
+          protocol: TCP
+        volumeMounts:
+        - mountPath: /home/postgres/pgdata
+          name: pgdata
+        env:
+        - name: PATRONI_KUBERNETES_POD_IP
+          valueFrom:
+            fieldRef:
+              fieldPath: status.podIP
+        - name: PATRONI_KUBERNETES_NAMESPACE
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.namespace
+        - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+          value: 'true'
+        - name: PATRONI_KUBERNETES_USE_ENDPOINTS
+          value: 'true'
+        - name: PATRONI_KUBERNETES_LABELS
+          value: '{application: patroni, cluster-name: citusdemo}'
+        - name: PATRONI_CITUS_DATABASE
+          value: citus
+        - name: PATRONI_CITUS_GROUP
+          value: '0'
+        - name: PATRONI_SUPERUSER_USERNAME
+          value: postgres
+        - name: PATRONI_SUPERUSER_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: superuser-password
+        - name: PATRONI_REPLICATION_USERNAME
+          value: standby
+        - name: PATRONI_REPLICATION_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: replication-password
+        - name: PATRONI_SCOPE
+          value: citusdemo
+        - name: PATRONI_NAME
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.name
+        - name: PATRONI_POSTGRESQL_DATA_DIR
+          value: /home/postgres/pgdata/pgroot/data
+        - name: PATRONI_POSTGRESQL_PGPASS
+          value: /tmp/pgpass
+        - name: PATRONI_POSTGRESQL_LISTEN
+          value: '0.0.0.0:5432'
+        - name: PATRONI_RESTAPI_LISTEN
+          value: '0.0.0.0:8008'
+      terminationGracePeriodSeconds: 0
+      volumes:
+      - name: pgdata
+        emptyDir: {}
+#  volumeClaimTemplates:
+#  - metadata:
+#      labels:
+#        application: spilo
+#        spilo-cluster: *cluster_name
+#      annotations:
+#        volume.alpha.kubernetes.io/storage-class: anything
+#      name: pgdata
+#    spec:
+#      accessModes:
+#      - ReadWriteOnce
+#      resources:
+#        requests:
+#          storage: 5Gi
+
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: &cluster_name citusdemo-1
+  labels: &labels
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '1'
+    citus-type: worker
+spec:
+  replicas: 2
+  serviceName: *cluster_name
+  selector:
+    matchLabels:
+      <<: *labels
+  template:
+    metadata:
+      labels:
+        <<: *labels
+    spec:
+      serviceAccountName: citusdemo
+      containers:
+      - name: *cluster_name
+        image: patroni-citus-k8s  # docker build -f Dockerfile.citus -t patroni-citus-k8s .
+        imagePullPolicy: IfNotPresent
+        readinessProbe:
+          httpGet:
+            scheme: HTTP
+            path: /readiness
+            port: 8008
+          initialDelaySeconds: 3
+          periodSeconds: 10
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 3
+        ports:
+        - containerPort: 8008
+          protocol: TCP
+        - containerPort: 5432
+          protocol: TCP
+        volumeMounts:
+        - mountPath: /home/postgres/pgdata
+          name: pgdata
+        env:
+        - name: PATRONI_KUBERNETES_POD_IP
+          valueFrom:
+            fieldRef:
+              fieldPath: status.podIP
+        - name: PATRONI_KUBERNETES_NAMESPACE
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.namespace
+        - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+          value: 'true'
+        - name: PATRONI_KUBERNETES_USE_ENDPOINTS
+          value: 'true'
+        - name: PATRONI_KUBERNETES_LABELS
+          value: '{application: patroni, cluster-name: citusdemo}'
+        - name: PATRONI_CITUS_DATABASE
+          value: citus
+        - name: PATRONI_CITUS_GROUP
+          value: '1'
+        - name: PATRONI_SUPERUSER_USERNAME
+          value: postgres
+        - name: PATRONI_SUPERUSER_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: superuser-password
+        - name: PATRONI_REPLICATION_USERNAME
+          value: standby
+        - name: PATRONI_REPLICATION_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: replication-password
+        - name: PATRONI_SCOPE
+          value: citusdemo
+        - name: PATRONI_NAME
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.name
+        - name: PATRONI_POSTGRESQL_DATA_DIR
+          value: /home/postgres/pgdata/pgroot/data
+        - name: PATRONI_POSTGRESQL_PGPASS
+          value: /tmp/pgpass
+        - name: PATRONI_POSTGRESQL_LISTEN
+          value: '0.0.0.0:5432'
+        - name: PATRONI_RESTAPI_LISTEN
+          value: '0.0.0.0:8008'
+      terminationGracePeriodSeconds: 0
+      volumes:
+      - name: pgdata
+        emptyDir: {}
+#  volumeClaimTemplates:
+#  - metadata:
+#      labels:
+#        application: spilo
+#        spilo-cluster: *cluster_name
+#      annotations:
+#        volume.alpha.kubernetes.io/storage-class: anything
+#      name: pgdata
+#    spec:
+#      accessModes:
+#      - ReadWriteOnce
+#      resources:
+#        requests:
+#          storage: 5Gi
+
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: &cluster_name citusdemo-2
+  labels: &labels
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '2'
+    citus-type: worker
+spec:
+  replicas: 2
+  serviceName: *cluster_name
+  selector:
+    matchLabels:
+      <<: *labels
+  template:
+    metadata:
+      labels:
+        <<: *labels
+    spec:
+      serviceAccountName: citusdemo
+      containers:
+      - name: *cluster_name
+        image: patroni-citus-k8s  # docker build -f Dockerfile.citus -t patroni-citus-k8s .
+        imagePullPolicy: IfNotPresent
+        readinessProbe:
+          httpGet:
+            scheme: HTTP
+            path: /readiness
+            port: 8008
+          initialDelaySeconds: 3
+          periodSeconds: 10
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 3
+        ports:
+        - containerPort: 8008
+          protocol: TCP
+        - containerPort: 5432
+          protocol: TCP
+        volumeMounts:
+        - mountPath: /home/postgres/pgdata
+          name: pgdata
+        env:
+        - name: PATRONI_KUBERNETES_POD_IP
+          valueFrom:
+            fieldRef:
+              fieldPath: status.podIP
+        - name: PATRONI_KUBERNETES_NAMESPACE
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.namespace
+        - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+          value: 'true'
+        - name: PATRONI_KUBERNETES_USE_ENDPOINTS
+          value: 'true'
+        - name: PATRONI_KUBERNETES_LABELS
+          value: '{application: patroni, cluster-name: citusdemo}'
+        - name: PATRONI_CITUS_DATABASE
+          value: citus
+        - name: PATRONI_CITUS_GROUP
+          value: '2'
+        - name: PATRONI_SUPERUSER_USERNAME
+          value: postgres
+        - name: PATRONI_SUPERUSER_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: superuser-password
+        - name: PATRONI_REPLICATION_USERNAME
+          value: standby
+        - name: PATRONI_REPLICATION_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: citusdemo
+              key: replication-password
+        - name: PATRONI_SCOPE
+          value: citusdemo
+        - name: PATRONI_NAME
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.name
+        - name: PATRONI_POSTGRESQL_DATA_DIR
+          value: /home/postgres/pgdata/pgroot/data
+        - name: PATRONI_POSTGRESQL_PGPASS
+          value: /tmp/pgpass
+        - name: PATRONI_POSTGRESQL_LISTEN
+          value: '0.0.0.0:5432'
+        - name: PATRONI_RESTAPI_LISTEN
+          value: '0.0.0.0:8008'
+      terminationGracePeriodSeconds: 0
+      volumes:
+      - name: pgdata
+        emptyDir: {}
+#  volumeClaimTemplates:
+#  - metadata:
+#      labels:
+#        application: spilo
+#        spilo-cluster: *cluster_name
+#      annotations:
+#        volume.alpha.kubernetes.io/storage-class: anything
+#      name: pgdata
+#    spec:
+#      accessModes:
+#      - ReadWriteOnce
+#      resources:
+#        requests:
+#          storage: 5Gi
+          
+---
+apiVersion: v1
+kind: Endpoints
+metadata:
+  name: citusdemo-0
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '0'
+    citus-type: coordinator
+subsets: []
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-0
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '0'
+    citus-type: coordinator
+spec:
+  type: ClusterIP
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Endpoints
+metadata:
+  name: citusdemo-1
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '1'
+    citus-type: worker
+subsets: []
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-1
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '1'
+    citus-type: worker
+spec:
+  type: ClusterIP
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Endpoints
+metadata:
+  name: citusdemo-2
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '2'
+    citus-type: worker
+subsets: []
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-2
+  labels:
+    application: patroni
+    cluster-name: citusdemo
+    citus-group: '2'
+    citus-type: worker
+spec:
+  type: ClusterIP
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: citusdemo-workers
+  labels: &labels
+    application: patroni
+    cluster-name: citusdemo
+    citus-type: worker
+    role: master
+spec:
+  type: ClusterIP
+  selector:
+    <<: *labels
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  name: &cluster_name citusdemo
+  labels:
+    application: patroni
+    cluster-name: *cluster_name
+type: Opaque
+data:
+  superuser-password: emFsYW5kbw==
+  replication-password: cmVwLXBhc3M=
+
+---
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: citusdemo
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: Role
+metadata:
+  name: citusdemo
+rules:
+- apiGroups:
+  - ""
+  resources:
+  - configmaps
+  verbs:
+  - create
+  - get
+  - list
+  - patch
+  - update
+  - watch
+  # delete and deletecollection are required only for 'patronictl remove'
+  - delete
+  - deletecollection
+- apiGroups:
+  - ""
+  resources:
+  - endpoints
+  verbs:
+  - get
+  - patch
+  - update
+  # the following three privileges are necessary only when using endpoints
+  - create
+  - list
+  - watch
+  # delete and deletecollection are required only for for 'patronictl remove'
+  - delete
+  - deletecollection
+- apiGroups:
+  - ""
+  resources:
+  - pods
+  verbs:
+  - get
+  - list
+  - patch
+  - update
+  - watch
+# The following privilege is only necessary for creation of headless service
+# for citusdemo-config endpoint, in order to prevent cleaning it up by the
+# k8s master. You can avoid giving this privilege by explicitly creating the
+# service like it is done in this manifest (lines 2..10)
+- apiGroups:
+  - ""
+  resources:
+  - services
+  verbs:
+  - create
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: RoleBinding
+metadata:
+  name: citusdemo
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: Role
+  name: citusdemo
+subjects:
+- kind: ServiceAccount
+  name: citusdemo
+
+# Following privileges are only required if deployed not in the "default"
+# namespace and you want Patroni to bypass kubernetes service
+# (PATRONI_KUBERNETES_BYPASS_API_SERVICE=true)
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: patroni-k8s-ep-access
+rules:
+- apiGroups:
+  - ""
+  resources:
+  - endpoints
+  resourceNames:
+  - kubernetes
+  verbs:
+  - get
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: patroni-k8s-ep-access
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: patroni-k8s-ep-access
+subjects:
+- kind: ServiceAccount
+  name: citusdemo
+# The namespace must be specified explicitly.
+# If deploying to the different namespace you have to change it.
+  namespace: default

kubernetes/entrypoint.sh

@@ -0,0 +1,37 @@
+#!/bin/bash
+
+if [[ $UID -ge 10000 ]]; then
+    GID=$(id -g)
+    sed -e "s/^postgres:x:[^:]*:[^:]*:/postgres:x:$UID:$GID:/" /etc/passwd > /tmp/passwd
+    cat /tmp/passwd > /etc/passwd
+    rm /tmp/passwd
+fi
+
+cat > /home/postgres/patroni.yml <<__EOF__
+bootstrap:
+  dcs:
+    postgresql:
+      use_pg_rewind: true
+  initdb:
+  - auth-host: md5
+  - auth-local: trust
+  - encoding: UTF8
+  - locale: en_US.UTF-8
+  - data-checksums
+  pg_hba:
+  - host all all 0.0.0.0/0 md5
+  - host replication ${PATRONI_REPLICATION_USERNAME} ${PATRONI_KUBERNETES_POD_IP}/16 md5
+restapi:
+  connect_address: '${PATRONI_KUBERNETES_POD_IP}:8008'
+postgresql:
+  connect_address: '${PATRONI_KUBERNETES_POD_IP}:5432'
+  authentication:
+    superuser:
+      password: '${PATRONI_SUPERUSER_PASSWORD}'
+    replication:
+      password: '${PATRONI_REPLICATION_PASSWORD}'
+__EOF__
+
+unset PATRONI_SUPERUSER_PASSWORD PATRONI_REPLICATION_PASSWORD
+
+exec /usr/bin/python3 /usr/local/bin/patroni /home/postgres/patroni.yml

kubernetes/openshift-example/README.md

@@ -0,0 +1,49 @@
+# Patroni OpenShift Configuration
+Patroni can be run in OpenShift. Based on the kubernetes configuration, the Dockerfile and Entrypoint has been modified to support the dynamic UID/GID configuration that is applied in OpenShift. This can be run under the standard `restricted` SCC.
+
+# Examples
+
+## Create test project
+
+```
+oc new-project patroni-test
+```
+
+## Build the image
+
+Note: Update the references when merged upstream.
+Note: If deploying as a template for multiple users, the following commands should be performed in a shared namespace like `openshift`.
+
+```
+oc import-image postgres:10 --confirm -n openshift
+oc new-build https://github.com/zalando/patroni --context-dir=kubernetes -n openshift
+```
+
+## Deploy the Image
+Two configuration templates exist in [templates](templates) directory:
+-   Patroni Ephemeral
+-   Patroni Persistent
+
+The only difference is whether or not the statefulset requests persistent storage.
+
+## Create the Template
+Install the template into the `openshift` namespace if this should be shared across projects:
+
+```
+oc create -f templates/template_patroni_ephemeral.yml -n openshift
+```
+
+Then, from your own project:
+
+```
+oc new-app patroni-pgsql-ephemeral
+```
+
+Once the pods are running, two configmaps should be available:
+
+```
+$ oc get configmap
+NAME                DATA      AGE
+patroniocp-config   0         1m
+patroniocp-leader   0         1m
+```

kubernetes/openshift-example/templates/template_patroni_ephemeral.yml

@@ -0,0 +1,327 @@
+apiVersion: v1
+kind: Template
+metadata:
+  name: patroni-pgsql-ephemeral
+  annotations:
+    description: |-
+      Patroni Postgresql database cluster, without persistent storage.
+
+      WARNING: Any data stored will be lost upon pod destruction. Only use this template for testing.
+    iconClass: icon-postgresql
+    openshift.io/display-name: Patroni Postgresql (Ephemeral)
+    openshift.io/long-description: This template deploys a a patroni postgresql HA cluster without persistent storage.
+    tags: postgresql
+objects:
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_CLUSTER_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_MASTER_SERVICE_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    selector:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+      role: master
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: v1
+  kind: Secret
+  metadata:
+    name: ${PATRONI_CLUSTER_NAME}
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+  stringData:
+    superuser-password: ${PATRONI_SUPERUSER_PASSWORD}
+    replication-password: ${PATRONI_REPLICATION_PASSWORD}
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_REPLICA_SERVICE_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    selector:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+      role: replica
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: apps/v1
+  kind: StatefulSet
+  metadata:
+    creationTimestamp: null
+    generation: 3
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${APPLICATION_NAME}
+  spec:
+    podManagementPolicy: OrderedReady
+    replicas: 3
+    revisionHistoryLimit: 10
+    selector:
+      matchLabels:
+        application: ${APPLICATION_NAME}
+        cluster-name: ${PATRONI_CLUSTER_NAME}
+    serviceName: ${APPLICATION_NAME}
+    template:
+      metadata:
+        creationTimestamp: null
+        labels:
+          application: ${APPLICATION_NAME}
+          cluster-name: ${PATRONI_CLUSTER_NAME}
+      spec:
+        containers:
+        - env:
+          - name: PATRONI_KUBERNETES_POD_IP
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: status.podIP
+          - name: PATRONI_KUBERNETES_NAMESPACE
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.namespace
+          - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+            value: 'true'
+          - name: PATRONI_KUBERNETES_LABELS
+            value: '{application: ${APPLICATION_NAME}, cluster-name: ${PATRONI_CLUSTER_NAME}}'
+          - name: PATRONI_SUPERUSER_USERNAME
+            value: ${PATRONI_SUPERUSER_USERNAME}
+          - name: PATRONI_SUPERUSER_PASSWORD
+            valueFrom:
+              secretKeyRef:
+                key: superuser-password
+                name: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_REPLICATION_USERNAME
+            value: ${PATRONI_REPLICATION_USERNAME}
+          - name: PATRONI_REPLICATION_PASSWORD
+            valueFrom:
+              secretKeyRef:
+                key: replication-password
+                name: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_SCOPE
+            value: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_NAME
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.name
+          - name: PATRONI_POSTGRESQL_DATA_DIR
+            value: /home/postgres/pgdata/pgroot/data
+          - name: PATRONI_POSTGRESQL_PGPASS
+            value: /tmp/pgpass
+          - name: PATRONI_POSTGRESQL_LISTEN
+            value: 0.0.0.0:5432
+          - name: PATRONI_RESTAPI_LISTEN
+            value: 0.0.0.0:8008
+          image: docker-registry.default.svc:5000/${NAMESPACE}/patroni:latest
+          imagePullPolicy: IfNotPresent
+          name: ${APPLICATION_NAME}
+          readinessProbe:
+            httpGet:
+              scheme: HTTP
+              path: /readiness
+              port: 8008
+            initialDelaySeconds: 3
+            periodSeconds: 10
+            timeoutSeconds: 5
+            successThreshold: 1
+            failureThreshold: 3
+          ports:
+          - containerPort: 8008
+            protocol: TCP
+          - containerPort: 5432
+            protocol: TCP
+          resources: {}
+          terminationMessagePath: /dev/termination-log
+          terminationMessagePolicy: File
+          volumeMounts:
+          - mountPath: /home/postgres/pgdata
+            name: pgdata
+        dnsPolicy: ClusterFirst
+        restartPolicy: Always
+        schedulerName: default-scheduler
+        securityContext: {}
+        serviceAccount: ${SERVICE_ACCOUNT}
+        serviceAccountName: ${SERVICE_ACCOUNT}
+        terminationGracePeriodSeconds: 0
+        volumes:
+        - name: pgdata
+          emptyDir: {}
+    updateStrategy:
+      type: OnDelete
+- apiVersion: v1
+  kind: Endpoints
+  metadata:
+    name: ${APPLICATION_NAME}
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+  subsets: []
+- apiVersion: v1
+  kind: ServiceAccount
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: Role
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+  rules:
+  - apiGroups:
+    - ""
+    resources:
+    - configmaps
+    verbs:
+    - create
+    - get
+    - list
+    - patch
+    - update
+    - watch
+    # delete is required only for 'patronictl remove'
+    - delete
+  - apiGroups:
+    - ""
+    resources:
+    - endpoints
+    verbs:
+    - get
+    - patch
+    - update
+    # the following three privileges are necessary only when using endpoints
+    - create
+    - list
+    - watch
+    # delete is required only for for 'patronictl remove'
+    - delete
+  - apiGroups:
+    - ""
+    resources:
+    - pods
+    verbs:
+    - get
+    - list
+    - patch
+    - update
+    - watch
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: RoleBinding
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+  roleRef:
+    apiGroup: rbac.authorization.k8s.io
+    kind: Role
+    name: ${SERVICE_ACCOUNT}
+  subjects:
+  - kind: ServiceAccount
+    name: ${SERVICE_ACCOUNT}
+# Following privileges are only required if deployed not in the "default"
+# namespace and you want Patroni to bypass kubernetes service
+# (PATRONI_KUBERNETES_BYPASS_API_SERVICE=true)
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: ClusterRole
+  metadata:
+    name: patroni-k8s-ep-access
+  rules:
+  - apiGroups:
+    - ""
+    resources:
+    - endpoints
+    resourceNames:
+    - kubernetes
+    verbs:
+    - get
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: ClusterRoleBinding
+  metadata:
+    name: ${NAMESPACE}-${SERVICE_ACCOUNT}-k8s-ep-access
+  roleRef:
+    apiGroup: rbac.authorization.k8s.io
+    kind: ClusterRole
+    name: patroni-k8s-ep-access
+  subjects:
+  - kind: ServiceAccount
+    name: ${SERVICE_ACCOUNT}
+    namespace: ${NAMESPACE}
+parameters:
+- description: The name of the application for labelling all artifacts.
+  displayName: Application Name
+  name: APPLICATION_NAME
+  value: patroni-ephemeral
+- description: The name of the patroni-pgsql cluster.
+  displayName: Cluster Name
+  name: PATRONI_CLUSTER_NAME
+  value: patroni-ephemeral
+- description: The name of the OpenShift Service exposed for the patroni-ephemeral-master container.
+  displayName: Master service name.  
+  name: PATRONI_MASTER_SERVICE_NAME
+  value: patroni-ephemeral-master
+- description: The name of the OpenShift Service exposed for the patroni-ephemeral-replica containers.
+  displayName: Replica service name. 
+  name: PATRONI_REPLICA_SERVICE_NAME
+  value: patroni-ephemeral-replica
+- description: Maximum amount of memory the container can use.
+  displayName: Memory Limit
+  name: MEMORY_LIMIT
+  value: 512Mi
+- description: The OpenShift Namespace where the patroni and postgresql ImageStream resides.
+  displayName: ImageStream Namespace
+  name: NAMESPACE
+  value: openshift
+- description: Username of the superuser account for initialization.
+  displayName: Superuser Username
+  name: PATRONI_SUPERUSER_USERNAME
+  value: postgres
+- description: Password of the superuser account for initialization.
+  displayName: Superuser Passsword
+  name: PATRONI_SUPERUSER_PASSWORD
+  value: postgres
+- description: Username of the replication account for initialization.
+  displayName: Replication Username
+  name: PATRONI_REPLICATION_USERNAME
+  value: postgres
+- description: Password of the replication account for initialization.
+  displayName: Repication Passsword
+  name: PATRONI_REPLICATION_PASSWORD
+  value: postgres
+- description: Service account name used for pods and rolebindings to form a cluster in the project. 
+  displayName: Service Account
+  name: SERVICE_ACCOUNT
+  value: patroniocp

kubernetes/openshift-example/templates/template_patroni_persistent.yaml

@@ -0,0 +1,355 @@
+apiVersion: v1
+kind: Template
+metadata:
+  name: patroni-pgsql-persistent
+  annotations:
+    description: |-
+      Patroni Postgresql database cluster, with persistent storage.
+    iconClass: icon-postgresql
+    openshift.io/display-name: Patroni Postgresql (Persistent)
+    openshift.io/long-description: This template deploys a a patroni postgresql HA cluster with persistent storage.
+    tags: postgresql
+objects:
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_CLUSTER_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_MASTER_SERVICE_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    selector:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+      role: master
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: v1
+  kind: Secret
+  metadata:
+    name: ${PATRONI_CLUSTER_NAME}
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+  stringData:
+    superuser-password: ${PATRONI_SUPERUSER_PASSWORD}
+    replication-password: ${PATRONI_REPLICATION_PASSWORD}
+- apiVersion: v1
+  kind: Service
+  metadata:
+    creationTimestamp: null
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${PATRONI_REPLICA_SERVICE_NAME}
+  spec:
+    ports:
+    - port: 5432
+      protocol: TCP
+      targetPort: 5432
+    selector:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+      role: replica
+    sessionAffinity: None
+    type: ClusterIP
+  status:
+    loadBalancer: {}
+- apiVersion: apps/v1
+  kind: StatefulSet
+  metadata:
+    creationTimestamp: null
+    generation: 3
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+    name: ${APPLICATION_NAME}
+  spec:
+    podManagementPolicy: OrderedReady
+    replicas: 3
+    revisionHistoryLimit: 10
+    selector:
+      matchLabels:
+        application: ${APPLICATION_NAME}
+        cluster-name: ${PATRONI_CLUSTER_NAME}
+    serviceName: ${APPLICATION_NAME}
+    template:
+      metadata:
+        creationTimestamp: null
+        labels:
+          application: ${APPLICATION_NAME}
+          cluster-name: ${PATRONI_CLUSTER_NAME}
+      spec:
+        initContainers:
+        - command:
+          - sh
+          - -c 
+          - "mkdir -p /home/postgres/pgdata/pgroot/data && chmod 0700 /home/postgres/pgdata/pgroot/data"
+          image: docker-registry.default.svc:5000/${NAMESPACE}/patroni:latest
+          imagePullPolicy: IfNotPresent
+          name: fix-perms
+          resources: {}
+          terminationMessagePath: /dev/termination-log
+          terminationMessagePolicy: File
+          volumeMounts:
+          - mountPath: /home/postgres/pgdata
+            name: ${APPLICATION_NAME}
+        containers:
+        - env:
+          - name: PATRONI_KUBERNETES_POD_IP
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: status.podIP
+          - name: PATRONI_KUBERNETES_NAMESPACE
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.namespace
+          - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+            value: 'true'
+          - name: PATRONI_KUBERNETES_LABELS
+            value: '{application: ${APPLICATION_NAME}, cluster-name: ${PATRONI_CLUSTER_NAME}}'
+          - name: PATRONI_SUPERUSER_USERNAME
+            value: ${PATRONI_SUPERUSER_USERNAME}
+          - name: PATRONI_SUPERUSER_PASSWORD
+            valueFrom:
+              secretKeyRef:
+                key: superuser-password
+                name: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_REPLICATION_USERNAME
+            value: ${PATRONI_REPLICATION_USERNAME}
+          - name: PATRONI_REPLICATION_PASSWORD
+            valueFrom:
+              secretKeyRef:
+                key: replication-password
+                name: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_SCOPE
+            value: ${PATRONI_CLUSTER_NAME}
+          - name: PATRONI_NAME
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.name
+          - name: PATRONI_POSTGRESQL_DATA_DIR
+            value: /home/postgres/pgdata/pgroot/data
+          - name: PATRONI_POSTGRESQL_PGPASS
+            value: /tmp/pgpass
+          - name: PATRONI_POSTGRESQL_LISTEN
+            value: 0.0.0.0:5432
+          - name: PATRONI_RESTAPI_LISTEN
+            value: 0.0.0.0:8008
+          image: docker-registry.default.svc:5000/${NAMESPACE}/patroni:latest
+          imagePullPolicy: IfNotPresent
+          name: ${APPLICATION_NAME}
+          readinessProbe:
+            httpGet:
+              scheme: HTTP
+              path: /readiness
+              port: 8008
+            initialDelaySeconds: 3
+            periodSeconds: 10
+            timeoutSeconds: 5
+            successThreshold: 1
+            failureThreshold: 3
+          ports:
+          - containerPort: 8008
+            protocol: TCP
+          - containerPort: 5432
+            protocol: TCP
+          resources: {}
+          terminationMessagePath: /dev/termination-log
+          terminationMessagePolicy: File
+          volumeMounts:
+          - mountPath: /home/postgres/pgdata
+            name: ${APPLICATION_NAME}
+        dnsPolicy: ClusterFirst
+        restartPolicy: Always
+        schedulerName: default-scheduler
+        securityContext: {}
+        serviceAccount: ${SERVICE_ACCOUNT}
+        serviceAccountName: ${SERVICE_ACCOUNT}
+        terminationGracePeriodSeconds: 0
+        volumes:
+        - name: ${APPLICATION_NAME}
+          persistentVolumeClaim: 
+            claimName: ${APPLICATION_NAME}
+    volumeClaimTemplates:
+    - metadata:
+        labels:
+          application: ${APPLICATION_NAME}
+        name: ${APPLICATION_NAME}
+      spec:
+        accessModes:
+        - ReadWriteOnce
+        resources:
+          requests:
+            storage: ${PVC_SIZE}
+    updateStrategy:
+      type: OnDelete
+- apiVersion: v1
+  kind: Endpoints
+  metadata:
+    name: ${APPLICATION_NAME}
+    labels:
+      application: ${APPLICATION_NAME}
+      cluster-name: ${PATRONI_CLUSTER_NAME}
+  subsets: []
+- apiVersion: v1
+  kind: ServiceAccount
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: Role
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+  rules:
+  - apiGroups:
+    - ""
+    resources:
+    - configmaps
+    verbs:
+    - create
+    - get
+    - list
+    - patch
+    - update
+    - watch
+    # delete is required only for 'patronictl remove'
+    - delete
+  - apiGroups:
+    - ""
+    resources:
+    - endpoints
+    verbs:
+    - get
+    - patch
+    - update
+    # the following three privileges are necessary only when using endpoints
+    - create
+    - list
+    - watch
+    # delete is required only for for 'patronictl remove'
+    - delete
+  - apiGroups:
+    - ""
+    resources:
+    - pods
+    verbs:
+    - get
+    - list
+    - patch
+    - update
+    - watch
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: RoleBinding
+  metadata:
+    name: ${SERVICE_ACCOUNT}
+  roleRef:
+    apiGroup: rbac.authorization.k8s.io
+    kind: Role
+    name: ${SERVICE_ACCOUNT}
+  subjects:
+  - kind: ServiceAccount
+    name: ${SERVICE_ACCOUNT}
+# Following privileges are only required if deployed not in the "default"
+# namespace and you want Patroni to bypass kubernetes service
+# (PATRONI_KUBERNETES_BYPASS_API_SERVICE=true)
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: ClusterRole
+  metadata:
+    name: patroni-k8s-ep-access
+  rules:
+  - apiGroups:
+    - ""
+    resources:
+    - endpoints
+    resourceNames:
+    - kubernetes
+    verbs:
+    - get
+- apiVersion: rbac.authorization.k8s.io/v1
+  kind: ClusterRoleBinding
+  metadata:
+    name: ${NAMESPACE}-${SERVICE_ACCOUNT}-k8s-ep-access
+  roleRef:
+    apiGroup: rbac.authorization.k8s.io
+    kind: ClusterRole
+    name: patroni-k8s-ep-access
+  subjects:
+  - kind: ServiceAccount
+    name: ${SERVICE_ACCOUNT}
+    namespace: ${NAMESPACE}
+parameters:
+- description: The name of the application for labelling all artifacts.
+  displayName: Application Name
+  name: APPLICATION_NAME
+  value: patroni-persistent
+- description: The name of the patroni-pgsql cluster.
+  displayName: Cluster Name
+  name: PATRONI_CLUSTER_NAME
+  value: patroni-persistent
+- description: The name of the OpenShift Service exposed for the patroni-persistent-master container.
+  displayName: Master service name.  
+  name: PATRONI_MASTER_SERVICE_NAME
+  value: patroni-persistent-master
+- description: The name of the OpenShift Service exposed for the patroni-persistent-replica containers.
+  displayName: Replica service name. 
+  name: PATRONI_REPLICA_SERVICE_NAME
+  value: patroni-persistent-replica
+- description: Maximum amount of memory the container can use.
+  displayName: Memory Limit
+  name: MEMORY_LIMIT
+  value: 512Mi
+- description: The OpenShift Namespace where the patroni and postgresql ImageStream resides.
+  displayName: ImageStream Namespace
+  name: NAMESPACE
+  value: openshift
+- description: Username of the superuser account for initialization.
+  displayName: Superuser Username
+  name: PATRONI_SUPERUSER_USERNAME
+  value: postgres
+- description: Password of the superuser account for initialization.
+  displayName: Superuser Passsword
+  name: PATRONI_SUPERUSER_PASSWORD
+  value: postgres
+- description: Username of the replication account for initialization.
+  displayName: Replication Username
+  name: PATRONI_REPLICATION_USERNAME
+  value: postgres
+- description: Password of the replication account for initialization.
+  displayName: Repication Passsword
+  name: PATRONI_REPLICATION_PASSWORD
+  value: postgres
+- description: Service account name used for pods and rolebindings to form a cluster in the project. 
+  displayName: Service Account
+  name: SERVICE_ACCOUNT
+  value: patroni-persistent
+- description: The size of the persistent volume to create. 
+  displayName: Persistent Volume Size
+  name: PVC_SIZE
+  value: 5Gi

kubernetes/openshift-example/test/Jenkinsfile

@@ -0,0 +1,43 @@
+pipeline {
+  agent any
+  stages {
+    stage ('Deploy test pod'){
+      when {
+          expression {
+            openshift.withCluster() {
+              openshift.withProject() {
+                return !openshift.selector( "dc", "pgbench" ).exists()
+              }
+            }
+          }
+        }
+      steps {
+        script {
+            openshift.withCluster() {
+                openshift.withProject() {
+                  def pgbench = openshift.newApp( "https://github.com/stewartshea/docker-pgbench/", "--name=pgbench", "-e PGPASSWORD=postgres", "-e PGUSER=postgres", "-e PGHOST=patroni-persistent-master", "-e PGDATABASE=postgres", "-e TEST_CLIENT_COUNT=20", "-e TEST_DURATION=120" )
+                  def pgbenchdc = openshift.selector( "dc", "pgbench" )
+                  timeout(5) { 
+                    pgbenchdc.rollout().status()
+                    }
+                  }
+                }
+            }
+        }
+    }
+    stage ('Run benchmark Test'){
+      steps {
+          sh '''
+          oc exec $(oc get pods -l app=pgbench | grep Running | awk '{print $1}')  ./test.sh
+          '''
+      }
+    }
+    stage ('Clean up pgtest pod'){
+      steps {
+          sh '''
+          oc delete all -l app=pgbench
+          '''
+      }
+    }
+  }
+}

kubernetes/openshift-example/test/README.md

@@ -0,0 +1,2 @@
+# Jenkins Test
+This pipeline test will create a separate deployment config for a pgbench pod and execute a test against the patroni cluster. This is a sample and should be customized.

kubernetes/patroni_k8s.yaml

@@ -0,0 +1,281 @@
+# headless service to avoid deletion of patronidemo-config endpoint
+apiVersion: v1
+kind: Service
+metadata:
+  name: patronidemo-config
+  labels:
+    application: patroni
+    cluster-name: patronidemo
+spec:
+  clusterIP: None
+
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: &cluster_name patronidemo
+  labels:
+    application: patroni
+    cluster-name: *cluster_name
+spec:
+  replicas: 3
+  serviceName: *cluster_name
+  selector:
+    matchLabels:
+      application: patroni
+      cluster-name: *cluster_name
+  template:
+    metadata:
+      labels:
+        application: patroni
+        cluster-name: *cluster_name
+    spec:
+      serviceAccountName: patronidemo
+      containers:
+      - name: *cluster_name
+        image: patroni  # docker build -t patroni .
+        imagePullPolicy: IfNotPresent
+        readinessProbe:
+          httpGet:
+            scheme: HTTP
+            path: /readiness
+            port: 8008
+          initialDelaySeconds: 3
+          periodSeconds: 10
+          timeoutSeconds: 5
+          successThreshold: 1
+          failureThreshold: 3
+        ports:
+        - containerPort: 8008
+          protocol: TCP
+        - containerPort: 5432
+          protocol: TCP
+        volumeMounts:
+        - mountPath: /home/postgres/pgdata
+          name: pgdata
+        env:
+        - name: PATRONI_KUBERNETES_POD_IP
+          valueFrom:
+            fieldRef:
+              fieldPath: status.podIP
+        - name: PATRONI_KUBERNETES_NAMESPACE
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.namespace
+        - name: PATRONI_KUBERNETES_BYPASS_API_SERVICE
+          value: 'true'
+        - name: PATRONI_KUBERNETES_USE_ENDPOINTS
+          value: 'true'
+        - name: PATRONI_KUBERNETES_LABELS
+          value: '{application: patroni, cluster-name: patronidemo}'
+        - name: PATRONI_SUPERUSER_USERNAME
+          value: postgres
+        - name: PATRONI_SUPERUSER_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: *cluster_name
+              key: superuser-password
+        - name: PATRONI_REPLICATION_USERNAME
+          value: standby
+        - name: PATRONI_REPLICATION_PASSWORD
+          valueFrom:
+            secretKeyRef:
+              name: *cluster_name
+              key: replication-password
+        - name: PATRONI_SCOPE
+          value: *cluster_name
+        - name: PATRONI_NAME
+          valueFrom:
+            fieldRef:
+              fieldPath: metadata.name
+        - name: PATRONI_POSTGRESQL_DATA_DIR
+          value: /home/postgres/pgdata/pgroot/data
+        - name: PATRONI_POSTGRESQL_PGPASS
+          value: /tmp/pgpass
+        - name: PATRONI_POSTGRESQL_LISTEN
+          value: '0.0.0.0:5432'
+        - name: PATRONI_RESTAPI_LISTEN
+          value: '0.0.0.0:8008'
+      terminationGracePeriodSeconds: 0
+      volumes:
+      - name: pgdata
+        emptyDir: {}
+#  volumeClaimTemplates:
+#  - metadata:
+#      labels:
+#        application: spilo
+#        spilo-cluster: *cluster_name
+#      annotations:
+#        volume.alpha.kubernetes.io/storage-class: anything
+#      name: pgdata
+#    spec:
+#      accessModes:
+#      - ReadWriteOnce
+#      resources:
+#        requests:
+#          storage: 5Gi
+
+---
+apiVersion: v1
+kind: Endpoints
+metadata:
+  name: &cluster_name patronidemo
+  labels:
+    application: patroni
+    cluster-name: *cluster_name
+subsets: []
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: &cluster_name patronidemo
+  labels:
+    application: patroni
+    cluster-name: *cluster_name
+spec:
+  type: ClusterIP
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: patronidemo-repl
+  labels:
+    application: patroni
+    cluster-name:  &cluster_name patronidemo
+    role: replica
+spec:
+  type: ClusterIP
+  selector:
+    application: patroni
+    cluster-name: *cluster_name
+    role: replica
+  ports:
+  - port: 5432
+    targetPort: 5432
+
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  name: &cluster_name patronidemo
+  labels:
+    application: patroni
+    cluster-name: *cluster_name
+type: Opaque
+data:
+  superuser-password: emFsYW5kbw==
+  replication-password: cmVwLXBhc3M=
+
+---
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: patronidemo
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: Role
+metadata:
+  name: patronidemo
+rules:
+- apiGroups:
+  - ""
+  resources:
+  - configmaps
+  verbs:
+  - create
+  - get
+  - list
+  - patch
+  - update
+  - watch
+  # delete and deletecollection are required only for 'patronictl remove'
+  - delete
+  - deletecollection
+- apiGroups:
+  - ""
+  resources:
+  - endpoints
+  verbs:
+  - get
+  - patch
+  - update
+  # the following three privileges are necessary only when using endpoints
+  - create
+  - list
+  - watch
+  # delete and deletecollection are required only for for 'patronictl remove'
+  - delete
+  - deletecollection
+- apiGroups:
+  - ""
+  resources:
+  - pods
+  verbs:
+  - get
+  - list
+  - patch
+  - update
+  - watch
+# The following privilege is only necessary for creation of headless service
+# for patronidemo-config endpoint, in order to prevent cleaning it up by the
+# k8s master. You can avoid giving this privilege by explicitly creating the
+# service like it is done in this manifest (lines 2..10)
+- apiGroups:
+  - ""
+  resources:
+  - services
+  verbs:
+  - create
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: RoleBinding
+metadata:
+  name: patronidemo
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: Role
+  name: patronidemo
+subjects:
+- kind: ServiceAccount
+  name: patronidemo
+
+# Following privileges are only required if deployed not in the "default"
+# namespace and you want Patroni to bypass kubernetes service
+# (PATRONI_KUBERNETES_BYPASS_API_SERVICE=true)
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: patroni-k8s-ep-access
+rules:
+- apiGroups:
+  - ""
+  resources:
+  - endpoints
+  resourceNames:
+  - kubernetes
+  verbs:
+  - get
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: patroni-k8s-ep-access
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: patroni-k8s-ep-access
+subjects:
+- kind: ServiceAccount
+  name: patronidemo
+# The namespace must be specified explicitly.
+# If deploying to the different namespace you have to change it.
+  namespace: default

mkbinary.sh

@@ -1,5 +1,5 @@
 #!/bin/sh
 set -e
 
-pip install --ignore-installed setuptools==19.2 pyinstaller
-pyinstaller --clean --onefile patroni.spec
+pip install --ignore-installed pyinstaller
+pyinstaller --clean patroni.spec

patroni.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python
-from patroni import main
+from patroni.__main__ import main
 
 
 if __name__ == '__main__':

patroni.spec

@@ -8,7 +8,7 @@ def hiddenimports():
     sys.path.insert(0, '.')
     try:
         import patroni.dcs
-        return patroni.dcs.dcs_modules()
+        return patroni.dcs.dcs_modules() + ['http.server']
     finally:
         sys.path.pop(0)
 

patroni/__init__.py

@@ -1,200 +1,41 @@
-import logging
-import os
-import signal
 import sys
-import time
 
-logger = logging.getLogger(__name__)
+PATRONI_ENV_PREFIX = 'PATRONI_'
+KUBERNETES_ENV_PREFIX = 'KUBERNETES_'
+MIN_PSYCOPG2 = (2, 5, 4)
 
 
-class Patroni(object):
+def fatal(string, *args):
+    sys.stderr.write('FATAL: ' + string.format(*args) + '\n')
+    sys.exit(1)
 
-    def __init__(self):
-        from patroni.api import RestApiServer
-        from patroni.config import Config
-        from patroni.dcs import get_dcs
-        from patroni.ha import Ha
-        from patroni.postgresql import Postgresql
-        from patroni.version import __version__
 
-        self.setup_signal_handlers()
-
-        self.version = __version__
-        self.config = Config()
-        self.dcs = get_dcs(self.config)
-        self.load_dynamic_configuration()
-
-        self.postgresql = Postgresql(self.config['postgresql'])
-        self.api = RestApiServer(self, self.config['restapi'])
-        self.ha = Ha(self)
-
-        self.tags = self.get_tags()
-        self.next_run = time.time()
-        self.scheduled_restart = {}
-
-    def load_dynamic_configuration(self):
-        from patroni.exceptions import DCSError
-        while True:
+def parse_version(version):
+    def _parse_version(version):
+        for e in version.split('.'):
             try:
-                cluster = self.dcs.get_cluster()
-                if cluster and cluster.config:
-                    if self.config.set_dynamic_configuration(cluster.config):
-                        self.dcs.reload_config(self.config)
-                elif not self.config.dynamic_configuration and 'bootstrap' in self.config:
-                    if self.config.set_dynamic_configuration(self.config['bootstrap']['dcs']):
-                        self.dcs.reload_config(self.config)
+                yield int(e)
+            except ValueError:
                 break
-            except DCSError:
-                logger.warning('Can not get cluster from dcs')
-
-    def get_tags(self):
-        return {tag: value for tag, value in self.config.get('tags', {}).items()
-                if tag not in ('clonefrom', 'nofailover', 'noloadbalance', 'nosync') or value}
-
-    @property
-    def nofailover(self):
-        return bool(self.tags.get('nofailover', False))
-
-    @property
-    def nosync(self):
-        return bool(self.tags.get('nosync', False))
-
-    def reload_config(self):
-        try:
-            self.tags = self.get_tags()
-            self.dcs.reload_config(self.config)
-            self.api.reload_config(self.config['restapi'])
-            self.postgresql.reload_config(self.config['postgresql'])
-        except Exception:
-            logger.exception('Failed to reload config_file=%s', self.config.config_file)
-
-    @property
-    def replicatefrom(self):
-        return self.tags.get('replicatefrom')
-
-    def sighup_handler(self, *args):
-        self._received_sighup = True
-
-    def sigterm_handler(self, *args):
-        if not self._received_sigterm:
-            self._received_sigterm = True
-            sys.exit()
-
-    @property
-    def noloadbalance(self):
-        return bool(self.tags.get('noloadbalance', False))
-
-    def schedule_next_run(self):
-        self.next_run += self.dcs.loop_wait
-        current_time = time.time()
-        nap_time = self.next_run - current_time
-        if nap_time <= 0:
-            self.next_run = current_time
-            # Release the GIL so we don't starve anyone waiting on async_executor lock
-            time.sleep(0.001)
-            # Warn user that Patroni is not keeping up
-            logger.warning("Loop time exceeded, rescheduling immediately.")
-        elif self.ha.watch(nap_time):
-            self.next_run = time.time()
-
-    def run(self):
-        self.api.start()
-        self.next_run = time.time()
-
-        while not self._received_sigterm:
-            if self._received_sighup:
-                self._received_sighup = False
-                if self.config.reload_local_configuration():
-                    self.reload_config()
-
-            logger.info(self.ha.run_cycle())
-
-            cluster = self.dcs.cluster
-            if cluster and cluster.config and self.config.set_dynamic_configuration(cluster.config):
-                self.reload_config()
-
-            if not self.postgresql.data_directory_empty():
-                self.config.save_cache()
-
-            self.schedule_next_run()
-
-    def setup_signal_handlers(self):
-        self._received_sighup = False
-        self._received_sigterm = False
-        signal.signal(signal.SIGHUP, self.sighup_handler)
-        signal.signal(signal.SIGTERM, self.sigterm_handler)
+    return tuple(_parse_version(version.split(' ')[0]))
 
 
-def patroni_main():
-    logging.basicConfig(format='%(asctime)s %(levelname)s: %(message)s', level=logging.INFO)
-    logging.getLogger('requests').setLevel(logging.WARNING)
+# We pass MIN_PSYCOPG2 and parse_version as arguments to simplify usage of check_psycopg from the setup.py
+def check_psycopg(_min_psycopg2=MIN_PSYCOPG2, _parse_version=parse_version):
+    min_psycopg2_str = '.'.join(map(str, _min_psycopg2))
 
-    patroni = Patroni()
     try:
-        patroni.run()
-    except KeyboardInterrupt:
-        pass
-    finally:
-        patroni.api.shutdown()
-        if patroni.ha.is_paused():
-            logger.info('Leader key is not deleted and Postgresql is not stopped due paused state')
-        else:
-            patroni.ha.while_not_sync_standby(lambda: patroni.postgresql.stop(checkpoint=False))
-            patroni.dcs.delete_leader()
+        from psycopg2 import __version__
+        if _parse_version(__version__) >= _min_psycopg2:
+            return
+        version_str = __version__.split(' ')[0]
+    except ImportError:
+        version_str = None
 
-
-def pg_ctl_start(args):
-    import subprocess
-    postmaster = subprocess.Popen(args)
-    print(postmaster.pid)
-
-
-def call_self(args, **kwargs):
-    """This function executes Patroni once again with provided arguments.
-
-    :args: list of arguments to call Patroni with.
-    :returns: `Popen` object"""
-
-    exe = [sys.executable]
-    if not getattr(sys, 'frozen', False):  # Binary distribution?
-        exe.append(sys.argv[0])
-
-    import subprocess
-    return subprocess.Popen(exe + args, **kwargs)
-
-
-def main():
-    if os.getpid() != 1:
-        if len(sys.argv) > 5 and sys.argv[1] == 'pg_ctl_start':
-            return pg_ctl_start(sys.argv[2:])
-        return patroni_main()
-
-    pid = 0
-
-    # Looks like we are in a docker, so we will act like init
-    def sigchld_handler(signo, stack_frame):
-        try:
-            while True:
-                ret = os.waitpid(-1, os.WNOHANG)
-                if ret == (0, 0):
-                    break
-                elif ret[0] != pid:
-                    logging.info('Reaped pid=%s, exit status=%s', *ret)
-        except OSError:
-            pass
-
-    def passtochild(signo, stack_frame):
-        if pid:
-            os.kill(pid, signo)
-
-    signal.signal(signal.SIGCHLD, sigchld_handler)
-    signal.signal(signal.SIGHUP, passtochild)
-    signal.signal(signal.SIGINT, passtochild)
-    signal.signal(signal.SIGUSR1, passtochild)
-    signal.signal(signal.SIGUSR2, passtochild)
-    signal.signal(signal.SIGQUIT, passtochild)
-    signal.signal(signal.SIGTERM, passtochild)
-
-    patroni = call_self(sys.argv[1:])
-    pid = patroni.pid
-    patroni.wait()
+    try:
+        from psycopg import __version__
+    except ImportError:
+        error = 'Patroni requires psycopg2>={0}, psycopg2-binary, or psycopg>=3.0'.format(min_psycopg2_str)
+        if version_str:
+            error += ', but only psycopg2=={0} is available'.format(version_str)
+        fatal(error)

patroni/__main__.py

@@ -1,4 +1,182 @@
-from patroni import main
+import logging
+import os
+import signal
+import time
+
+from patroni.daemon import AbstractPatroniDaemon, abstract_main
+
+logger = logging.getLogger(__name__)
+
+
+class Patroni(AbstractPatroniDaemon):
+
+    def __init__(self, config):
+        from patroni.api import RestApiServer
+        from patroni.dcs import get_dcs
+        from patroni.ha import Ha
+        from patroni.postgresql import Postgresql
+        from patroni.request import PatroniRequest
+        from patroni.version import __version__
+        from patroni.watchdog import Watchdog
+
+        super(Patroni, self).__init__(config)
+
+        self.version = __version__
+        self.dcs = get_dcs(self.config)
+        self.watchdog = Watchdog(self.config)
+        self.load_dynamic_configuration()
+
+        self.postgresql = Postgresql(self.config['postgresql'])
+        self.api = RestApiServer(self, self.config['restapi'])
+        self.request = PatroniRequest(self.config, True)
+        self.ha = Ha(self)
+
+        self.tags = self.get_tags()
+        self.next_run = time.time()
+        self.scheduled_restart = {}
+
+    def load_dynamic_configuration(self):
+        from patroni.exceptions import DCSError
+        while True:
+            try:
+                cluster = self.dcs.get_cluster()
+                if cluster and cluster.config and cluster.config.data:
+                    if self.config.set_dynamic_configuration(cluster.config):
+                        self.dcs.reload_config(self.config)
+                        self.watchdog.reload_config(self.config)
+                elif not self.config.dynamic_configuration and 'bootstrap' in self.config:
+                    if self.config.set_dynamic_configuration(self.config['bootstrap']['dcs']):
+                        self.dcs.reload_config(self.config)
+                        self.watchdog.reload_config(self.config)
+                break
+            except DCSError:
+                logger.warning('Can not get cluster from dcs')
+                time.sleep(5)
+
+    def get_tags(self):
+        return {tag: value for tag, value in self.config.get('tags', {}).items()
+                if tag not in ('clonefrom', 'nofailover', 'noloadbalance', 'nosync') or value}
+
+    @property
+    def nofailover(self):
+        return bool(self.tags.get('nofailover', False))
+
+    @property
+    def nosync(self):
+        return bool(self.tags.get('nosync', False))
+
+    def reload_config(self, sighup=False, local=False):
+        try:
+            super(Patroni, self).reload_config(sighup, local)
+            if local:
+                self.tags = self.get_tags()
+                self.request.reload_config(self.config)
+            if local or sighup and self.api.reload_local_certificate():
+                self.api.reload_config(self.config['restapi'])
+            self.watchdog.reload_config(self.config)
+            self.postgresql.reload_config(self.config['postgresql'], sighup)
+            self.dcs.reload_config(self.config)
+        except Exception:
+            logger.exception('Failed to reload config_file=%s', self.config.config_file)
+
+    @property
+    def replicatefrom(self):
+        return self.tags.get('replicatefrom')
+
+    @property
+    def noloadbalance(self):
+        return bool(self.tags.get('noloadbalance', False))
+
+    def schedule_next_run(self):
+        self.next_run += self.dcs.loop_wait
+        current_time = time.time()
+        nap_time = self.next_run - current_time
+        if nap_time <= 0:
+            self.next_run = current_time
+            # Release the GIL so we don't starve anyone waiting on async_executor lock
+            time.sleep(0.001)
+            # Warn user that Patroni is not keeping up
+            logger.warning("Loop time exceeded, rescheduling immediately.")
+        elif self.ha.watch(nap_time):
+            self.next_run = time.time()
+
+    def run(self):
+        self.api.start()
+        self.next_run = time.time()
+        super(Patroni, self).run()
+
+    def _run_cycle(self):
+        logger.info(self.ha.run_cycle())
+
+        if self.dcs.cluster and self.dcs.cluster.config and self.dcs.cluster.config.data \
+                and self.config.set_dynamic_configuration(self.dcs.cluster.config):
+            self.reload_config()
+
+        if self.postgresql.role != 'uninitialized':
+            self.config.save_cache()
+
+        self.schedule_next_run()
+
+    def _shutdown(self):
+        try:
+            self.api.shutdown()
+        except Exception:
+            logger.exception('Exception during RestApi.shutdown')
+        try:
+            self.ha.shutdown()
+        except Exception:
+            logger.exception('Exception during Ha.shutdown')
+
+
+def patroni_main():
+    from multiprocessing import freeze_support
+    from patroni.validator import schema
+
+    freeze_support()
+    abstract_main(Patroni, schema)
+
+
+def main():
+    if os.getpid() != 1:
+        from patroni import check_psycopg
+
+        check_psycopg()
+        return patroni_main()
+
+    # Patroni started with PID=1, it looks like we are in the container
+    pid = 0
+
+    # Looks like we are in a docker, so we will act like init
+    def sigchld_handler(signo, stack_frame):
+        try:
+            while True:
+                ret = os.waitpid(-1, os.WNOHANG)
+                if ret == (0, 0):
+                    break
+                elif ret[0] != pid:
+                    logger.info('Reaped pid=%s, exit status=%s', *ret)
+        except OSError:
+            pass
+
+    def passtochild(signo, stack_frame):
+        if pid:
+            os.kill(pid, signo)
+
+    if os.name != 'nt':
+        signal.signal(signal.SIGCHLD, sigchld_handler)
+        signal.signal(signal.SIGHUP, passtochild)
+        signal.signal(signal.SIGQUIT, passtochild)
+        signal.signal(signal.SIGUSR1, passtochild)
+        signal.signal(signal.SIGUSR2, passtochild)
+    signal.signal(signal.SIGINT, passtochild)
+    signal.signal(signal.SIGABRT, passtochild)
+    signal.signal(signal.SIGTERM, passtochild)
+
+    import multiprocessing
+    patroni = multiprocessing.Process(target=patroni_main)
+    patroni.start()
+    pid = patroni.pid
+    patroni.join()
 
 
 if __name__ == '__main__':

patroni/async_executor.py

@@ -1,26 +1,78 @@
 import logging
-from threading import RLock, Thread
+from threading import Event, Lock, RLock, Thread
 
 logger = logging.getLogger(__name__)
 
 
+class CriticalTask(object):
+    """Represents a critical task in a background process that we either need to cancel or get the result of.
+
+    Fields of this object may be accessed only when holding a lock on it. To perform the critical task the background
+    thread must, while holding lock on this object, check `is_cancelled` flag, run the task and mark the task as
+    complete using `complete()`.
+
+    The main thread must hold async lock to prevent the task from completing, hold lock on critical task object,
+    call cancel. If the task has completed `cancel()` will return False and `result` field will contain the result of
+    the task. When cancel returns True it is guaranteed that the background task will notice the `is_cancelled` flag.
+    """
+    def __init__(self):
+        self._lock = Lock()
+        self.is_cancelled = False
+        self.result = None
+
+    def reset(self):
+        """Must be called every time the background task is finished.
+
+        Must be called from async thread. Caller must hold lock on async executor when calling."""
+        self.is_cancelled = False
+        self.result = None
+
+    def cancel(self):
+        """Tries to cancel the task, returns True if the task has already run.
+
+        Caller must hold lock on async executor and the task when calling."""
+        if self.result is not None:
+            return False
+        self.is_cancelled = True
+        return True
+
+    def complete(self, result):
+        """Mark task as completed along with a result.
+
+        Must be called from async thread. Caller must hold lock on task when calling."""
+        self.result = result
+
+    def __enter__(self):
+        self._lock.acquire()
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self._lock.release()
+
+
 class AsyncExecutor(object):
 
-    def __init__(self, ha_wakeup):
+    def __init__(self, cancellable, ha_wakeup):
+        self._cancellable = cancellable
         self._ha_wakeup = ha_wakeup
         self._thread_lock = RLock()
         self._scheduled_action = None
         self._scheduled_action_lock = RLock()
+        self._is_cancelled = False
+        self._finish_event = Event()
+        self.critical_task = CriticalTask()
 
     @property
     def busy(self):
         return self.scheduled_action is not None
 
-    def schedule(self, action, immediately=False):
+    def schedule(self, action):
         with self._scheduled_action_lock:
             if self._scheduled_action is not None:
                 return self._scheduled_action
             self._scheduled_action = action
+            self._is_cancelled = False
+            self._finish_event.set()
         return None
 
     @property
@@ -35,20 +87,49 @@ class AsyncExecutor(object):
     def run(self, func, args=()):
         wakeup = False
         try:
+            with self:
+                if self._is_cancelled:
+                    return
+                self._finish_event.clear()
+
+            self._cancellable.reset_is_cancelled()
             # if the func returned something (not None) - wake up main HA loop
             wakeup = func(*args) if args else func()
             return wakeup
-        except:
+        except Exception:
             logger.exception('Exception during execution of long running task %s', self.scheduled_action)
         finally:
             with self:
                 self.reset_scheduled_action()
+                self._finish_event.set()
+                with self.critical_task:
+                    self.critical_task.reset()
             if wakeup is not None:
                 self._ha_wakeup()
 
     def run_async(self, func, args=()):
         Thread(target=self.run, args=(func, args)).start()
 
+    def try_run_async(self, action, func, args=()):
+        prev = self.schedule(action)
+        if prev is None:
+            return self.run_async(func, args)
+        return 'Failed to run {0}, {1} is already in progress'.format(action, prev)
+
+    def cancel(self):
+        with self:
+            with self._scheduled_action_lock:
+                if self._scheduled_action is None:
+                    return
+                logger.warning('Cancelling long running task %s', self._scheduled_action)
+            self._is_cancelled = True
+
+        self._cancellable.cancel()
+        self._finish_event.wait()
+
+        with self:
+            self.reset_scheduled_action()
+
     def __enter__(self):
         self._thread_lock.acquire()
 

patroni/callback_executor.py

@@ -1,37 +0,0 @@
-import logging
-import subprocess
-from threading import Event, Lock, Thread
-
-logger = logging.getLogger(__name__)
-
-
-class CallbackExecutor(Thread):
-
-    def __init__(self):
-        super(CallbackExecutor, self).__init__()
-        self.daemon = True
-        self._lock = Lock()
-        self._cmd = None
-        self._process = None
-        self._callback_event = Event()
-        self.start()
-
-    def call(self, cmd):
-        with self._lock:
-            if self._process and self._process.poll() is None:
-                self._process.kill()
-                logger.warning('Killed the old callback process because it was still running: %s', self._cmd)
-        self._cmd = cmd
-        self._callback_event.set()
-
-    def run(self):
-        while True:
-            self._callback_event.wait()
-            self._callback_event.clear()
-            with self._lock:
-                try:
-                    self._process = subprocess.Popen(self._cmd, close_fds=True)
-                except Exception:
-                    logger.exception('Failed to execute %s',  self._cmd)
-                    continue
-            self._process.wait()

patroni/config.py

@@ -1,18 +1,40 @@
 import json
 import logging
 import os
-import sys
+import shutil
+import six
 import tempfile
 import yaml
 
 from collections import defaultdict
 from copy import deepcopy
+from patroni import PATRONI_ENV_PREFIX
+from patroni.exceptions import ConfigParseError
 from patroni.dcs import ClusterConfig
-from patroni.postgresql import Postgresql
-from patroni.utils import deep_compare, parse_int, patch_config
+from patroni.postgresql.config import CaseInsensitiveDict, ConfigHandler
+from patroni.utils import deep_compare, parse_bool, parse_int, patch_config
 
 logger = logging.getLogger(__name__)
 
+_AUTH_ALLOWED_PARAMETERS = (
+    'username',
+    'password',
+    'sslmode',
+    'sslcert',
+    'sslkey',
+    'sslpassword',
+    'sslrootcert',
+    'sslcrl',
+    'sslcrldir',
+    'gssencmode',
+    'channel_binding'
+)
+
+
+def default_validator(conf):
+    if not conf:
+        raise ConfigParseError("Config is empty.")
+
 
 class Config(object):
     """
@@ -34,43 +56,64 @@ class Config(object):
          to work with it as with the old `config` object.
     """
 
-    PATRONI_ENV_PREFIX = 'PATRONI_'
     PATRONI_CONFIG_VARIABLE = PATRONI_ENV_PREFIX + 'CONFIGURATION'
 
     __CACHE_FILENAME = 'patroni.dynamic.json'
+    __REMAP_KEYS = {
+        'master_start_timeout': 'primary_start_timeout',
+        'master_stop_timeout': 'primary_stop_timeout'
+    }
     __DEFAULT_CONFIG = {
         'ttl': 30, 'loop_wait': 10, 'retry_timeout': 10,
         'maximum_lag_on_failover': 1048576,
-        'master_start_timeout': 300,
+        'maximum_lag_on_syncnode': -1,
+        'check_timeline': False,
+        'primary_start_timeout': 300,
+        'primary_stop_timeout': 0,
         'synchronous_mode': False,
+        'synchronous_mode_strict': False,
+        'synchronous_node_count': 1,
+        'failsafe_mode': False,
+        'standby_cluster': {
+            'create_replica_methods': '',
+            'host': '',
+            'port': '',
+            'primary_slot_name': '',
+            'restore_command': '',
+            'archive_cleanup_command': '',
+            'recovery_min_apply_delay': ''
+        },
         'postgresql': {
             'bin_dir': '',
             'use_slots': True,
-            'parameters': {p: v[0] for p, v in Postgresql.CMDLINE_OPTIONS.items()}
+            'parameters': CaseInsensitiveDict({p: v[0] for p, v in ConfigHandler.CMDLINE_OPTIONS.items()
+                                               if p not in ('wal_keep_segments', 'wal_keep_size')})
+        },
+        'watchdog': {
+            'mode': 'automatic',
         }
     }
 
-    def __init__(self):
+    def __init__(self, configfile, validator=default_validator):
         self._modify_index = -1
         self._dynamic_configuration = {}
 
         self.__environment_configuration = self._build_environment_configuration()
 
         # Patroni reads the configuration from the command-line argument if it exists, otherwise from the environment
-        self._config_file = len(sys.argv) >= 2 and os.path.isfile(sys.argv[1]) and sys.argv[1]
+        self._config_file = configfile and os.path.exists(configfile) and configfile
         if self._config_file:
             self._local_configuration = self._load_config_file()
         else:
             config_env = os.environ.pop(self.PATRONI_CONFIG_VARIABLE, None)
             self._local_configuration = config_env and yaml.safe_load(config_env) or self.__environment_configuration
-            if not self._local_configuration:
-                print('Usage: {0} config.yml'.format(sys.argv[0]))
-                print('\tPatroni may also read the configuration from the {0} environment variable'.
-                      format(self.PATRONI_CONFIG_VARIABLE))
-                sys.exit(1)
+        if validator:
+            errors = validator(self._local_configuration)
+            if errors:
+                raise ConfigParseError("\n".join(errors))
 
         self.__effective_configuration = self._build_effective_configuration({}, self._local_configuration)
-        self._data_dir = self.__effective_configuration['postgresql']['data_dir']
+        self._data_dir = self.__effective_configuration.get('postgresql', {}).get('data_dir', "")
         self._cache_file = os.path.join(self._data_dir, self.__CACHE_FILENAME)
         self._load_cache()
         self._cache_needs_saving = False
@@ -83,12 +126,35 @@ class Config(object):
     def dynamic_configuration(self):
         return deepcopy(self._dynamic_configuration)
 
+    def check_mode(self, mode):
+        return bool(parse_bool(self._dynamic_configuration.get(mode)))
+
+    def _load_config_path(self, path):
+        """
+        If path is a file, loads the yml file pointed to by path.
+        If path is a directory, loads all yml files in that directory in alphabetical order
+        """
+        if os.path.isfile(path):
+            files = [path]
+        elif os.path.isdir(path):
+            files = [os.path.join(path, f) for f in sorted(os.listdir(path))
+                     if (f.endswith('.yml') or f.endswith('.yaml')) and os.path.isfile(os.path.join(path, f))]
+        else:
+            logger.error('config path %s is neither directory nor file', path)
+            raise ConfigParseError('invalid config path')
+
+        overall_config = {}
+        for fname in files:
+            with open(fname) as f:
+                config = yaml.safe_load(f)
+                patch_config(overall_config, config)
+        return overall_config
+
     def _load_config_file(self):
         """Loads config.yaml from filesystem and applies some values which were set via ENV"""
-        with open(self._config_file) as f:
-            config = yaml.safe_load(f)
-            patch_config(config, self.__environment_configuration)
-            return config
+        config = self._load_config_path(self._config_file)
+        patch_config(config, self.__environment_configuration)
+        return config
 
     def _load_cache(self):
         if os.path.isfile(self._cache_file):
@@ -106,7 +172,7 @@ class Config(object):
                 with os.fdopen(fd, 'w') as f:
                     fd = None
                     json.dump(self.dynamic_configuration, f)
-                tmpfile = os.rename(tmpfile, self._cache_file)
+                tmpfile = shutil.move(tmpfile, self._cache_file)
                 self._cache_needs_saving = False
             except Exception:
                 logger.exception('Exception when saving file: %s', self._cache_file)
@@ -139,42 +205,46 @@ class Config(object):
             except Exception:
                 logger.exception('Exception when setting dynamic_configuration')
 
-    def reload_local_configuration(self, dry_run=False):
+    def reload_local_configuration(self):
         if self.config_file:
             try:
                 configuration = self._load_config_file()
                 if not deep_compare(self._local_configuration, configuration):
                     new_configuration = self._build_effective_configuration(self._dynamic_configuration, configuration)
-                    if dry_run:
-                        return not deep_compare(new_configuration, self.__effective_configuration)
                     self._local_configuration = configuration
                     self.__effective_configuration = new_configuration
                     return True
+                else:
+                    logger.info('No local configuration items changed.')
             except Exception:
                 logger.exception('Exception when reloading local configuration from %s', self.config_file)
-                if dry_run:
-                    raise
 
     @staticmethod
     def _process_postgresql_parameters(parameters, is_local=False):
-        ret = {}
-        for name, value in (parameters or {}).items():
-            if name not in Postgresql.CMDLINE_OPTIONS or not is_local and Postgresql.CMDLINE_OPTIONS[name][1](value):
-                ret[name] = value
-        return ret
+        return {name: value for name, value in (parameters or {}).items()
+                if name not in ConfigHandler.CMDLINE_OPTIONS or
+                not is_local and ConfigHandler.CMDLINE_OPTIONS[name][1](value)}
 
     def _safe_copy_dynamic_configuration(self, dynamic_configuration):
         config = deepcopy(self.__DEFAULT_CONFIG)
 
         for name, value in dynamic_configuration.items():
+            # allow copying master_start_timeout->primary_start_timeout when the latter isn't in dynamic_configuration
+            if name in self.__REMAP_KEYS and self.__REMAP_KEYS[name] not in dynamic_configuration:
+                name = self.__REMAP_KEYS[name]
             if name == 'postgresql':
                 for name, value in (value or {}).items():
                     if name == 'parameters':
                         config['postgresql'][name].update(self._process_postgresql_parameters(value))
-                    elif name not in ('connect_address', 'listen', 'data_dir', 'pgpass', 'authentication'):
+                    elif name not in ('connect_address', 'proxy_address', 'listen',
+                                      'config_dir', 'data_dir', 'pgpass', 'authentication'):
                         config['postgresql'][name] = deepcopy(value)
-            elif name in config:  # only variables present in __DEFAULT_CONFIG allowed to be overriden from DCS
-                if name == 'synchronous_mode':
+            elif name == 'standby_cluster':
+                for name, value in (value or {}).items():
+                    if name in self.__DEFAULT_CONFIG['standby_cluster']:
+                        config['standby_cluster'][name] = deepcopy(value)
+            elif name in config:  # only variables present in __DEFAULT_CONFIG allowed to be overridden from DCS
+                if name in ('synchronous_mode', 'synchronous_mode_strict', 'failsafe_mode'):
                     config[name] = value
                 else:
                     config[name] = int(value)
@@ -185,44 +255,51 @@ class Config(object):
         ret = defaultdict(dict)
 
         def _popenv(name):
-            return os.environ.pop(Config.PATRONI_ENV_PREFIX + name.upper(), None)
+            return os.environ.pop(PATRONI_ENV_PREFIX + name.upper(), None)
 
         for param in ('name', 'namespace', 'scope'):
             value = _popenv(param)
             if value:
                 ret[param] = value
 
+        def _fix_log_env(name, oldname):
+            value = _popenv(oldname)
+            name = PATRONI_ENV_PREFIX + 'LOG_' + name.upper()
+            if value and name not in os.environ:
+                os.environ[name] = value
+
+        for name, oldname in (('level', 'loglevel'), ('format', 'logformat'), ('dateformat', 'log_datefmt')):
+            _fix_log_env(name, oldname)
+
         def _set_section_values(section, params):
             for param in params:
                 value = _popenv(section + '_' + param)
                 if value:
                     ret[section][param] = value
 
-        _set_section_values('restapi', ['listen', 'connect_address', 'certfile', 'keyfile'])
-        _set_section_values('postgresql', ['listen', 'connect_address', 'data_dir', 'pgpass', 'bin_dir'])
+        _set_section_values('restapi', ['listen', 'connect_address', 'certfile', 'keyfile', 'keyfile_password',
+                                        'cafile', 'ciphers', 'verify_client', 'http_extra_headers',
+                                        'https_extra_headers', 'allowlist', 'allowlist_include_members'])
+        _set_section_values('ctl', ['insecure', 'cacert', 'certfile', 'keyfile', 'keyfile_password'])
+        _set_section_values('postgresql', ['listen', 'connect_address', 'proxy_address',
+                                           'config_dir', 'data_dir', 'pgpass', 'bin_dir'])
+        _set_section_values('log', ['level', 'traceback_level', 'format', 'dateformat', 'max_queue_size',
+                                    'dir', 'file_size', 'file_num', 'loggers'])
+        _set_section_values('raft', ['data_dir', 'self_addr', 'partner_addrs', 'password', 'bind_addr'])
 
-        def _get_auth(name):
-            ret = {}
-            for param in ('username', 'password'):
-                value = _popenv(name + '_' + param)
-                if value:
-                    ret[param] = value
-            return ret
+        for first, second in (('restapi', 'allowlist_include_members'), ('ctl', 'insecure')):
+            value = ret.get(first, {}).pop(second, None)
+            if value:
+                value = parse_bool(value)
+                if value is not None:
+                    ret[first][second] = value
 
-        restapi_auth = _get_auth('restapi')
-        if restapi_auth:
-            ret['restapi']['authentication'] = restapi_auth
-
-        authentication = {}
-        for user_type in ('replication', 'superuser'):
-            entry = _get_auth(user_type)
-            if entry:
-                authentication[user_type] = entry
-
-        if authentication:
-            ret['postgresql']['authentication'] = authentication
-
-        users = {}
+        for second in ('max_queue_size', 'file_size', 'file_num'):
+            value = ret.get('log', {}).pop(second, None)
+            if value:
+                value = parse_int(value)
+                if value is not None:
+                    ret['log'][second] = value
 
         def _parse_list(value):
             if not (value.strip().startswith('-') or '[' in value):
@@ -233,30 +310,95 @@ class Config(object):
                 logger.exception('Exception when parsing list %s', value)
                 return None
 
+        for first, second in (('raft', 'partner_addrs'), ('restapi', 'allowlist')):
+            value = ret.get(first, {}).pop(second, None)
+            if value:
+                value = _parse_list(value)
+                if value:
+                    ret[first][second] = value
+
+        def _parse_dict(value):
+            if not value.strip().startswith('{'):
+                value = '{{{0}}}'.format(value)
+            try:
+                return yaml.safe_load(value)
+            except Exception:
+                logger.exception('Exception when parsing dict %s', value)
+                return None
+
+        for first, params in (('restapi', ('http_extra_headers', 'https_extra_headers')), ('log', ('loggers',))):
+            for second in params:
+                value = ret.get(first, {}).pop(second, None)
+                if value:
+                    value = _parse_dict(value)
+                    if value:
+                        ret[first][second] = value
+
+        def _get_auth(name, params=None):
+            ret = {}
+            for param in params or _AUTH_ALLOWED_PARAMETERS[:2]:
+                value = _popenv(name + '_' + param)
+                if value:
+                    ret[param] = value
+            return ret
+
+        restapi_auth = _get_auth('restapi')
+        if restapi_auth:
+            ret['restapi']['authentication'] = restapi_auth
+
+        authentication = {}
+        for user_type in ('replication', 'superuser', 'rewind'):
+            entry = _get_auth(user_type, _AUTH_ALLOWED_PARAMETERS)
+            if entry:
+                authentication[user_type] = entry
+
+        if authentication:
+            ret['postgresql']['authentication'] = authentication
+
         for param in list(os.environ.keys()):
-            if param.startswith(Config.PATRONI_ENV_PREFIX):
+            if param.startswith(PATRONI_ENV_PREFIX):
+                # PATRONI_(ETCD|CONSUL|ZOOKEEPER|EXHIBITOR|...)_(HOSTS?|PORT|..)
+                name, suffix = (param[8:].split('_', 1) + [''])[:2]
+                if suffix in ('HOST', 'HOSTS', 'PORT', 'USE_PROXIES', 'PROTOCOL', 'SRV', 'SRV_SUFFIX', 'URL', 'PROXY',
+                              'CACERT', 'CERT', 'KEY', 'VERIFY', 'TOKEN', 'CHECKS', 'DC', 'CONSISTENCY',
+                              'REGISTER_SERVICE', 'SERVICE_CHECK_INTERVAL', 'SERVICE_CHECK_TLS_SERVER_NAME',
+                              'SERVICE_TAGS', 'NAMESPACE', 'CONTEXT', 'USE_ENDPOINTS', 'SCOPE_LABEL', 'ROLE_LABEL',
+                              'POD_IP', 'PORTS', 'LABELS', 'BYPASS_API_SERVICE', 'KEY_PASSWORD', 'USE_SSL', 'SET_ACLS',
+                              'GROUP', 'DATABASE') and name:
+                    value = os.environ.pop(param)
+                    if name == 'CITUS':
+                        if suffix == 'GROUP':
+                            value = parse_int(value)
+                        elif suffix != 'DATABASE':
+                            continue
+                    elif suffix == 'PORT':
+                        value = value and parse_int(value)
+                    elif suffix in ('HOSTS', 'PORTS', 'CHECKS', 'SERVICE_TAGS'):
+                        value = value and _parse_list(value)
+                    elif suffix in ('LABELS', 'SET_ACLS'):
+                        value = _parse_dict(value)
+                    elif suffix in ('USE_PROXIES', 'REGISTER_SERVICE', 'USE_ENDPOINTS', 'BYPASS_API_SERVICE', 'VERIFY'):
+                        value = parse_bool(value)
+                    if value is not None:
+                        ret[name.lower()][suffix.lower()] = value
+        for dcs in ('etcd', 'etcd3'):
+            if dcs in ret:
+                ret[dcs].update(_get_auth(dcs))
+
+        users = {}
+        for param in list(os.environ.keys()):
+            if param.startswith(PATRONI_ENV_PREFIX):
                 name, suffix = (param[8:].rsplit('_', 1) + [''])[:2]
-                if name and suffix:
-                    # PATRONI_(ETCD|CONSUL|ZOOKEEPER|EXHIBITOR|...)_(HOSTS?|PORT|..)
-                    if suffix in ('HOST', 'HOSTS', 'PORT', 'SRV', 'URL', 'PROXY', 'CACERT', 'CERT', 'KEY') \
-                            and '_' not in name:
-                        value = os.environ.pop(param)
-                        if suffix == 'PORT':
-                            value = value and parse_int(value)
-                        elif suffix == 'HOSTS':
-                            value = value and _parse_list(value)
-                        if value:
-                            ret[name.lower()][suffix.lower()] = value
-                    # PATRONI_<username>_PASSWORD=<password>, PATRONI_<username>_OPTIONS=<option1,option2,...>
-                    # CREATE USER "<username>" WITH <OPTIONS> PASSWORD '<password>'
-                    elif suffix == 'PASSWORD':
-                        password = os.environ.pop(param)
-                        if password:
-                            users[name] = {'password': password}
-                            options = os.environ.pop(param[:-9] + '_OPTIONS', None)
-                            options = options and _parse_list(options)
-                            if options:
-                                users[name]['options'] = options
+                # PATRONI_<username>_PASSWORD=<password>, PATRONI_<username>_OPTIONS=<option1,option2,...>
+                # CREATE USER "<username>" WITH <OPTIONS> PASSWORD '<password>'
+                if name and suffix == 'PASSWORD':
+                    password = os.environ.pop(param)
+                    if password:
+                        users[name] = {'password': password}
+                        options = os.environ.pop(param[:-9] + '_OPTIONS', None)
+                        options = options and _parse_list(options)
+                        if options:
+                            users[name]['options'] = options
         if users:
             ret['bootstrap']['users'] = users
 
@@ -265,17 +407,21 @@ class Config(object):
     def _build_effective_configuration(self, dynamic_configuration, local_configuration):
         config = self._safe_copy_dynamic_configuration(dynamic_configuration)
         for name, value in local_configuration.items():
-            if name == 'postgresql':
+            if name == 'citus':  # remove invalid citus configuration
+                if isinstance(value, dict) and isinstance(value.get('group'), six.integer_types)\
+                        and isinstance(value.get('database'), six.string_types):
+                    config[name] = value
+            elif name == 'postgresql':
                 for name, value in (value or {}).items():
                     if name == 'parameters':
                         config['postgresql'][name].update(self._process_postgresql_parameters(value, True))
                     elif name != 'use_slots':  # replication slots must be enabled/disabled globally
                         config['postgresql'][name] = deepcopy(value)
-            elif name not in config:
+            elif name not in config or name in ['watchdog']:
                 config[name] = deepcopy(value) if value else {}
 
         # restapi server expects to get restapi.auth = 'username:password'
-        if 'authentication' in config['restapi']:
+        if 'restapi' in config and 'authentication' in config['restapi']:
             config['restapi']['auth'] = '{username}:{password}'.format(**config['restapi']['authentication'])
 
         # special treatment for old config
@@ -294,12 +440,33 @@ class Config(object):
         if 'superuser' not in pg_config['authentication'] and 'pg_rewind' in pg_config:
             pg_config['authentication']['superuser'] = pg_config['pg_rewind']
 
+        # handle setting additional connection parameters that may be available
+        # in the configuration file, such as SSL connection parameters
+        for name, value in pg_config['authentication'].items():
+            pg_config['authentication'][name] = {n: v for n, v in value.items() if n in _AUTH_ALLOWED_PARAMETERS}
+
         # no 'name' in config
         if 'name' not in config and 'name' in pg_config:
             config['name'] = pg_config['name']
 
-        pg_config.update({p: config[p] for p in ('name', 'scope', 'retry_timeout',
-                          'synchronous_mode', 'maximum_lag_on_failover') if p in config})
+        # when bootstrapping the new Citus cluster (coordinator/worker) enable sync replication in global configuration
+        if 'citus' in config:
+            bootstrap = config.setdefault('bootstrap', {})
+            dcs = bootstrap.setdefault('dcs', {})
+            dcs.setdefault('synchronous_mode', True)
+
+        updated_fields = (
+            'name',
+            'scope',
+            'retry_timeout',
+            'synchronous_mode',
+            'synchronous_mode_strict',
+            'synchronous_node_count',
+            'maximum_lag_on_syncnode',
+            'citus'
+        )
+
+        pg_config.update({p: config[p] for p in updated_fields if p in config})
 
         return config
 

patroni/daemon.py

@@ -0,0 +1,112 @@
+from __future__ import print_function
+
+import abc
+import os
+import signal
+import six
+import sys
+
+from threading import Lock
+
+
+@six.add_metaclass(abc.ABCMeta)
+class AbstractPatroniDaemon(object):
+
+    def __init__(self, config):
+        from patroni.log import PatroniLogger
+
+        self.setup_signal_handlers()
+
+        self.logger = PatroniLogger()
+        self.config = config
+        AbstractPatroniDaemon.reload_config(self, local=True)
+
+    def sighup_handler(self, *args):
+        self._received_sighup = True
+
+    def api_sigterm(self):
+        with self._sigterm_lock:
+            if not self._received_sigterm:
+                self._received_sigterm = True
+                return True
+
+    def sigterm_handler(self, *args):
+        if self.api_sigterm():
+            sys.exit()
+
+    def setup_signal_handlers(self):
+        self._received_sighup = False
+        self._sigterm_lock = Lock()
+        self._received_sigterm = False
+        if os.name != 'nt':
+            signal.signal(signal.SIGHUP, self.sighup_handler)
+        signal.signal(signal.SIGTERM, self.sigterm_handler)
+
+    @property
+    def received_sigterm(self):
+        with self._sigterm_lock:
+            return self._received_sigterm
+
+    def reload_config(self, sighup=False, local=False):
+        if local:
+            self.logger.reload_config(self.config.get('log', {}))
+
+    @abc.abstractmethod
+    def _run_cycle(self):
+        """_run_cycle"""
+
+    def run(self):
+        self.logger.start()
+        while not self.received_sigterm:
+            if self._received_sighup:
+                self._received_sighup = False
+                self.reload_config(True, self.config.reload_local_configuration())
+
+            self._run_cycle()
+
+    @abc.abstractmethod
+    def _shutdown(self):
+        """_shutdown"""
+
+    def shutdown(self):
+        with self._sigterm_lock:
+            self._received_sigterm = True
+        self._shutdown()
+        self.logger.shutdown()
+
+
+def abstract_main(cls, validator=None):
+    import argparse
+
+    from .config import Config, ConfigParseError
+    from .version import __version__
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--version', action='version', version='%(prog)s {0}'.format(__version__))
+    if validator:
+        parser.add_argument('--validate-config', action='store_true', help='Run config validator and exit')
+    parser.add_argument('configfile', nargs='?', default='',
+                        help='Patroni may also read the configuration from the {0} environment variable'
+                        .format(Config.PATRONI_CONFIG_VARIABLE))
+    args = parser.parse_args()
+    validate_config = validator and args.validate_config
+    try:
+        if validate_config:
+            Config(args.configfile, validator=validator)
+            sys.exit()
+
+        config = Config(args.configfile)
+    except ConfigParseError as e:
+        if e.value:
+            print(e.value, file=sys.stderr)
+        if not validate_config:
+            parser.print_help()
+        sys.exit(1)
+
+    controller = cls(config)
+    try:
+        controller.run()
+    except KeyboardInterrupt:
+        pass
+    finally:
+        controller.shutdown()

patroni/dcs/__init__.py

@@ -1,19 +1,46 @@
 import abc
-import dateutil
+import dateutil.parser
 import importlib
 import inspect
 import json
+import logging
 import os
 import pkgutil
+import re
 import six
 import sys
+import time
 
-from collections import namedtuple
-from patroni.exceptions import PatroniException
+from collections import defaultdict, namedtuple
+from copy import deepcopy
 from random import randint
 from six.moves.urllib_parse import urlparse, urlunparse, parse_qsl
 from threading import Event, Lock
 
+from ..exceptions import PatroniFatalException
+from ..utils import deep_compare, parse_bool, uri
+
+CITUS_COORDINATOR_GROUP_ID = 0
+citus_group_re = re.compile('^(0|[1-9][0-9]*)$')
+slot_name_re = re.compile('^[a-z0-9_]{1,63}$')
+logger = logging.getLogger(__name__)
+
+
+def slot_name_from_member_name(member_name):
+    """Translate member name to valid PostgreSQL slot name.
+
+    PostgreSQL replication slot names must be valid PostgreSQL names. This function maps the wider space of
+    member names to valid PostgreSQL names. Names are lowercased, dashes and periods common in hostnames
+    are replaced with underscores, other characters are encoded as their unicode codepoint. Name is truncated
+    to 64 characters. Multiple different member names may map to a single slot name."""
+
+    def replace_char(match):
+        c = match.group(0)
+        return '_' if c in '-.' else "u{:04d}".format(ord(c))
+
+    slot_name = re.sub('[^a-z0-9_]', replace_char, member_name.lower())
+    return slot_name[0:63]
+
 
 def parse_connection_string(value):
     """Original Governor stores connection strings for each cluster members if a following format:
@@ -42,29 +69,51 @@ def dcs_modules():
     module_prefix = __package__ + '.'
 
     if getattr(sys, 'frozen', False):
-        importer = pkgutil.get_importer(dcs_dirname)
-        return [module for module in list(importer.toc) if module.startswith(module_prefix) and module.count('.') == 2]
+        toc = set()
+        # dcs_dirname may contain a dot, which causes pkgutil.iter_importers()
+        # to misinterpret the path as a package name. This can be avoided
+        # altogether by not passing a path at all, because PyInstaller's
+        # FrozenImporter is a singleton and registered as top-level finder.
+        for importer in pkgutil.iter_importers():
+            if hasattr(importer, 'toc'):
+                toc |= importer.toc
+        return [module for module in toc if module.startswith(module_prefix) and module.count('.') == 2]
     else:
         return [module_prefix + name for _, name, is_pkg in pkgutil.iter_modules([dcs_dirname]) if not is_pkg]
 
 
 def get_dcs(config):
-    available_implementations = set()
-    for module_name in dcs_modules():
-        module = importlib.import_module(module_name)
-        for name in filter(lambda name: not name.startswith('__'), dir(module)):  # iterate through module content
-            value = getattr(module, name)
-            name = name.lower()
-            # try to find implementation of AbstractDCS interface, class name must match with module_name
-            if inspect.isclass(value) and issubclass(value, AbstractDCS) and __package__ + '.' + name == module_name:
-                available_implementations.add(name)
-                if name in config:  # which has configuration section in the config file
-                    # propagate some parameters
-                    config[name].update({p: config[p] for p in ('namespace', 'name', 'scope', 'loop_wait',
-                                         'patronictl', 'ttl', 'retry_timeout') if p in config})
-                    return value(config[name])
-    raise PatroniException("""Can not find suitable configuration of distributed configuration store
-Available implementations: """ + ', '.join(available_implementations))
+    modules = dcs_modules()
+
+    for module_name in modules:
+        name = module_name.split('.')[-1]
+        if name in config:  # we will try to import only modules which have configuration section in the config file
+            try:
+                module = importlib.import_module(module_name)
+                for key, item in module.__dict__.items():  # iterate through the module content
+                    # try to find implementation of AbstractDCS interface, class name must match with module_name
+                    if key.lower() == name and inspect.isclass(item) and issubclass(item, AbstractDCS):
+                        # propagate some parameters
+                        config[name].update({p: config[p] for p in ('namespace', 'name', 'scope', 'loop_wait',
+                                             'patronictl', 'ttl', 'retry_timeout') if p in config})
+                        # From citus section we only need "group" parameter, but will propagate everything just in case.
+                        if isinstance(config.get('citus'), dict):
+                            config[name].update(config['citus'])
+                        return item(config[name])
+            except ImportError:
+                logger.debug('Failed to import %s', module_name)
+
+    available_implementations = []
+    for module_name in modules:
+        name = module_name.split('.')[-1]
+        try:
+            module = importlib.import_module(module_name)
+            available_implementations.extend(name for key, item in module.__dict__.items() if key.lower() == name
+                                             and inspect.isclass(item) and issubclass(item, AbstractDCS))
+        except ImportError:
+            logger.info('Failed to import %s', module_name)
+    raise PatroniFatalException("""Can not find suitable configuration of distributed configuration store
+Available implementations: """ + ', '.join(sorted(set(available_implementations))))
 
 
 class Member(namedtuple('Member', 'index,name,session,data')):
@@ -94,32 +143,51 @@ class Member(namedtuple('Member', 'index,name,session,data')):
         else:
             try:
                 data = json.loads(data)
+                if not isinstance(data, dict):
+                    data = {}
             except (TypeError, ValueError):
                 data = {}
         return Member(index, name, session, data)
 
     @property
     def conn_url(self):
-        return self.data.get('conn_url')
+        conn_url = self.data.get('conn_url')
+        if conn_url:
+            return conn_url
+
+        conn_kwargs = self.data.get('conn_kwargs')
+        if conn_kwargs:
+            conn_url = uri('postgresql', (conn_kwargs.get('host'), conn_kwargs.get('port', 5432)))
+            self.data['conn_url'] = conn_url
+            return conn_url
 
     def conn_kwargs(self, auth=None):
+        defaults = {
+            "host": None,
+            "port": None,
+            "dbname": None
+        }
         ret = self.data.get('conn_kwargs')
         if ret:
-            ret = ret.copy()
+            defaults.update(ret)
+            ret = defaults
         else:
-            r = urlparse(self.conn_url)
+            conn_url = self.conn_url
+            if not conn_url:
+                return {}  # due to the invalid conn_url we don't care about authentication parameters
+            r = urlparse(conn_url)
             ret = {
                 'host': r.hostname,
                 'port': r.port or 5432,
-                'database': r.path[1:]
+                'dbname': r.path[1:]
             }
             self.data['conn_kwargs'] = ret.copy()
 
+        # apply any remaining authentication parameters
         if auth and isinstance(auth, dict):
+            ret.update({k: v for k, v in auth.items() if v is not None})
             if 'username' in auth:
-                ret['user'] = auth['username']
-            if 'password' in auth:
-                ret['password'] = auth['password']
+                ret['user'] = ret.pop('username')
         return ret
 
     @property
@@ -150,6 +218,34 @@ class Member(namedtuple('Member', 'index,name,session,data')):
     def is_running(self):
         return self.state == 'running'
 
+    @property
+    def version(self):
+        version = self.data.get('version')
+        if version:
+            try:
+                return tuple(map(int, version.split('.')))
+            except Exception:
+                logger.debug('Failed to parse Patroni version %s', version)
+
+
+class RemoteMember(Member):
+    """Represents a remote member (typically a primary) for a standby cluster"""
+    def __new__(cls, name, data):
+        return super(RemoteMember, cls).__new__(cls, None, name, None, data)
+
+    @staticmethod
+    def allowed_keys():
+        return ('primary_slot_name',
+                'create_replica_methods',
+                'restore_command',
+                'archive_cleanup_command',
+                'recovery_min_apply_delay',
+                'no_replication_slot')
+
+    def __getattr__(self, name):
+        if name in RemoteMember.allowed_keys():
+            return self.data.get(name)
+
 
 class Leader(namedtuple('Leader', 'index,session,member')):
 
@@ -170,40 +266,62 @@ class Leader(namedtuple('Leader', 'index,session,member')):
     def conn_url(self):
         return self.member.conn_url
 
+    @property
+    def data(self):
+        return self.member.data
+
+    @property
+    def timeline(self):
+        return self.data.get('timeline')
+
+    @property
+    def checkpoint_after_promote(self):
+        """
+        >>> Leader(1, '', Member.from_node(1, '', '', '{"version":"z"}')).checkpoint_after_promote
+        """
+        version = self.member.version
+        # 1.5.6 is the last version which doesn't expose checkpoint_after_promote: false
+        if version and version > (1, 5, 6):
+            return self.data.get('role') in ('master', 'primary') and 'checkpoint_after_promote' not in self.data
+
 
 class Failover(namedtuple('Failover', 'index,leader,candidate,scheduled_at')):
 
     """
     >>> 'Failover' in str(Failover.from_node(1, '{"leader": "cluster_leader"}'))
     True
+    >>> 'Failover' in str(Failover.from_node(1, {"leader": "cluster_leader"}))
+    True
     >>> 'Failover' in str(Failover.from_node(1, '{"leader": "cluster_leader", "member": "cluster_candidate"}'))
     True
     >>> Failover.from_node(1, 'null') is None
-    True
+    False
     >>> n = '{"leader": "cluster_leader", "member": "cluster_candidate", "scheduled_at": "2016-01-14T10:09:57.1394Z"}'
     >>> 'tzinfo=' in str(Failover.from_node(1, n))
     True
     >>> Failover.from_node(1, None) is None
-    True
+    False
     >>> Failover.from_node(1, '{}') is None
-    True
+    False
     >>> 'abc' in Failover.from_node(1, 'abc:def')
     True
     """
     @staticmethod
     def from_node(index, value):
-        if not value:
-            return None
-
-        try:
-            data = json.loads(value)
-            if not data:
-                return None
-        except ValueError:
-            t = [a.strip() for a in value.split(':')]
-            leader = t[0]
-            candidate = t[1] if len(t) > 1 else None
-            return Failover(index, leader, candidate, None) if leader or candidate else None
+        if isinstance(value, dict):
+            data = value
+        elif value:
+            try:
+                data = json.loads(value)
+                if not isinstance(data, dict):
+                    data = {}
+            except ValueError:
+                t = [a.strip() for a in value.split(':')]
+                leader = t[0]
+                candidate = t[1] if len(t) > 1 else None
+                return Failover(index, leader, candidate, None) if leader or candidate else None
+        else:
+            data = {}
 
         if data.get('scheduled_at'):
             data['scheduled_at'] = dateutil.parser.parse(data['scheduled_at'])
@@ -220,14 +338,32 @@ class ClusterConfig(namedtuple('ClusterConfig', 'index,data,modify_index')):
     def from_node(index, data, modify_index=None):
         """
         >>> ClusterConfig.from_node(1, '{') is None
-        True
+        False
         """
 
         try:
             data = json.loads(data)
         except (TypeError, ValueError):
-            return None
-        return ClusterConfig(index, data, modify_index or index)
+            data = None
+            modify_index = 0
+        if not isinstance(data, dict):
+            data = {}
+        return ClusterConfig(index, data, index if modify_index is None else modify_index)
+
+    @property
+    def permanent_slots(self):
+        return isinstance(self.data, dict) and (
+                self.data.get('permanent_replication_slots') or
+                self.data.get('permanent_slots') or self.data.get('slots')
+        ) or {}
+
+    @property
+    def ignore_slots_matchers(self):
+        return isinstance(self.data, dict) and self.data.get('ignore_slots') or []
+
+    @property
+    def max_timelines_history(self):
+        return self.data.get('max_timelines_history', 0)
 
 
 class SyncState(namedtuple('SyncState', 'index,leader,sync_standby')):
@@ -235,7 +371,7 @@ class SyncState(namedtuple('SyncState', 'index,leader,sync_standby')):
 
     :param index: modification index of a synchronization key in a Configuration Store
     :param leader: reference to member that was leader
-    :param sync_standby: standby that was last synchronized to leader
+    :param sync_standby: synchronous standby list (comma delimited) which are last synchronized to leader
     """
 
     @staticmethod
@@ -251,8 +387,12 @@ class SyncState(namedtuple('SyncState', 'index,leader,sync_standby')):
         True
         >>> SyncState.from_node(1, '{"leader": "leader"}').leader == "leader"
         True
+        >>> SyncState.from_node(1, {"leader": "leader"}).leader == "leader"
+        True
         """
-        if value:
+        if isinstance(value, dict):
+            data = value
+        elif value:
             try:
                 data = json.loads(value)
                 if not isinstance(data, dict):
@@ -263,15 +403,22 @@ class SyncState(namedtuple('SyncState', 'index,leader,sync_standby')):
             data = {}
         return SyncState(index, data.get('leader'), data.get('sync_standby'))
 
+    @property
+    def members(self):
+        """ Returns sync_standby in list """
+        return self.sync_standby and self.sync_standby.split(',') or []
+
     def matches(self, name):
         """
         Returns if a node name matches one of the nodes in the sync state
 
-        >>> s = SyncState(1, 'foo', 'bar')
+        >>> s = SyncState(1, 'foo', 'bar,zoo')
         >>> s.matches('foo')
         True
         >>> s.matches('bar')
         True
+        >>> s.matches('zoo')
+        True
         >>> s.matches('baz')
         False
         >>> s.matches(None)
@@ -279,25 +426,58 @@ class SyncState(namedtuple('SyncState', 'index,leader,sync_standby')):
         >>> SyncState(1, None, None).matches('foo')
         False
         """
-        return name is not None and name in (self.leader, self.sync_standby)
+        return name is not None and name in [self.leader] + self.members
 
 
-class Cluster(namedtuple('Cluster', 'initialize,config,leader,last_leader_operation,members,failover,sync')):
+class TimelineHistory(namedtuple('TimelineHistory', 'index,value,lines')):
+    """Object representing timeline history file"""
+
+    @staticmethod
+    def from_node(index, value):
+        """
+        >>> h = TimelineHistory.from_node(1, 2)
+        >>> h.lines
+        []
+        """
+        try:
+            lines = json.loads(value)
+        except (TypeError, ValueError):
+            lines = None
+        if not isinstance(lines, list):
+            lines = []
+        return TimelineHistory(index, value, lines)
+
+
+class Cluster(namedtuple('Cluster', 'initialize,config,leader,last_lsn,members,'
+                                    'failover,sync,history,slots,failsafe,workers')):
 
     """Immutable object (namedtuple) which represents PostgreSQL cluster.
     Consists of the following fields:
     :param initialize: shows whether this cluster has initialization key stored in DC or not.
     :param config: global dynamic configuration, reference to `ClusterConfig` object
     :param leader: `Leader` object which represents current leader of the cluster
-    :param last_leader_operation: int or long object containing position of last known leader operation.
-        This value is stored in `/optime/leader` key
+    :param last_lsn: int or long object containing position of last known leader LSN.
+        This value is stored in the `/status` key or `/optime/leader` (legacy) key
     :param members: list of Member object, all PostgreSQL cluster members including leader
     :param failover: reference to `Failover` object
     :param sync: reference to `SyncState` object, last observed synchronous replication state.
-    """
+    :param history: reference to `TimelineHistory` object
+    :param slots: state of permanent logical replication slots on the primary in the format: {"slot_name": int}
+    :param failsafe: failsafe topology. Node is allowed to become the leader only if its name is found in this list.
+    :param workers: workers of the Citus cluster, optional. Format: {int(group): Cluster()}"""
+
+    def __new__(cls, *args):
+        # Make workers argument optional
+        if len(cls._fields) == len(args) + 1:
+            args = args + ({},)
+        return super(Cluster, cls).__new__(cls, *args)
+
+    @property
+    def leader_name(self):
+        return self.leader and self.leader.name
 
     def is_unlocked(self):
-        return not (self.leader and self.leader.name)
+        return not self.leader_name
 
     def has_member(self, member_name):
         return any(m for m in self.members if m.name == member_name)
@@ -310,8 +490,168 @@ class Cluster(namedtuple('Cluster', 'initialize,config,leader,last_leader_operat
         candidates = [m for m in self.members if m.clonefrom and m.is_running and m.name not in exclude]
         return candidates[randint(0, len(candidates) - 1)] if candidates else self.leader
 
+    def check_mode(self, mode):
+        return bool(self.config and parse_bool(self.config.data.get(mode)))
+
     def is_paused(self):
-        return self.config and self.config.data.get('pause', False) or False
+        return self.check_mode('pause')
+
+    def is_synchronous_mode(self):
+        return self.check_mode('synchronous_mode')
+
+    @property
+    def __permanent_slots(self):
+        return self.config and self.config.permanent_slots or {}
+
+    @property
+    def __permanent_physical_slots(self):
+        return {name: value for name, value in self.__permanent_slots.items()
+                if not value or isinstance(value, dict) and value.get('type', 'physical') == 'physical'}
+
+    @property
+    def __permanent_logical_slots(self):
+        return {name: value for name, value in self.__permanent_slots.items() if isinstance(value, dict)
+                and value.get('type', 'logical') == 'logical' and value.get('database') and value.get('plugin')}
+
+    @property
+    def use_slots(self):
+        return self.config and (self.config.data.get('postgresql') or {}).get('use_slots', True)
+
+    def get_replication_slots(self, my_name, role, nofailover, major_version, show_error=False):
+        # if the replicatefrom tag is set on the member - we should not create the replication slot for it on
+        # the current primary, because that member would replicate from elsewhere. We still create the slot if
+        # the replicatefrom destination member is currently not a member of the cluster (fallback to the
+        # primary), or if replicatefrom destination member happens to be the current primary
+        use_slots = self.use_slots
+        if role in ('master', 'primary', 'standby_leader'):
+            slot_members = [m.name for m in self.members if use_slots and m.name != my_name and
+                            (m.replicatefrom is None or m.replicatefrom == my_name or
+                             not self.has_member(m.replicatefrom))]
+            permanent_slots = self.__permanent_slots if use_slots and \
+                role in ('master', 'primary') else self.__permanent_physical_slots
+        else:
+            # only manage slots for replicas that replicate from this one, except for the leader among them
+            slot_members = [m.name for m in self.members if use_slots and
+                            m.replicatefrom == my_name and m.name != self.leader_name]
+            permanent_slots = self.__permanent_logical_slots if use_slots and not nofailover else {}
+
+        slots = {slot_name_from_member_name(name): {'type': 'physical'} for name in slot_members}
+
+        if len(slots) < len(slot_members):
+            # Find which names are conflicting for a nicer error message
+            slot_conflicts = defaultdict(list)
+            for name in slot_members:
+                slot_conflicts[slot_name_from_member_name(name)].append(name)
+            logger.error("Following cluster members share a replication slot name: %s",
+                         "; ".join("{} map to {}".format(", ".join(v), k)
+                                   for k, v in slot_conflicts.items() if len(v) > 1))
+
+        # "merge" replication slots for members with permanent_replication_slots
+        disabled_permanent_logical_slots = []
+        for name, value in permanent_slots.items():
+            if not slot_name_re.match(name):
+                logger.error("Invalid permanent replication slot name '%s'", name)
+                logger.error("Slot name may only contain lower case letters, numbers, and the underscore chars")
+                continue
+
+            value = deepcopy(value) if value else {'type': 'physical'}
+            if isinstance(value, dict):
+                if 'type' not in value:
+                    value['type'] = 'logical' if value.get('database') and value.get('plugin') else 'physical'
+
+                if value['type'] == 'physical':
+                    # Don't try to create permanent physical replication slot for yourself
+                    if name != slot_name_from_member_name(my_name):
+                        slots[name] = value
+                    continue
+                elif value['type'] == 'logical' and value.get('database') and value.get('plugin'):
+                    if major_version < 110000:
+                        disabled_permanent_logical_slots.append(name)
+                    elif name in slots:
+                        logger.error("Permanent logical replication slot {'%s': %s} is conflicting with" +
+                                     " physical replication slot for cluster member", name, value)
+                    else:
+                        slots[name] = value
+                    continue
+
+            logger.error("Bad value for slot '%s' in permanent_slots: %s", name, permanent_slots[name])
+
+        if disabled_permanent_logical_slots and show_error:
+            logger.error("Permanent logical replication slots supported by Patroni only starting from PostgreSQL 11. "
+                         "Following slots will not be created: %s.", disabled_permanent_logical_slots)
+
+        return slots
+
+    def has_permanent_logical_slots(self, my_name, nofailover, major_version=110000):
+        if major_version < 110000:
+            return False
+        slots = self.get_replication_slots(my_name, 'replica', nofailover, major_version).values()
+        return any(v for v in slots if v.get("type") == "logical")
+
+    def should_enforce_hot_standby_feedback(self, my_name, nofailover, major_version):
+        """
+        The hot_standby_feedback must be enabled if the current replica has logical slots
+        or it is working as a cascading replica for the other node that has logical slots.
+        """
+
+        if major_version < 110000:
+            return False
+
+        if self.has_permanent_logical_slots(my_name, nofailover, major_version):
+            return True
+
+        if self.use_slots:
+            members = [m for m in self.members if m.replicatefrom == my_name and m.name != self.leader_name]
+            return any(self.should_enforce_hot_standby_feedback(m.name, m.nofailover, major_version) for m in members)
+        return False
+
+    def get_my_slot_name_on_primary(self, my_name, replicatefrom):
+        """
+        P <-- I <-- L
+        In case of cascading replication we have to check not our physical slot,
+        but slot of the replica that connects us to the primary.
+        """
+
+        m = self.get_member(replicatefrom, False) if replicatefrom else None
+        return self.get_my_slot_name_on_primary(m.name, m.replicatefrom) if m else slot_name_from_member_name(my_name)
+
+    @property
+    def timeline(self):
+        """
+        >>> Cluster(0, 0, 0, 0, 0, 0, 0, 0, 0, None).timeline
+        0
+        >>> Cluster(0, 0, 0, 0, 0, 0, 0, TimelineHistory.from_node(1, '[]'), 0, None).timeline
+        1
+        >>> Cluster(0, 0, 0, 0, 0, 0, 0, TimelineHistory.from_node(1, '[["a"]]'), 0, None).timeline
+        0
+        """
+        if self.history:
+            if self.history.lines:
+                try:
+                    return int(self.history.lines[-1][0]) + 1
+                except Exception:
+                    logger.error('Failed to parse cluster history from DCS: %s', self.history.lines)
+            elif self.history.value == '[]':
+                return 1
+        return 0
+
+    @property
+    def min_version(self):
+        return next(iter(sorted(filter(lambda v: v, [m.version for m in self.members])) + [None]))
+
+
+class ReturnFalseException(Exception):
+    pass
+
+
+def catch_return_false_exception(func):
+    def wrapper(*args, **kwargs):
+        try:
+            return func(*args, **kwargs)
+        except ReturnFalseException:
+            return False
+
+    return wrapper
 
 
 @six.add_metaclass(abc.ABCMeta)
@@ -321,10 +661,13 @@ class AbstractDCS(object):
     _CONFIG = 'config'
     _LEADER = 'leader'
     _FAILOVER = 'failover'
+    _HISTORY = 'history'
     _MEMBERS = 'members/'
     _OPTIME = 'optime'
-    _LEADER_OPTIME = _OPTIME + '/' + _LEADER
+    _STATUS = 'status'  # JSON, contains "leader_lsn" and confirmed_flush_lsn of logical "slots" on the leader
+    _LEADER_OPTIME = _OPTIME + '/' + _LEADER  # legacy
     _SYNC = 'sync'
+    _FAILSAFE = 'failsafe'
 
     def __init__(self, config):
         """
@@ -332,18 +675,26 @@ class AbstractDCS(object):
             i.e.: `zookeeper` for zookeeper, `etcd` for etcd, etc...
         """
         self._name = config['name']
-        self._namespace = '/{0}'.format(config.get('namespace', '/service/').strip('/'))
-        self._base_path = '/'.join([self._namespace, config['scope']])
+        self._base_path = re.sub('/+', '/', '/'.join(['', config.get('namespace', 'service'), config['scope']]))
+        self._citus_group = str(config['group']) if isinstance(config.get('group'), six.integer_types) else None
         self._set_loop_wait(config.get('loop_wait', 10))
 
         self._ctl = bool(config.get('patronictl', False))
         self._cluster = None
+        self._cluster_valid_till = 0
         self._cluster_thread_lock = Lock()
-        self._last_leader_operation = ''
+        self._last_lsn = ''
+        self._last_seen = 0
+        self._last_status = {}
+        self._last_failsafe = {}
         self.event = Event()
 
     def client_path(self, path):
-        return '/'.join([self._base_path, path.lstrip('/')])
+        components = [self._base_path]
+        if self._citus_group:
+            components.append(self._citus_group)
+        components.append(path.lstrip('/'))
+        return '/'.join(components)
 
     @property
     def initialize_path(self):
@@ -369,6 +720,14 @@ class AbstractDCS(object):
     def failover_path(self):
         return self.client_path(self._FAILOVER)
 
+    @property
+    def history_path(self):
+        return self.client_path(self._HISTORY)
+
+    @property
+    def status_path(self):
+        return self.client_path(self._STATUS)
+
     @property
     def leader_optime_path(self):
         return self.client_path(self._LEADER_OPTIME)
@@ -377,10 +736,18 @@ class AbstractDCS(object):
     def sync_path(self):
         return self.client_path(self._SYNC)
 
+    @property
+    def failsafe_path(self):
+        return self.client_path(self._FAILSAFE)
+
     @abc.abstractmethod
     def set_ttl(self, ttl):
         """Set the new ttl value for leader key"""
 
+    @abc.abstractmethod
+    def ttl(self):
+        """Get new ttl value"""
+
     @abc.abstractmethod
     def set_retry_timeout(self, retry_timeout):
         """Set the new value for retry_timeout"""
@@ -397,64 +764,173 @@ class AbstractDCS(object):
     def loop_wait(self):
         return self._loop_wait
 
+    @property
+    def last_seen(self):
+        return self._last_seen
+
     @abc.abstractmethod
-    def _load_cluster(self):
-        """Internally this method should build  `Cluster` object which
-           represents current state and topology of the cluster in DCS.
-           this method supposed to be called only by `get_cluster` method.
+    def _cluster_loader(self, path):
+        """Load and build the `Cluster` object from DCS, which
+        represents a single Patroni cluster.
 
-           raise `~DCSError` in case of communication or other problems with DCS.
-           If the current node was running as a master and exception raised,
-           instance would be demoted."""
+        :param path: the path in DCS where to load Cluster(s) from.
+        :returns: `Cluster`"""
+
+    def _citus_cluster_loader(self, path):
+        """Load and build `Cluster` onjects from DCS that represent all
+        Patroni clusters from a single Citus cluster.
+
+        :param path: the path in DCS where to load Cluster(s) from.
+        :returns: all Citus groups as `dict`, with group ids as keys"""
+
+    @abc.abstractmethod
+    def _load_cluster(self, path, loader):
+        """Internally this method should call the `loader` method that
+        will build `Cluster` object which represents current state and
+        topology of the cluster in DCS. This method supposed to be
+        called only by `get_cluster` method.
+
+        :param path: the path in DCS where to load Cluster(s) from.
+        :param loader: one of `_cluster_loader` or `_citus_cluster_loader`
+        :raise: `~DCSError` in case of communication problems with DCS.
+        If the current node was running as a primary and exception
+        raised, instance would be demoted."""
+
+    def _bypass_caches(self):
+        """Used only in zookeeper"""
+
+    def is_citus_coordinator(self):
+        return self._citus_group == str(CITUS_COORDINATOR_GROUP_ID)
+
+    def get_citus_coordinator(self):
+        try:
+            path = '{0}/{1}/'.format(self._base_path, CITUS_COORDINATOR_GROUP_ID)
+            return self._load_cluster(path, self._cluster_loader)
+        except Exception as e:
+            logger.error('Failed to load Citus coordinator cluster from %s: %r', self.__class__.__name__, e)
+
+    def _get_citus_cluster(self):
+        groups = self._load_cluster(self._base_path + '/', self._citus_cluster_loader)
+        if isinstance(groups, Cluster):  # Zookeeper could return a cached version
+            cluster = groups
+        else:
+            cluster = groups.pop(CITUS_COORDINATOR_GROUP_ID,
+                                 Cluster(None, None, None, None, [], None, None, None, None, None))
+            cluster.workers.update(groups)
+        return cluster
+
+    def get_cluster(self, force=False):
+        if force:
+            self._bypass_caches()
+        try:
+            cluster = self._get_citus_cluster() if self.is_citus_coordinator()\
+                else self._load_cluster(self.client_path(''), self._cluster_loader)
+        except Exception:
+            self.reset_cluster()
+            raise
+
+        self._last_seen = int(time.time())
+        self._last_status = {self._OPTIME: cluster.last_lsn, 'slots': cluster.slots}
+        self._last_failsafe = cluster.failsafe
 
-    def get_cluster(self):
         with self._cluster_thread_lock:
-            try:
-                self._load_cluster()
-            except:
-                self._cluster = None
-                raise
-            return self._cluster
+            self._cluster = cluster
+            self._cluster_valid_till = time.time() + self.ttl
+            return cluster
 
     @property
     def cluster(self):
         with self._cluster_thread_lock:
-            return self._cluster
+            return self._cluster if self._cluster_valid_till > time.time() else None
 
     def reset_cluster(self):
         with self._cluster_thread_lock:
             self._cluster = None
+            self._cluster_valid_till = 0
 
     @abc.abstractmethod
-    def _write_leader_optime(self, last_operation):
-        """write current xlog location into `/optime/leader` key in DCS
-        :param last_operation: absolute xlog location in bytes
+    def _write_leader_optime(self, last_lsn):
+        """write current WAL LSN into `/optime/leader` key in DCS
+
+        :param last_lsn: absolute WAL LSN in bytes
         :returns: `!True` on success."""
 
-    def write_leader_optime(self, last_operation):
-        if self._last_leader_operation != last_operation and self._write_leader_optime(last_operation):
-            self._last_leader_operation = last_operation
+    def write_leader_optime(self, last_lsn):
+        self.write_status({self._OPTIME: last_lsn})
 
     @abc.abstractmethod
-    def update_leader(self):
+    def _write_status(self, value):
+        """write current WAL LSN and confirmed_flush_lsn of permanent slots into the `/status` key in DCS
+
+        :param value: status serialized in JSON forman
+        :returns: `!True` on success."""
+
+    def write_status(self, value):
+        if not deep_compare(self._last_status, value) and self._write_status(json.dumps(value, separators=(',', ':'))):
+            self._last_status = value
+        cluster = self.cluster
+        min_version = cluster and cluster.min_version
+        if min_version and min_version < (2, 1, 0) and self._last_lsn != value[self._OPTIME]:
+            self._last_lsn = value[self._OPTIME]
+            self._write_leader_optime(str(value[self._OPTIME]))
+
+    @abc.abstractmethod
+    def _write_failsafe(self, value):
+        """Write current cluster topology to DCS that will be used by failsafe mechanism (if enabled).
+
+        :param value: failsafe topology serialized in JSON format
+        :returns: `!True` on success."""
+
+    def write_failsafe(self, value):
+        if not (isinstance(self._last_failsafe, dict) and deep_compare(self._last_failsafe, value))\
+                and self._write_failsafe(json.dumps(value, separators=(',', ':'))):
+            self._last_failsafe = value
+
+    @property
+    def failsafe(self):
+        return self._last_failsafe
+
+    @abc.abstractmethod
+    def _update_leader(self):
         """Update leader key (or session) ttl
 
         :returns: `!True` if leader key (or session) has been updated successfully.
-            If not, `!False` must be returned and current instance would be demoted.
 
         You have to use CAS (Compare And Swap) operation in order to update leader key,
-        for example for etcd `prevValue` parameter must be used."""
+        for example for etcd `prevValue` parameter must be used.
+        If update fails due to DCS not being accessible or because it is not able to
+        process requests (hopefuly temporary), the ~DCSError exception should be raised."""
+
+    def update_leader(self, last_lsn, slots=None, failsafe=None):
+        """Update leader key (or session) ttl and optime/leader
+
+        :param last_lsn: absolute WAL LSN in bytes
+        :param slots: dict with permanent slots confirmed_flush_lsn
+        :returns: `!True` if leader key (or session) has been updated successfully."""
+
+        ret = self._update_leader()
+        if ret and last_lsn:
+            status = {self._OPTIME: last_lsn}
+            if slots:
+                status['slots'] = slots
+            self.write_status(status)
+
+        if ret and failsafe is not None:
+            self.write_failsafe(failsafe)
+
+        return ret
 
     @abc.abstractmethod
-    def attempt_to_acquire_leader(self, permanent=False):
+    def attempt_to_acquire_leader(self):
         """Attempt to acquire leader lock
         This method should create `/leader` key with value=`~self._name`
-        :param permanent: if set to `!True`, the leader key will never expire.
-         Used in patronictl for the external master
         :returns: `!True` if key has been created successfully.
 
         Key must be created atomically. In case if key already exists it should not be
-        overwritten and `!False` must be returned"""
+        overwritten and `!False` must be returned.
+
+        If key creation fails due to DCS not being accessible or because it is not able to
+        process requests (hopefuly temporary), the ~DCSError exception should be raised"""
 
     @abc.abstractmethod
     def set_failover_value(self, value, index=None):
@@ -470,7 +946,6 @@ class AbstractDCS(object):
 
         if scheduled_at:
             failover_value['scheduled_at'] = scheduled_at.isoformat()
-
         return self.set_failover_value(json.dumps(failover_value, separators=(',', ':')), index)
 
     @abc.abstractmethod
@@ -478,15 +953,13 @@ class AbstractDCS(object):
         """Create or update `/config` key"""
 
     @abc.abstractmethod
-    def touch_member(self, data, ttl=None, permanent=False):
+    def touch_member(self, data):
         """Update member key in DCS.
         This method should create or update key with the name = '/members/' + `~self._name`
         and value = data in a given DCS.
 
-        :param data: json serialized information about instance (including connection strings)
+        :param data: information about instance (including connection strings)
         :param ttl: ttl for member key, optional parameter. If it is None `~self.member_ttl will be used`
-        :param permanent: if set to `!True`, the member key will never expire.
-         Used in patronictl for the external master.
         :returns: `!True` on success otherwise `!False`
         """
 
@@ -508,10 +981,19 @@ class AbstractDCS(object):
         otherwise it should return `!False`"""
 
     @abc.abstractmethod
-    def delete_leader(self):
-        """Voluntarily remove leader key from DCS
+    def _delete_leader(self):
+        """Remove leader key from DCS.
         This method should remove leader key if current instance is the leader"""
 
+    def delete_leader(self, last_lsn=None):
+        """Update optime/leader and voluntarily remove leader key from DCS.
+        This method should remove leader key if current instance is the leader.
+        :param last_lsn: latest checkpoint location in bytes"""
+
+        if last_lsn:
+            self.write_status({self._OPTIME: last_lsn})
+        return self._delete_leader()
+
     @abc.abstractmethod
     def cancel_initialization(self):
         """ Removes the initialize key for a cluster """
@@ -520,8 +1002,20 @@ class AbstractDCS(object):
     def delete_cluster(self):
         """Delete cluster from DCS"""
 
+    @staticmethod
+    def sync_state(leader, sync_standby):
+        """Build sync_state dict
+           sync_standby dictionary key being kept for backward compatibility
+        """
+        return {'leader': leader, 'sync_standby': sync_standby and ','.join(sorted(sync_standby)) or None}
+
     def write_sync_state(self, leader, sync_standby, index=None):
-        return self.set_sync_state_value(json.dumps({'leader': leader, 'sync_standby': sync_standby}), index=index)
+        sync_value = self.sync_state(leader, sync_standby)
+        return self.set_sync_state_value(json.dumps(sync_value, separators=(',', ':')), index)
+
+    @abc.abstractmethod
+    def set_history_value(self, value):
+        """"""
 
     @abc.abstractmethod
     def set_sync_state_value(self, value, index=None):
@@ -532,7 +1026,7 @@ class AbstractDCS(object):
         """"""
 
     def watch(self, leader_index, timeout):
-        """If the current node is a master it should just sleep.
+        """If the current node is a leader it should just sleep.
         Any other node should watch for changes of leader key with a given timeout
 
         :param leader_index: index of a leader key
@@ -540,4 +1034,4 @@ class AbstractDCS(object):
         :returns: `!True` if you would like to reschedule the next run of ha cycle"""
 
         self.event.wait(timeout)
-        return self.event.isSet()
+        return self.event.is_set()

patroni/dcs/consul.py

@@ -1,18 +1,24 @@
 from __future__ import absolute_import
+import json
 import logging
 import os
+import re
 import socket
+import ssl
 import time
 import urllib3
 
+from collections import defaultdict, namedtuple
 from consul import ConsulException, NotFound, base
-from patroni.dcs import AbstractDCS, ClusterConfig, Cluster, Failover, Leader, Member, SyncState
-from patroni.exceptions import DCSError
-from patroni.utils import Retry, RetryFailedError
 from urllib3.exceptions import HTTPError
-from six.moves.urllib.parse import urlencode
+from six.moves.urllib.parse import urlencode, urlparse, quote
 from six.moves.http_client import HTTPException
 
+from . import AbstractDCS, Cluster, ClusterConfig, Failover, Leader, Member, SyncState,\
+        TimelineHistory, ReturnFalseException, catch_return_false_exception, citus_group_re
+from ..exceptions import DCSError
+from ..utils import deep_compare, parse_bool, Retry, RetryFailedError, split_host_port, uri, USER_AGENT
+
 logger = logging.getLogger(__name__)
 
 
@@ -24,21 +30,45 @@ class ConsulInternalError(ConsulException):
     """An internal Consul server error occurred"""
 
 
+class InvalidSessionTTL(ConsulException):
+    """Session TTL is too small or too big"""
+
+
+class InvalidSession(ConsulException):
+    """invalid session"""
+
+
+Response = namedtuple('Response', 'code,headers,body,content')
+
+
 class HTTPClient(object):
 
-    def __init__(self, host='127.0.0.1', port=8500, scheme='http', verify=True, timeout=10):
-        self.host = host
-        self.port = port
-        self.scheme = scheme
-        self.verify = verify
-        self.set_read_timeout(timeout)
-        self.base_uri = '{0}://{1}:{2}'.format(self.scheme, self.host, self.port)
-        self.http = urllib3.PoolManager(num_pools=10)
+    def __init__(self, host='127.0.0.1', port=8500, token=None, scheme='http', verify=True, cert=None, ca_cert=None):
+        self.token = token
+        self._read_timeout = 10
+        self.base_uri = uri(scheme, (host, port))
+        kwargs = {}
+        if cert:
+            if isinstance(cert, tuple):
+                # Key and cert are separate
+                kwargs['cert_file'] = cert[0]
+                kwargs['key_file'] = cert[1]
+            else:
+                # combined certificate
+                kwargs['cert_file'] = cert
+        if ca_cert:
+            kwargs['ca_certs'] = ca_cert
+        kwargs['cert_reqs'] = ssl.CERT_REQUIRED if verify or ca_cert else ssl.CERT_NONE
+        self.http = urllib3.PoolManager(num_pools=10, maxsize=10, **kwargs)
         self._ttl = None
 
     def set_read_timeout(self, timeout):
         self._read_timeout = timeout/3.0
 
+    @property
+    def ttl(self):
+        return self._ttl
+
     def set_ttl(self, ttl):
         ret = self._ttl != ttl
         self._ttl = ttl
@@ -46,10 +76,17 @@ class HTTPClient(object):
 
     @staticmethod
     def response(response):
-        data = response.data.decode('utf-8')
+        content = response.data
+        body = content.decode('utf-8')
         if response.status == 500:
-            raise ConsulInternalError('{0} {1}'.format(response.status, data))
-        return base.Response(response.status, response.headers, data)
+            msg = '{0} {1}'.format(response.status, body)
+            if body.startswith('Invalid Session TTL'):
+                raise InvalidSessionTTL(msg)
+            elif body.startswith('invalid session'):
+                raise InvalidSession(msg)
+            else:
+                raise ConsulInternalError(msg)
+        return Response(response.status, response.headers, body, content)
 
     def uri(self, path, params=None):
         return '{0}{1}{2}'.format(self.base_uri, path, params and '?' + urlencode(params) or '')
@@ -58,7 +95,7 @@ class HTTPClient(object):
         if method not in ('get', 'post', 'put', 'delete'):
             raise AttributeError("HTTPClient instance has no attribute '{0}'".format(method))
 
-        def wrapper(callback, path, params=None, data=''):
+        def wrapper(callback, path, params=None, data='', headers=None):
             # python-consul doesn't allow to specify ttl smaller then 10 seconds
             # because session_ttl_min defaults to 10s, so we have to do this ugly dirty hack...
             if method == 'put' and path == '/v1/session/create':
@@ -67,20 +104,52 @@ class HTTPClient(object):
                     data = '{' + ttl + '}'
                 else:
                     data = data[:-1] + ', ' + ttl + '}'
+            if isinstance(params, list):  # starting from v1.1.0 python-consul switched from `dict` to `list` for params
+                params = {k: v for k, v in params}
             kwargs = {'retries': 0, 'preload_content': False, 'body': data}
             if method == 'get' and isinstance(params, dict) and 'index' in params:
-                kwargs['timeout'] = (float(params['wait'][:-1]) if 'wait' in params else 300) + 1
+                timeout = float(params['wait'][:-1]) if 'wait' in params else 300
+                # According to the documentation a small random amount of additional wait time is added to the
+                # supplied maximum wait time to spread out the wake up time of any concurrent requests. This adds
+                # up to wait / 16 additional time to the maximum duration. Since our goal is actually getting a
+                # response rather read timeout we will add to the timeout a slightly bigger value.
+                kwargs['timeout'] = timeout + max(timeout/15.0, 1)
             else:
                 kwargs['timeout'] = self._read_timeout
+            kwargs['headers'] = (headers or {}).copy()
+            kwargs['headers'].update(urllib3.make_headers(user_agent=USER_AGENT))
+            token = params.pop('token', self.token) if isinstance(params, dict) else self.token
+            if token:
+                kwargs['headers']['X-Consul-Token'] = token
             return callback(self.response(self.http.request(method.upper(), self.uri(path, params), **kwargs)))
         return wrapper
 
 
 class ConsulClient(base.Consul):
 
-    @staticmethod
-    def connect(host, port, scheme, verify=True):
-        return HTTPClient(host, port, scheme, verify)
+    def __init__(self, *args, **kwargs):
+        self._cert = kwargs.pop('cert', None)
+        self._ca_cert = kwargs.pop('ca_cert', None)
+        self.token = kwargs.get('token')
+        super(ConsulClient, self).__init__(*args, **kwargs)
+
+    def http_connect(self, *args, **kwargs):
+        kwargs.update(dict(zip(['host', 'port', 'scheme', 'verify'], args)))
+        if self._cert:
+            kwargs['cert'] = self._cert
+        if self._ca_cert:
+            kwargs['ca_cert'] = self._ca_cert
+        if self.token:
+            kwargs['token'] = self.token
+        return HTTPClient(**kwargs)
+
+    def connect(self, *args, **kwargs):
+        return self.http_connect(*args, **kwargs)
+
+    def reload_config(self, config):
+        self.http.token = self.token = config.get('token')
+        self.consistency = config.get('consistency', 'default')
+        self.dc = config.get('dc')
 
 
 def catch_consul_errors(func):
@@ -92,10 +161,36 @@ def catch_consul_errors(func):
     return wrapper
 
 
+def force_if_last_failed(func):
+    def wrapper(*args, **kwargs):
+        if wrapper.last_result is False:
+            kwargs['force'] = True
+        wrapper.last_result = func(*args, **kwargs)
+        return wrapper.last_result
+
+    wrapper.last_result = None
+    return wrapper
+
+
+def service_name_from_scope_name(scope_name):
+    """Translate scope name to service name which can be used in dns.
+
+    230 = 253 - len('replica.') - len('.service.consul')
+    """
+
+    def replace_char(match):
+        c = match.group(0)
+        return '-' if c in '. _' else "u{:04d}".format(ord(c))
+
+    service_name = re.sub(r'[^a-z0-9\-]', replace_char, scope_name.lower())
+    return service_name[0:230]
+
+
 class Consul(AbstractDCS):
 
     def __init__(self, config):
         super(Consul, self).__init__(config)
+        self._base_path = self._base_path[1:]
         self._scope = config['scope']
         self._session = None
         self.__do_not_watch = False
@@ -103,14 +198,47 @@ class Consul(AbstractDCS):
                             retry_exceptions=(ConsulInternalError, HTTPException,
                                               HTTPError, socket.error, socket.timeout))
 
-        self._my_member_data = None
-        host, port = config.get('host', '127.0.0.1:8500').split(':')
-        self._client = ConsulClient(host=host, port=port)
+        kwargs = {}
+        if 'url' in config:
+            r = urlparse(config['url'])
+            config.update({'scheme': r.scheme, 'host': r.hostname, 'port': r.port or 8500})
+        elif 'host' in config:
+            host, port = split_host_port(config.get('host', '127.0.0.1:8500'), 8500)
+            config['host'] = host
+            if 'port' not in config:
+                config['port'] = int(port)
+
+        if config.get('cacert'):
+            config['ca_cert'] = config.pop('cacert')
+
+        if config.get('key') and config.get('cert'):
+            config['cert'] = (config['cert'], config['key'])
+
+        config_keys = ('host', 'port', 'token', 'scheme', 'cert', 'ca_cert', 'dc', 'consistency')
+        kwargs = {p: config.get(p) for p in config_keys if config.get(p)}
+
+        verify = config.get('verify')
+        if not isinstance(verify, bool):
+            verify = parse_bool(verify)
+        if isinstance(verify, bool):
+            kwargs['verify'] = verify
+
+        self._client = ConsulClient(**kwargs)
         self.set_retry_timeout(config['retry_timeout'])
         self.set_ttl(config.get('ttl') or 30)
         self._last_session_refresh = 0
+        self.__session_checks = config.get('checks', [])
+        self._register_service = config.get('register_service', False)
+        self._previous_loop_register_service = self._register_service
+        self._service_tags = sorted(config.get('service_tags', []))
+        self._previous_loop_service_tags = self._service_tags
+        if self._register_service:
+            self._set_service_name()
+        self._service_check_interval = config.get('service_check_interval', '5s')
+        self._service_check_tls_server_name = config.get('service_check_tls_server_name', None)
         if not self._ctl:
             self.create_session()
+        self._previous_loop_token = self._client.token
 
     def retry(self, *args, **kwargs):
         return self._retry.copy()(*args, **kwargs)
@@ -123,18 +251,46 @@ class Consul(AbstractDCS):
                 logger.info('waiting on consul')
                 time.sleep(5)
 
+    def reload_config(self, config):
+        super(Consul, self).reload_config(config)
+
+        consul_config = config.get('consul', {})
+        self._client.reload_config(consul_config)
+        self._previous_loop_service_tags = self._service_tags
+        self._service_tags = sorted(consul_config.get('service_tags', []))
+
+        should_register_service = consul_config.get('register_service', False)
+        if should_register_service and not self._register_service:
+            self._set_service_name()
+
+        self._previous_loop_register_service = self._register_service
+        self._register_service = should_register_service
+
     def set_ttl(self, ttl):
         if self._client.http.set_ttl(ttl/2.0):  # Consul multiplies the TTL by 2x
             self._session = None
             self.__do_not_watch = True
 
+    @property
+    def ttl(self):
+        return self._client.http.ttl * 2  # we multiply the value by 2 because it was divided in the `set_ttl()` method
+
     def set_retry_timeout(self, retry_timeout):
         self._retry.deadline = retry_timeout
         self._client.http.set_read_timeout(retry_timeout)
 
-    def _do_refresh_session(self):
+    def adjust_ttl(self):
+        try:
+            settings = self._client.agent.self()
+            min_ttl = (settings['Config']['SessionTTLMin'] or 10000000000)/1000000000.0
+            logger.warning('Changing Session TTL from %s to %s', self._client.http.ttl, min_ttl)
+            self._client.http.set_ttl(min_ttl)
+        except Exception:
+            logger.exception('adjust_ttl')
+
+    def _do_refresh_session(self, force=False):
         """:returns: `!True` if it had to create new session"""
-        if self._session and self._last_session_refresh + self._loop_wait > time.time():
+        if not force and self._session and self._last_session_refresh + self._loop_wait > time.time():
             return False
 
         if self._session:
@@ -144,8 +300,15 @@ class Consul(AbstractDCS):
                 self._session = None
         ret = not self._session
         if ret:
-            self._session = self._client.session.create(name=self._scope + '-' + self._name,
-                                                        lock_delay=0.001, behavior='delete')
+            try:
+                self._session = self._client.session.create(name=self._scope + '-' + self._name,
+                                                            checks=self.__session_checks,
+                                                            lock_delay=0.001, behavior='delete')
+            except InvalidSessionTTL:
+                logger.exception('session.create')
+                self.adjust_ttl()
+                raise
+
         self._last_session_refresh = time.time()
         return ret
 
@@ -156,103 +319,239 @@ class Consul(AbstractDCS):
             logger.exception('refresh_session')
         raise ConsulError('Failed to renew/create session')
 
-    def client_path(self, path):
-        return super(Consul, self).client_path(path)[1:]
-
     @staticmethod
     def member(node):
         return Member.from_node(node['ModifyIndex'], os.path.basename(node['Key']), node.get('Session'), node['Value'])
 
-    def _load_cluster(self):
+    def _cluster_from_nodes(self, nodes):
+        # get initialize flag
+        initialize = nodes.get(self._INITIALIZE)
+        initialize = initialize and initialize['Value']
+
+        # get global dynamic configuration
+        config = nodes.get(self._CONFIG)
+        config = config and ClusterConfig.from_node(config['ModifyIndex'], config['Value'])
+
+        # get timeline history
+        history = nodes.get(self._HISTORY)
+        history = history and TimelineHistory.from_node(history['ModifyIndex'], history['Value'])
+
+        # get last known leader lsn and slots
+        status = nodes.get(self._STATUS)
+        if status:
+            try:
+                status = json.loads(status['Value'])
+                last_lsn = status.get(self._OPTIME)
+                slots = status.get('slots')
+            except Exception:
+                slots = last_lsn = None
+        else:
+            last_lsn = nodes.get(self._LEADER_OPTIME)
+            last_lsn = last_lsn and last_lsn['Value']
+            slots = None
+
         try:
-            path = self.client_path('/')
-            _, results = self.retry(self._client.kv.get, path, recurse=True)
+            last_lsn = int(last_lsn)
+        except Exception:
+            last_lsn = 0
 
-            if results is None:
-                raise NotFound
+        # get list of members
+        members = [self.member(n) for k, n in nodes.items() if k.startswith(self._MEMBERS) and k.count('/') == 1]
 
-            nodes = {}
-            for node in results:
+        # get leader
+        leader = nodes.get(self._LEADER)
+
+        if leader:
+            member = Member(-1, leader['Value'], None, {})
+            member = ([m for m in members if m.name == leader['Value']] or [member])[0]
+            leader = Leader(leader['ModifyIndex'], leader.get('Session'), member)
+
+        # failover key
+        failover = nodes.get(self._FAILOVER)
+        if failover:
+            failover = Failover.from_node(failover['ModifyIndex'], failover['Value'])
+
+        # get synchronization state
+        sync = nodes.get(self._SYNC)
+        sync = SyncState.from_node(sync and sync['ModifyIndex'], sync and sync['Value'])
+
+        # get failsafe topology
+        failsafe = nodes.get(self._FAILSAFE)
+        try:
+            failsafe = json.loads(failsafe['Value']) if failsafe else None
+        except Exception:
+            failsafe = None
+
+        return Cluster(initialize, config, leader, last_lsn, members, failover, sync, history, slots, failsafe)
+
+    def _cluster_loader(self, path):
+        _, results = self.retry(self._client.kv.get, path, recurse=True)
+        if results is None:
+            raise NotFound
+        nodes = {}
+        for node in results:
+            node['Value'] = (node['Value'] or b'').decode('utf-8')
+            nodes[node['Key'][len(path):]] = node
+
+        return self._cluster_from_nodes(nodes)
+
+    def _citus_cluster_loader(self, path):
+        _, results = self.retry(self._client.kv.get, path, recurse=True)
+        clusters = defaultdict(dict)
+        for node in results or []:
+            key = node['Key'][len(path):].split('/', 1)
+            if len(key) == 2 and citus_group_re.match(key[0]):
                 node['Value'] = (node['Value'] or b'').decode('utf-8')
-                nodes[os.path.relpath(node['Key'], path)] = node
+                clusters[int(key[0])][key[1]] = node
+        return {group: self._cluster_from_nodes(nodes) for group, nodes in clusters.items()}
 
-            # get initialize flag
-            initialize = nodes.get(self._INITIALIZE)
-            initialize = initialize and initialize['Value']
-
-            # get global dynamic configuration
-            config = nodes.get(self._CONFIG)
-            config = config and ClusterConfig.from_node(config['ModifyIndex'], config['Value'])
-
-            # get last leader operation
-            last_leader_operation = nodes.get(self._LEADER_OPTIME)
-            last_leader_operation = 0 if last_leader_operation is None else int(last_leader_operation['Value'])
-
-            # get list of members
-            members = [self.member(n) for k, n in nodes.items() if k.startswith(self._MEMBERS) and k.count('/') == 1]
-
-            # get leader
-            leader = nodes.get(self._LEADER)
-            if not self._ctl and leader and leader['Value'] == self._name \
-                    and self._session != leader.get('Session', 'x'):
-                logger.info('I am leader but not owner of the session. Removing leader node')
-                self._client.kv.delete(self.leader_path, cas=leader['ModifyIndex'])
-                leader = None
-
-            if leader:
-                member = Member(-1, leader['Value'], None, {})
-                member = ([m for m in members if m.name == leader['Value']] or [member])[0]
-                leader = Leader(leader['ModifyIndex'], leader.get('Session'), member)
-
-            # failover key
-            failover = nodes.get(self._FAILOVER)
-            if failover:
-                failover = Failover.from_node(failover['ModifyIndex'], failover['Value'])
-
-            # get synchronization state
-            sync = nodes.get(self._SYNC)
-            sync = SyncState.from_node(sync and sync['ModifyIndex'], sync and sync['Value'])
-
-            self._cluster = Cluster(initialize, config, leader, last_leader_operation, members, failover, sync)
+    def _load_cluster(self, path, loader):
+        try:
+            return loader(path)
         except NotFound:
-            self._cluster = Cluster(None, None, None, None, [], None, None)
-        except:
+            return Cluster(None, None, None, None, [], None, None, None, None, None)
+        except Exception:
             logger.exception('get_cluster')
             raise ConsulError('Consul is not responding properly')
 
-    def touch_member(self, data, **kwargs):
+    @catch_consul_errors
+    def touch_member(self, data):
         cluster = self.cluster
         member = cluster and cluster.get_member(self._name, fallback_to_leader=False)
-        create_member = self.refresh_session()
+
+        try:
+            create_member = self.refresh_session()
+        except DCSError:
+            return False
 
         if member and (create_member or member.session != self._session):
-            try:
-                self._client.kv.delete(self.member_path)
-                create_member = True
-            except Exception:
-                return False
+            self._client.kv.delete(self.member_path)
+            create_member = True
 
-        if not create_member and member and data == self._my_member_data:
+        if self._register_service or self._previous_loop_register_service:
+            try:
+                self.update_service(not create_member and member and member.data or {}, data)
+            except Exception:
+                logger.exception('update_service')
+
+        if not create_member and member and deep_compare(data, member.data):
             return True
 
         try:
-            args = {} if kwargs.get('permanent', False) else {'acquire': self._session}
-            self._client.kv.put(self.member_path, data, **args)
-            self._my_member_data = data
+            self._client.kv.put(self.member_path, json.dumps(data, separators=(',', ':')), acquire=self._session)
             return True
+        except InvalidSession:
+            self._session = None
+            logger.error('Our session disappeared from Consul, can not "touch_member"')
         except Exception:
             logger.exception('touch_member')
         return False
 
-    @catch_consul_errors
-    def attempt_to_acquire_leader(self, permanent=False):
-        if not self._session and not permanent:
-            self.refresh_session()
+    def _set_service_name(self):
+        self._service_name = service_name_from_scope_name(self._scope)
+        if self._scope != self._service_name:
+            logger.warning('Using %s as consul service name instead of scope name %s', self._service_name, self._scope)
 
-        args = {} if permanent else {'acquire': self._session}
-        ret = self.retry(self._client.kv.put, self.leader_path, self._name, **args)
+    @catch_consul_errors
+    def register_service(self, service_name, **kwargs):
+        logger.info('Register service %s, params %s', service_name, kwargs)
+        return self._client.agent.service.register(service_name, **kwargs)
+
+    @catch_consul_errors
+    def deregister_service(self, service_id):
+        logger.info('Deregister service %s', service_id)
+        # service_id can contain special characters, but is used as part of uri in deregister request
+        service_id = quote(service_id)
+        return self._client.agent.service.deregister(service_id)
+
+    def _update_service(self, data):
+        service_name = self._service_name
+        role = data['role'].replace('_', '-')
+        state = data['state']
+        api_parts = urlparse(data['api_url'])
+        api_parts = api_parts._replace(path='/{0}'.format(role))
+        conn_parts = urlparse(data['conn_url'])
+        check = base.Check.http(api_parts.geturl(), self._service_check_interval,
+                                deregister='{0}s'.format(self._client.http.ttl * 10))
+        if self._service_check_tls_server_name is not None:
+            check['TLSServerName'] = self._service_check_tls_server_name
+        tags = self._service_tags[:]
+        tags.append(role)
+        if role == 'master':
+            tags.append('primary')
+        elif role == 'primary':
+            tags.append('master')
+        self._previous_loop_service_tags = self._service_tags
+        self._previous_loop_token = self._client.token
+
+        params = {
+            'service_id': '{0}/{1}'.format(self._scope, self._name),
+            'address': conn_parts.hostname,
+            'port': conn_parts.port,
+            'check': check,
+            'tags': tags,
+            'enable_tag_override': True,
+        }
+
+        if state == 'stopped' or (not self._register_service and self._previous_loop_register_service):
+            self._previous_loop_register_service = self._register_service
+            return self.deregister_service(params['service_id'])
+
+        self._previous_loop_register_service = self._register_service
+        if role in ['master', 'primary', 'replica', 'standby-leader']:
+            if state != 'running':
+                return
+            return self.register_service(service_name, **params)
+
+        logger.warning('Could not register service: unknown role type %s', role)
+
+    @force_if_last_failed
+    def update_service(self, old_data, new_data, force=False):
+        update = False
+
+        for key in ['role', 'api_url', 'conn_url', 'state']:
+            if key not in new_data:
+                logger.warning('Could not register service: not enough params in member data')
+                return
+            if old_data.get(key) != new_data[key]:
+                update = True
+
+        if (
+            force or update or self._register_service != self._previous_loop_register_service
+            or self._service_tags != self._previous_loop_service_tags
+            or self._client.token != self._previous_loop_token
+        ):
+            return self._update_service(new_data)
+
+    def _do_attempt_to_acquire_leader(self, retry):
+        try:
+            return retry(self._client.kv.put, self.leader_path, self._name, acquire=self._session)
+        except InvalidSession:
+            logger.error('Our session disappeared from Consul. Will try to get a new one and retry attempt')
+            self._session = None
+            retry.deadline = retry.stoptime - time.time()
+
+            retry(self._do_refresh_session)
+
+            retry.deadline = retry.stoptime - time.time()
+            if retry.deadline < 1:
+                raise ConsulError('_do_attempt_to_acquire_leader timeout')
+
+            return retry(self._client.kv.put, self.leader_path, self._name, acquire=self._session)
+
+    @catch_return_false_exception
+    def attempt_to_acquire_leader(self):
+        retry = self._retry.copy()
+        self._run_and_handle_exceptions(self._do_refresh_session, retry=retry)
+
+        retry.deadline = retry.stoptime - time.time()
+        if retry.deadline < 1:
+            raise ConsulError('attempt_to_acquire_leader timeout')
+
+        ret = self._run_and_handle_exceptions(self._do_attempt_to_acquire_leader, retry, retry=None)
         if not ret:
             logger.info('Could not take out TTL lock')
+
         return ret
 
     def take_leader(self):
@@ -267,14 +566,50 @@ class Consul(AbstractDCS):
         return self._client.kv.put(self.config_path, value, cas=index)
 
     @catch_consul_errors
-    def _write_leader_optime(self, last_operation):
-        return self._client.kv.put(self.leader_optime_path, last_operation)
+    def _write_leader_optime(self, last_lsn):
+        return self._client.kv.put(self.leader_optime_path, last_lsn)
 
     @catch_consul_errors
-    def update_leader(self):
+    def _write_status(self, value):
+        return self._client.kv.put(self.status_path, value)
+
+    @catch_consul_errors
+    def _write_failsafe(self, value):
+        return self._client.kv.put(self.failsafe_path, value)
+
+    @staticmethod
+    def _run_and_handle_exceptions(method, *args, **kwargs):
+        retry = kwargs.pop('retry', None)
+        try:
+            return retry(method, *args, **kwargs) if retry else method(*args, **kwargs)
+        except (RetryFailedError, InvalidSession, HTTPException, HTTPError, socket.error, socket.timeout) as e:
+            raise ConsulError(e)
+        except ConsulException:
+            raise ReturnFalseException
+
+    @catch_return_false_exception
+    def _update_leader(self):
+        retry = self._retry.copy()
+
+        self._run_and_handle_exceptions(self._do_refresh_session, True, retry=retry)
+
         if self._session:
-            self.retry(self._client.session.renew, self._session)
-            self._last_session_refresh = time.time()
+            cluster = self.cluster
+            leader_session = cluster and isinstance(cluster.leader, Leader) and cluster.leader.session
+            if leader_session != self._session:
+                retry.deadline = retry.stoptime - time.time()
+                if retry.deadline < 1:
+                    raise ConsulError('update_leader timeout')
+                logger.warning('Recreating the leader key due to session mismatch')
+                if cluster.leader:
+                    self._run_and_handle_exceptions(self._client.kv.delete, self.leader_path, cas=cluster.leader.index)
+
+                retry.deadline = retry.stoptime - time.time()
+                if retry.deadline < 0.5:
+                    raise ConsulError('update_leader timeout')
+                self._run_and_handle_exceptions(self._client.kv.put, self.leader_path,
+                                                self._name, acquire=self._session)
+
         return bool(self._session)
 
     @catch_consul_errors
@@ -291,20 +626,25 @@ class Consul(AbstractDCS):
         return self.retry(self._client.kv.delete, self.client_path(''), recurse=True)
 
     @catch_consul_errors
-    def delete_leader(self):
+    def set_history_value(self, value):
+        return self._client.kv.put(self.history_path, value)
+
+    @catch_consul_errors
+    def _delete_leader(self):
         cluster = self.cluster
         if cluster and isinstance(cluster.leader, Leader) and cluster.leader.name == self._name:
             return self._client.kv.delete(self.leader_path, cas=cluster.leader.index)
 
     @catch_consul_errors
     def set_sync_state_value(self, value, index=None):
-        return self._client.kv.put(self.sync_path, value, cas=index)
+        return self.retry(self._client.kv.put, self.sync_path, value, cas=index)
 
     @catch_consul_errors
     def delete_sync_state(self, index=None):
-        return self._client.kv.delete(self.sync_path, cas=index)
+        return self.retry(self._client.kv.delete, self.sync_path, cas=index)
 
     def watch(self, leader_index, timeout):
+        self._last_session_refresh = 0
         if self.__do_not_watch:
             self.__do_not_watch = False
             return True
@@ -316,7 +656,7 @@ class Consul(AbstractDCS):
                     idx, _ = self._client.kv.get(self.leader_path, index=leader_index, wait=str(timeout) + 's')
                     return str(idx) != str(leader_index)
                 except (ConsulException, HTTPException, HTTPError, socket.error, socket.timeout):
-                    logging.exception('watch')
+                    logger.exception('watch')
 
                 timeout = end_time - time.time()
 

patroni/dcs/etcd.py

@@ -1,28 +1,39 @@
 from __future__ import absolute_import
+import abc
 import etcd
+import json
 import logging
 import os
 import urllib3.util.connection
 import random
-import requests
+import six
 import socket
 import time
 
+from collections import defaultdict
+from copy import deepcopy
 from dns.exception import DNSException
 from dns import resolver
-from patroni.dcs import AbstractDCS, ClusterConfig, Cluster, Failover, Leader, Member, SyncState
-from patroni.exceptions import DCSError
-from patroni.utils import Retry, RetryFailedError
-from urllib3.exceptions import HTTPError, ReadTimeoutError
-from requests.exceptions import RequestException
+from urllib3 import Timeout
+from urllib3.exceptions import HTTPError, ReadTimeoutError, ProtocolError
 from six.moves.queue import Queue
 from six.moves.http_client import HTTPException
 from six.moves.urllib_parse import urlparse
 from threading import Thread
 
+from . import AbstractDCS, Cluster, ClusterConfig, Failover, Leader, Member, SyncState,\
+        TimelineHistory, ReturnFalseException, catch_return_false_exception, citus_group_re
+from ..exceptions import DCSError
+from ..request import get as requests_get
+from ..utils import Retry, RetryFailedError, split_host_port, uri, USER_AGENT
+
 logger = logging.getLogger(__name__)
 
 
+class EtcdRaftInternal(etcd.EtcdException):
+    """Raft Internal Error"""
+
+
 class EtcdError(DCSError):
     pass
 
@@ -63,16 +74,20 @@ class DnsCachingResolver(Thread):
     def resolve_async(self, host, port, attempt=0):
         self._resolve_queue.put(((host, port), attempt))
 
+    def remove(self, host, port):
+        self._cache.pop((host, port), None)
+
     @staticmethod
     def _do_resolve(host, port):
         try:
             return socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM, socket.IPPROTO_TCP)
-        except socket.gaierror:
-            logger.warning('failed to resolve host %s', host)
+        except Exception as e:
+            logger.warning('failed to resolve host %s: %s', host, e)
             return []
 
 
-class Client(etcd.Client):
+@six.add_metaclass(abc.ABCMeta)
+class AbstractEtcdClientWithFailover(etcd.Client):
 
     def __init__(self, config, dns_resolver, cache_ttl=300):
         self._dns_resolver = dns_resolver
@@ -80,31 +95,81 @@ class Client(etcd.Client):
         self._machines_cache_updated = 0
         args = {p: config.get(p) for p in ('host', 'port', 'protocol', 'use_proxies', 'username', 'password',
                                            'cert', 'ca_cert') if config.get(p)}
-        super(Client, self).__init__(read_timeout=config['retry_timeout'], **args)
+        super(AbstractEtcdClientWithFailover, self).__init__(read_timeout=config['retry_timeout'], **args)
+        # For some reason python3-etcd on debian and ubuntu are not based on the latest version
+        # Workaround for the case when https://github.com/jplana/python-etcd/pull/196 is not applied
+        self.http.connection_pool_kw.pop('ssl_version', None)
         self._config = config
+        self._initial_machines_cache = []
         self._load_machines_cache()
-        self._allow_reconnect = not self._use_proxies
+        self._allow_reconnect = True
+        # allow passing retry argument to api_execute in params
+        self._comparison_conditions.add('retry')
+        self._read_options.add('retry')
+        self._del_conditions.add('retry')
 
-    def _build_request_parameters(self):
-        kwargs = {'headers': self._get_headers(), 'redirect': self.allow_redirect}
+    def _calculate_timeouts(self, etcd_nodes, timeout=None):
+        """Calculate a request timeout and number of retries per single etcd node.
+        In case if the timeout per node is too small (less than one second) we will reduce the number of nodes.
+        For the cluster with only one node we will try to do 2 retries.
+        For clusters with 2 nodes we will try to do 1 retry for every node.
+        No retries for clusters with 3 or more nodes. We better rely on switching to a different node."""
 
-        # calculate the number of retries and timeout *per node*
-        # actual number of retries depends on the number of nodes
-        etcd_nodes = len(self._machines_cache) + 1
-        kwargs['retries'] = 0 if etcd_nodes > 3 else (1 if etcd_nodes > 1 else 2)
+        per_node_timeout = timeout = float(timeout or self.read_timeout)
 
-        # if etcd_nodes > 3:
-        #     kwargs.update({'retries': 0, 'timeout': float(self.read_timeout)/etcd_nodes})
-        # elif etcd_nodes > 1:
-        #     kwargs.update({'retries': 1, 'timeout': self.read_timeout/2.0/etcd_nodes})
-        # else:
-        #     kwargs.update({'retries': 2, 'timeout': self.read_timeout/3.0})
-        kwargs['timeout'] = self.read_timeout/float(kwargs['retries'] + 1)/etcd_nodes
+        max_retries = 4 - min(etcd_nodes, 3)
+        per_node_retries = 1
+        min_timeout = 1.0
+
+        while etcd_nodes > 0:
+            per_node_timeout = float(timeout) / etcd_nodes
+            if per_node_timeout >= min_timeout:
+                # for small clusters we will try to do more than on try on every node
+                while per_node_retries < max_retries and per_node_timeout / (per_node_retries + 1) >= min_timeout:
+                    per_node_retries += 1
+                per_node_timeout /= per_node_retries
+                break
+            # if the timeout per one node is to small try to reduce number of nodes
+            etcd_nodes -= 1
+            max_retries = 1
+
+        return etcd_nodes, per_node_timeout, per_node_retries - 1
+
+    def reload_config(self, config):
+        self.username = config.get('username')
+        self.password = config.get('password')
+
+    def _get_headers(self):
+        basic_auth = ':'.join((self.username, self.password)) if self.username and self.password else None
+        return urllib3.make_headers(basic_auth=basic_auth, user_agent=USER_AGENT)
+
+    def _prepare_common_parameters(self, etcd_nodes, timeout=None):
+        kwargs = {'headers': self._get_headers(), 'redirect': self.allow_redirect, 'preload_content': False}
+
+        if timeout is not None:
+            kwargs.update(retries=0, timeout=timeout)
+        else:
+            _, per_node_timeout, per_node_retries = self._calculate_timeouts(etcd_nodes)
+            connect_timeout = max(1, per_node_timeout/2)
+            kwargs.update(timeout=Timeout(connect=connect_timeout, total=per_node_timeout), retries=per_node_retries)
         return kwargs
 
     def set_machines_cache_ttl(self, cache_ttl):
         self._machines_cache_ttl = cache_ttl
 
+    @abc.abstractmethod
+    def _prepare_get_members(self, etcd_nodes):
+        """returns: request parameters"""
+
+    @abc.abstractmethod
+    def _get_members(self, base_uri, **kwargs):
+        """returns: list of clientURLs"""
+
+    @property
+    def machines_cache(self):
+        base_uri, cache = self._base_uri, self._machines_cache
+        return ([base_uri] if base_uri in cache else []) + [machine for machine in cache if machine != base_uri]
+
     @property
     def machines(self):
         """Original `machines` method(property) of `etcd.Client` class raise exception
@@ -117,98 +182,106 @@ class Client(etcd.Client):
         Also this method implements the same timeout-retry logic as `api_execute`, because
         the original method was retrying 2 times with the `read_timeout` on each node."""
 
-        kwargs = self._build_request_parameters()
+        machines_cache = self.machines_cache
+        kwargs = self._prepare_get_members(len(machines_cache))
 
-        while True:
+        for base_uri in machines_cache:
             try:
-                response = self.http.request(self._MGET, self._base_uri + self.version_prefix + '/machines', **kwargs)
-                machines = [n.strip() for n in self._handle_server_response(response).data.decode('utf-8').split(',')]
+                machines = list(set(self._get_members(base_uri, **kwargs)))
                 logger.debug("Retrieved list of machines: %s", machines)
-                random.shuffle(machines)
-                for url in machines:
-                    r = urlparse(url)
-                    port = r.port or (443 if r.scheme == 'https' else 80)
-                    self._dns_resolver.resolve_async(r.hostname, port)
-                return machines
+                if machines:
+                    random.shuffle(machines)
+                    if not self._use_proxies:
+                        self._update_dns_cache(self._dns_resolver.resolve_async, machines)
+                    return machines
             except Exception as e:
-                # We can't get the list of machines, if one server is in the
-                # machines cache, try on it
-                logger.error("Failed to get list of machines from %s%s: %r", self._base_uri, self.version_prefix, e)
-                if self._machines_cache:
-                    self._base_uri = self._machines_cache.pop(0)
-                    logger.info("Retrying on %s", self._base_uri)
-                elif self._update_machines_cache:
-                    raise etcd.EtcdException("Could not get the list of servers, "
-                                             "maybe you provided the wrong "
-                                             "host(s) to connect to?")
-                else:
-                    return []
+                self.http.clear()
+                logger.error("Failed to get list of machines from %s%s: %r", base_uri, self.version_prefix, e)
+
+        raise etcd.EtcdConnectionFailed('No more machines in the cluster')
 
     def set_read_timeout(self, timeout):
         self._read_timeout = timeout
 
-    def _do_http_request(self, request_executor, method, url, fields=None, **kwargs):
-        try:
-            response = request_executor(method, url, fields=fields, **kwargs)
-            response.data.decode('utf-8')
-            self._check_cluster_id(response)
-        except (HTTPError, HTTPException, socket.error, socket.timeout) as e:
-            if (isinstance(fields, dict) and fields.get("wait") == "true" and
-                    isinstance(e, ReadTimeoutError)):
-                logger.debug("Watch timed out.")
-                raise etcd.EtcdWatchTimedOut("Watch timed out: {0}".format(e), cause=e)
-            logger.error("Request to server %s failed: %r", self._base_uri, e)
-            logger.info("Reconnection allowed, looking for another server.")
-            self._base_uri = self._next_server(cause=e)
-            response = False
-        return response
+    def _do_http_request(self, retry, machines_cache, request_executor, method, path, fields=None, **kwargs):
+        if fields is not None:
+            kwargs['fields'] = fields
+        some_request_failed = False
+        for i, base_uri in enumerate(machines_cache):
+            if i > 0:
+                logger.info("Retrying on %s", base_uri)
+            try:
+                response = request_executor(method, base_uri + path, **kwargs)
+                response.data.decode('utf-8')
+                if some_request_failed:
+                    self.set_base_uri(base_uri)
+                    self._refresh_machines_cache()
+                return response
+            except (HTTPError, HTTPException, socket.error, socket.timeout) as e:
+                self.http.clear()
+                if not retry:
+                    if len(machines_cache) == 1:
+                        self.set_base_uri(self._base_uri)  # trigger Etcd3 watcher restart
+                    # switch to the next etcd node because we don't know exactly what happened,
+                    # whether the key didn't received an update or there is a network problem.
+                    elif i + 1 < len(machines_cache):
+                        self.set_base_uri(machines_cache[i + 1])
+                if (isinstance(fields, dict) and fields.get("wait") == "true" and
+                        isinstance(e, (ReadTimeoutError, ProtocolError))):
+                    logger.debug("Watch timed out.")
+                    raise etcd.EtcdWatchTimedOut("Watch timed out: {0}".format(e), cause=e)
+                logger.error("Request to server %s failed: %r", base_uri, e)
+                logger.info("Reconnection allowed, looking for another server.")
+                if not retry:
+                    raise etcd.EtcdException('{0} {1} request failed'.format(method, path))
+                some_request_failed = True
+
+        raise etcd.EtcdConnectionFailed('No more machines in the cluster')
+
+    @abc.abstractmethod
+    def _prepare_request(self, kwargs, params=None, method=None):
+        """returns: request_executor"""
 
     def api_execute(self, path, method, params=None, timeout=None):
-        if not path.startswith('/'):
-            raise ValueError('Path does not start with /')
-
-        kwargs = {'fields': params, 'preload_content': False}
-
-        if method in [self._MGET, self._MDELETE]:
-            request_executor = self.http.request
-        elif method in [self._MPUT, self._MPOST]:
-            request_executor = self.http.request_encode_body
-            kwargs['encode_multipart'] = False
-        else:
-            raise etcd.EtcdException('HTTP method {0} not supported'.format(method))
+        retry = params.pop('retry', None) if isinstance(params, dict) else None
 
         # Update machines_cache if previous attempt of update has failed
         if self._update_machines_cache:
             self._load_machines_cache()
-        elif time.time() - self._machines_cache_updated > self._machines_cache_ttl:
-            self._machines_cache = self.machines
-            if self._base_uri in self._machines_cache:
-                self._machines_cache.remove(self._base_uri)
-            self._machines_cache_updated = time.time()
+        elif not self._use_proxies and time.time() - self._machines_cache_updated > self._machines_cache_ttl:
+            self._refresh_machines_cache()
 
-        kwargs.update(self._build_request_parameters())
+        machines_cache = self.machines_cache
+        etcd_nodes = len(machines_cache)
 
-        if timeout is not None:
-            kwargs.update({'retries': 0, 'timeout': timeout})
+        kwargs = self._prepare_common_parameters(etcd_nodes, timeout)
+        request_executor = self._prepare_request(kwargs, params, method)
 
-        response = False
-
-        try:
-            some_request_failed = False
-            while not response:
-                response = self._do_http_request(request_executor, method, self._base_uri + path, **kwargs)
-
-                if response is False:
-                    some_request_failed = True
-            if some_request_failed and not self._use_proxies:
-                self._machines_cache = self.machines
-                if self._base_uri in self._machines_cache:
-                    self._machines_cache.remove(self._base_uri)
-        except etcd.EtcdConnectionFailed:
-            self._update_machines_cache = True
-            if not response:
+        while True:
+            try:
+                response = self._do_http_request(retry, machines_cache, request_executor, method, path, **kwargs)
+                return self._handle_server_response(response)
+            except etcd.EtcdWatchTimedOut:
                 raise
-        return self._handle_server_response(response)
+            except etcd.EtcdConnectionFailed as ex:
+                try:
+                    if self._load_machines_cache():
+                        machines_cache = self.machines_cache
+                        etcd_nodes = len(machines_cache)
+                except Exception as e:
+                    logger.debug('Failed to update list of etcd nodes: %r', e)
+                sleeptime = retry.sleeptime
+                remaining_time = retry.stoptime - sleeptime - time.time()
+                nodes, timeout, retries = self._calculate_timeouts(etcd_nodes, remaining_time)
+                if nodes == 0:
+                    self._update_machines_cache = True
+                    self.set_base_uri(self._base_uri)  # trigger Etcd3 watcher restart
+                    raise ex
+                retry.sleep_func(sleeptime)
+                retry.update_delay()
+                # We still have some time left. Partially reduce `machines_cache` and retry request
+                kwargs.update(timeout=Timeout(connect=max(1, timeout/2), total=timeout), retries=retries)
+                machines_cache = machines_cache[:nodes]
 
     @staticmethod
     def get_srv_record(host):
@@ -217,25 +290,26 @@ class Client(etcd.Client):
         except DNSException:
             return []
 
-    def _get_machines_cache_from_srv(self, srv):
+    def _get_machines_cache_from_srv(self, srv, srv_suffix=None):
         """Fetch list of etcd-cluster member by resolving _etcd-server._tcp. SRV record.
         This record should contain list of host and peer ports which could be used to run
         'GET http://{host}:{port}/members' request (peer protocol)"""
 
         ret = []
         for r in ['-client-ssl', '-client', '-ssl', '', '-server-ssl', '-server']:
+            r = '{0}-{1}'.format(r, srv_suffix) if srv_suffix else r
             protocol = 'https' if '-ssl' in r else 'http'
             endpoint = '/members' if '-server' in r else ''
             for host, port in self.get_srv_record('_etcd{0}._tcp.{1}'.format(r, srv)):
-                url = '{0}://{1}:{2}{3}'.format(protocol, host, port, endpoint)
+                url = uri(protocol, (host, port), endpoint)
                 if endpoint:
                     try:
-                        response = requests.get(url, timeout=self.read_timeout, verify=False)
-                        if response.ok:
-                            for member in response.json():
+                        response = requests_get(url, timeout=self.read_timeout, verify=False)
+                        if response.status < 400:
+                            for member in json.loads(response.data.decode('utf-8')):
                                 ret.extend(member['clientURLs'])
                             break
-                    except RequestException:
+                    except Exception:
                         logger.exception('GET %s', url)
                 else:
                     ret.append(url)
@@ -249,15 +323,32 @@ class Client(etcd.Client):
     def _get_machines_cache_from_dns(self, host, port):
         """One host might be resolved into multiple ip addresses. We will make list out of it"""
         if self.protocol == 'http':
-            ret = []
-            for af, _, _, _, sa in self._dns_resolver.resolve(host, port):
-                host, port = sa[:2]
-                if af == socket.AF_INET6:
-                    host = '[{0}]'.format(host)
-                ret.append('{0}://{1}:{2}'.format(self.protocol, host, port))
+            ret = map(lambda res: uri(self.protocol, res[-1][:2]), self._dns_resolver.resolve(host, port))
             if ret:
                 return list(set(ret))
-        return ['{0}://{1}:{2}'.format(self.protocol, host, port)]
+        return [uri(self.protocol, (host, port))]
+
+    def _get_machines_cache_from_config(self):
+        if 'proxy' in self._config:
+            return [uri(self.protocol, (self._config['host'], self._config['port']))]
+
+        machines_cache = []
+        if 'srv' in self._config:
+            machines_cache = self._get_machines_cache_from_srv(self._config['srv'], self._config.get('srv_suffix'))
+
+        if not machines_cache and 'hosts' in self._config:
+            machines_cache = list(self._config['hosts'])
+
+        if not machines_cache and 'host' in self._config:
+            machines_cache = self._get_machines_cache_from_dns(self._config['host'], self._config['port'])
+        return machines_cache
+
+    @staticmethod
+    def _update_dns_cache(func, machines):
+        for url in machines:
+            r = urlparse(url)
+            port = r.port or (443 if r.scheme == 'https' else 80)
+            func(r.hostname, port)
 
     def _load_machines_cache(self):
         """This method should fill up `_machines_cache` from scratch.
@@ -267,50 +358,99 @@ class Client(etcd.Client):
 
         self._update_machines_cache = True
 
-        if 'srv' not in self._config and 'host' not in self._config:
-            raise Exception('Neither srv nor host url are defined in etcd section of config')
-
-        if self._use_proxies:
-            self._machines_cache = ['{0}://{1}:{2}'.format(self.protocol, self._config['host'], self._config['port'])]
-        else:
-            self._machines_cache = []
-
-            if 'srv' in self._config:
-                self._machines_cache = self._get_machines_cache_from_srv(self._config['srv'])
-
-            if not self._machines_cache and 'host' in self._config:
-                self._machines_cache = self._get_machines_cache_from_dns(self._config['host'], self._config['port'])
+        if 'srv' not in self._config and 'host' not in self._config and 'hosts' not in self._config:
+            raise Exception('Neither srv, hosts, host nor url are defined in etcd section of config')
 
+        machines_cache = self._get_machines_cache_from_config()
         # Can not bootstrap list of etcd-cluster members, giving up
-        if not self._machines_cache:
+        if not machines_cache:
             raise etcd.EtcdException
 
-        # After filling up initial list of machines_cache we should ask etcd-cluster about actual list
-        self._base_uri = self._next_server()
-        self._machines_cache = self.machines
+        # enforce resolving dns name,they might get new ips
+        self._update_dns_cache(self._dns_resolver.remove, machines_cache)
 
-        if self._base_uri in self._machines_cache:
-            self._machines_cache.remove(self._base_uri)
+        # The etcd cluster could change its topology over time and depending on how we resolve the initial
+        # topology (list of hosts in the Patroni config or DNS records, A or SRV) we might get into the situation
+        # the the real topology doesn't match anymore with the topology resolved from the configuration file.
+        # In case if the "initial" topology is the same as before we will not override the `_machines_cache`.
+        ret = set(machines_cache) != set(self._initial_machines_cache)
+        if ret:
+            self._initial_machines_cache = self._machines_cache = machines_cache
+
+            # After filling up the initial list of machines_cache we should ask etcd-cluster about actual list
+            self._refresh_machines_cache(True)
 
         self._update_machines_cache = False
+        return ret
+
+    def _refresh_machines_cache(self, updating_cache=False):
+        if self._use_proxies:
+            self._machines_cache = self._get_machines_cache_from_config()
+        else:
+            try:
+                self._machines_cache = self.machines
+            except etcd.EtcdConnectionFailed:
+                if updating_cache:
+                    raise etcd.EtcdException("Could not get the list of servers, "
+                                             "maybe you provided the wrong "
+                                             "host(s) to connect to?")
+                return
+
+        if self._base_uri not in self._machines_cache:
+            self.set_base_uri(self._machines_cache[0])
         self._machines_cache_updated = time.time()
 
+    def set_base_uri(self, value):
+        if self._base_uri != value:
+            logger.info('Selected new etcd server %s', value)
+            self._base_uri = value
 
-class Etcd(AbstractDCS):
 
-    def __init__(self, config):
-        super(Etcd, self).__init__(config)
-        self._ttl = int(config.get('ttl') or 30)
+class EtcdClient(AbstractEtcdClientWithFailover):
+
+    ERROR_CLS = EtcdError
+
+    def __del__(self):
+        if self.http is not None:
+            try:
+                self.http.clear()
+            except (ReferenceError, TypeError, AttributeError):
+                pass
+
+    def _prepare_get_members(self, etcd_nodes):
+        return self._prepare_common_parameters(etcd_nodes)
+
+    def _get_members(self, base_uri, **kwargs):
+        response = self.http.request(self._MGET, base_uri + self.version_prefix + '/machines', **kwargs)
+        data = self._handle_server_response(response).data.decode('utf-8')
+        return [m.strip() for m in data.split(',') if m.strip()]
+
+    def _prepare_request(self, kwargs, params=None, method=None):
+        kwargs['fields'] = params
+        if method in (self._MPOST, self._MPUT):
+            kwargs['encode_multipart'] = False
+        return self.http.request
+
+
+class AbstractEtcd(AbstractDCS):
+
+    def __init__(self, config, client_cls, retry_errors_cls):
+        super(AbstractEtcd, self).__init__(config)
         self._retry = Retry(deadline=config['retry_timeout'], max_delay=1, max_tries=-1,
-                            retry_exceptions=(etcd.EtcdLeaderElectionInProgress,
-                                              etcd.EtcdWatcherCleared,
-                                              etcd.EtcdEventIndexCleared))
-        self._client = self.get_etcd_client(config)
+                            retry_exceptions=retry_errors_cls)
+        self._ttl = int(config.get('ttl') or 30)
+        self._client = self.get_etcd_client(config, client_cls)
         self.__do_not_watch = False
         self._has_failed = False
 
+    def reload_config(self, config):
+        super(AbstractEtcd, self).reload_config(config)
+        self._client.reload_config(config.get(self.__class__.__name__.lower(), {}))
+
     def retry(self, *args, **kwargs):
-        return self._retry.copy()(*args, **kwargs)
+        retry = self._retry.copy()
+        kwargs['retry'] = retry
+        return retry(*args, **kwargs)
 
     def _handle_exception(self, e, name='', do_sleep=False, raise_ex=None):
         if not self._has_failed:
@@ -323,22 +463,26 @@ class Etcd(AbstractDCS):
         if isinstance(raise_ex, Exception):
             raise raise_ex
 
-    def catch_etcd_errors(func):
-        def wrapper(self, *args, **kwargs):
-            try:
-                retval = func(self, *args, **kwargs) is not None
-                self._has_failed = False
-                return retval
-            except (RetryFailedError, etcd.EtcdException) as e:
-                self._handle_exception(e)
-                return False
-            except Exception as e:
-                self._handle_exception(e, raise_ex=EtcdError('unexpected error'))
-
-        return wrapper
+    def _run_and_handle_exceptions(self, method, *args, **kwargs):
+        retry = kwargs.pop('retry', self.retry)
+        try:
+            return retry(method, *args, **kwargs) if retry else method(*args, **kwargs)
+        except (RetryFailedError, etcd.EtcdConnectionFailed) as e:
+            raise self._client.ERROR_CLS(e)
+        except etcd.EtcdException as e:
+            self._handle_exception(e)
+            raise ReturnFalseException
+        except Exception as e:
+            self._handle_exception(e, raise_ex=self._client.ERROR_CLS('unexpected error'))
 
     @staticmethod
-    def get_etcd_client(config):
+    def set_socket_options(sock, socket_options):
+        if socket_options:
+            for opt in socket_options:
+                sock.setsockopt(*opt)
+
+    def get_etcd_client(self, config, client_cls):
+        config = deepcopy(config)
         if 'proxy' in config:
             config['use_proxies'] = True
             config['url'] = config['proxy']
@@ -347,8 +491,20 @@ class Etcd(AbstractDCS):
             r = urlparse(config['url'])
             config.update({'protocol': r.scheme, 'host': r.hostname, 'port': r.port or 2379,
                            'username': r.username, 'password': r.password})
+        elif 'hosts' in config:
+            hosts = config.pop('hosts')
+            default_port = config.pop('port', 2379)
+            protocol = config.get('protocol', 'http')
+
+            if isinstance(hosts, six.string_types):
+                hosts = hosts.split(',')
+
+            config['hosts'] = []
+            for value in hosts:
+                if isinstance(value, six.string_types):
+                    config['hosts'].append(uri(protocol, split_host_port(value.strip(), default_port)))
         elif 'host' in config:
-            host, port = (config['host'] + ':2379').split(':')[:2]
+            host, port = split_host_port(config['host'], 2379)
             config['host'] = host
             if 'port' not in config:
                 config['port'] = int(port)
@@ -375,9 +531,7 @@ class Etcd(AbstractDCS):
                 sock = None
                 try:
                     sock = socket.socket(af, socktype, proto)
-                    if socket_options:
-                        for opt in socket_options:
-                            sock.setsockopt(*opt)
+                    self.set_socket_options(sock, socket_options)
                     if timeout is not socket._GLOBAL_DEFAULT_TIMEOUT:
                         sock.settimeout(timeout)
                     if source_address:
@@ -401,7 +555,9 @@ class Etcd(AbstractDCS):
         client = None
         while not client:
             try:
-                client = Client(config, dns_resolver)
+                client = client_cls(config, dns_resolver)
+                if 'use_proxies' in config and not client.machines:
+                    raise etcd.EtcdException
             except etcd.EtcdException:
                 logger.info('waiting on etcd')
                 time.sleep(5)
@@ -409,82 +565,153 @@ class Etcd(AbstractDCS):
 
     def set_ttl(self, ttl):
         ttl = int(ttl)
-        self.__do_not_watch = self._ttl != ttl
+        ret = self._ttl != ttl
         self._ttl = ttl
         self._client.set_machines_cache_ttl(ttl*10)
+        return ret
+
+    @property
+    def ttl(self):
+        return self._ttl
 
     def set_retry_timeout(self, retry_timeout):
         self._retry.deadline = retry_timeout
         self._client.set_read_timeout(retry_timeout)
 
+
+def catch_etcd_errors(func):
+    def wrapper(self, *args, **kwargs):
+        try:
+            retval = func(self, *args, **kwargs) is not None
+            self._has_failed = False
+            return retval
+        except (RetryFailedError, etcd.EtcdException) as e:
+            self._handle_exception(e)
+            return False
+        except Exception as e:
+            self._handle_exception(e, raise_ex=self._client.ERROR_CLS('unexpected error'))
+
+    return wrapper
+
+
+class Etcd(AbstractEtcd):
+
+    def __init__(self, config):
+        super(Etcd, self).__init__(config, EtcdClient, (etcd.EtcdLeaderElectionInProgress, EtcdRaftInternal))
+        self.__do_not_watch = False
+
+    def set_ttl(self, ttl):
+        self.__do_not_watch = super(Etcd, self).set_ttl(ttl)
+
     @staticmethod
     def member(node):
         return Member.from_node(node.modifiedIndex, os.path.basename(node.key), node.ttl, node.value)
 
-    def _load_cluster(self):
+    def _cluster_from_nodes(self, etcd_index, nodes):
+        # get initialize flag
+        initialize = nodes.get(self._INITIALIZE)
+        initialize = initialize and initialize.value
+
+        # get global dynamic configuration
+        config = nodes.get(self._CONFIG)
+        config = config and ClusterConfig.from_node(config.modifiedIndex, config.value)
+
+        # get timeline history
+        history = nodes.get(self._HISTORY)
+        history = history and TimelineHistory.from_node(history.modifiedIndex, history.value)
+
+        # get last know leader lsn and slots
+        status = nodes.get(self._STATUS)
+        if status:
+            try:
+                status = json.loads(status.value)
+                last_lsn = status.get(self._OPTIME)
+                slots = status.get('slots')
+            except Exception:
+                slots = last_lsn = None
+        else:
+            last_lsn = nodes.get(self._LEADER_OPTIME)
+            last_lsn = last_lsn and last_lsn.value
+            slots = None
+
         try:
-            result = self.retry(self._client.read, self.client_path(''), recursive=True)
-            nodes = {os.path.relpath(node.key, result.key): node for node in result.leaves}
+            last_lsn = int(last_lsn)
+        except Exception:
+            last_lsn = 0
 
-            # get initialize flag
-            initialize = nodes.get(self._INITIALIZE)
-            initialize = initialize and initialize.value
+        # get list of members
+        members = [self.member(n) for k, n in nodes.items() if k.startswith(self._MEMBERS) and k.count('/') == 1]
 
-            # get global dynamic configuration
-            config = nodes.get(self._CONFIG)
-            config = config and ClusterConfig.from_node(config.modifiedIndex, config.value)
+        # get leader
+        leader = nodes.get(self._LEADER)
+        if leader:
+            member = Member(-1, leader.value, None, {})
+            member = ([m for m in members if m.name == leader.value] or [member])[0]
+            index = etcd_index if etcd_index > leader.modifiedIndex else leader.modifiedIndex + 1
+            leader = Leader(index, leader.ttl, member)
 
-            # get last leader operation
-            last_leader_operation = nodes.get(self._LEADER_OPTIME)
-            last_leader_operation = 0 if last_leader_operation is None else int(last_leader_operation.value)
+        # failover key
+        failover = nodes.get(self._FAILOVER)
+        if failover:
+            failover = Failover.from_node(failover.modifiedIndex, failover.value)
 
-            # get list of members
-            members = [self.member(n) for k, n in nodes.items() if k.startswith(self._MEMBERS) and k.count('/') == 1]
+        # get synchronization state
+        sync = nodes.get(self._SYNC)
+        sync = SyncState.from_node(sync and sync.modifiedIndex, sync and sync.value)
 
-            # get leader
-            leader = nodes.get(self._LEADER)
-            if leader:
-                member = Member(-1, leader.value, None, {})
-                member = ([m for m in members if m.name == leader.value] or [member])[0]
-                index = result.etcd_index if result.etcd_index > leader.modifiedIndex else leader.modifiedIndex + 1
-                leader = Leader(index, leader.ttl, member)
+        # get failsafe topology
+        failsafe = nodes.get(self._FAILSAFE)
+        try:
+            failsafe = json.loads(failsafe.value) if failsafe else None
+        except Exception:
+            failsafe = None
 
-            # failover key
-            failover = nodes.get(self._FAILOVER)
-            if failover:
-                failover = Failover.from_node(failover.modifiedIndex, failover.value)
+        return Cluster(initialize, config, leader, last_lsn, members, failover, sync, history, slots, failsafe)
 
-            # get synchronization state
-            sync = nodes.get(self._SYNC)
-            sync = SyncState.from_node(sync and sync.modifiedIndex, sync and sync.value)
+    def _cluster_loader(self, path):
+        result = self.retry(self._client.read, path, recursive=True)
+        nodes = {node.key[len(result.key):].lstrip('/'): node for node in result.leaves}
+        return self._cluster_from_nodes(result.etcd_index, nodes)
 
-            self._cluster = Cluster(initialize, config, leader, last_leader_operation, members, failover, sync)
+    def _citus_cluster_loader(self, path):
+        clusters = defaultdict(dict)
+        result = self.retry(self._client.read, path, recursive=True)
+        for node in result.leaves:
+            key = node.key[len(result.key):].lstrip('/').split('/', 1)
+            if len(key) == 2 and citus_group_re.match(key[0]):
+                clusters[int(key[0])][key[1]] = node
+        return {group: self._cluster_from_nodes(result.etcd_index, nodes) for group, nodes in clusters.items()}
+
+    def _load_cluster(self, path, loader):
+        cluster = None
+        try:
+            cluster = loader(path)
         except etcd.EtcdKeyNotFound:
-            self._cluster = Cluster(None, None, None, None, [], None, None)
+            cluster = Cluster(None, None, None, None, [], None, None, None, None, None)
         except Exception as e:
             self._handle_exception(e, 'get_cluster', raise_ex=EtcdError('Etcd is not responding properly'))
         self._has_failed = False
+        return cluster
 
     @catch_etcd_errors
-    def touch_member(self, data, ttl=None, permanent=False):
-        return self.retry(self._client.set, self.member_path, data, None if permanent else ttl or self._ttl)
+    def touch_member(self, data):
+        data = json.dumps(data, separators=(',', ':'))
+        return self._client.set(self.member_path, data, self._ttl)
 
     @catch_etcd_errors
     def take_leader(self):
-        return self.retry(self._client.set, self.leader_path, self._name, self._ttl)
+        return self.retry(self._client.write, self.leader_path, self._name, ttl=self._ttl)
 
-    def attempt_to_acquire_leader(self, permanent=False):
+    def _do_attempt_to_acquire_leader(self):
         try:
-            return bool(self.retry(self._client.write,
-                                   self.leader_path,
-                                   self._name,
-                                   ttl=None if permanent else self._ttl,
-                                   prevExist=False))
+            return bool(self.retry(self._client.write, self.leader_path, self._name, ttl=self._ttl, prevExist=False))
         except etcd.EtcdAlreadyExist:
             logger.info('Could not take out TTL lock')
-        except (RetryFailedError, etcd.EtcdException):
-            pass
-        return False
+            return False
+
+    @catch_return_false_exception
+    def attempt_to_acquire_leader(self):
+        return self._run_and_handle_exceptions(self._do_attempt_to_acquire_leader, retry=None)
 
     @catch_etcd_errors
     def set_failover_value(self, value, index=None):
@@ -495,19 +722,34 @@ class Etcd(AbstractDCS):
         return self._client.write(self.config_path, value, prevIndex=index or 0)
 
     @catch_etcd_errors
-    def _write_leader_optime(self, last_operation):
-        return self._client.set(self.leader_optime_path, last_operation)
+    def _write_leader_optime(self, last_lsn):
+        return self._client.set(self.leader_optime_path, last_lsn)
 
     @catch_etcd_errors
-    def update_leader(self):
-        return self.retry(self._client.test_and_set, self.leader_path, self._name, self._name, self._ttl)
+    def _write_status(self, value):
+        return self._client.set(self.status_path, value)
+
+    def _do_update_leader(self):
+        try:
+            return self.retry(self._client.write, self.leader_path, self._name,
+                              prevValue=self._name, ttl=self._ttl) is not None
+        except etcd.EtcdKeyNotFound:
+            return self._do_attempt_to_acquire_leader()
+
+    @catch_etcd_errors
+    def _write_failsafe(self, value):
+        return self._client.set(self.failsafe_path, value)
+
+    @catch_return_false_exception
+    def _update_leader(self):
+        return self._run_and_handle_exceptions(self._do_update_leader, retry=None)
 
     @catch_etcd_errors
     def initialize(self, create_new=True, sysid=""):
         return self.retry(self._client.write, self.initialize_path, sysid, prevExist=(not create_new))
 
     @catch_etcd_errors
-    def delete_leader(self):
+    def _delete_leader(self):
         return self._client.delete(self.leader_path, prevValue=self._name)
 
     @catch_etcd_errors
@@ -518,9 +760,13 @@ class Etcd(AbstractDCS):
     def delete_cluster(self):
         return self.retry(self._client.delete, self.client_path(''), recursive=True)
 
+    @catch_etcd_errors
+    def set_history_value(self, value):
+        return self._client.write(self.history_path, value)
+
     @catch_etcd_errors
     def set_sync_state_value(self, value, index=None):
-        return self._client.write(self.sync_path, value, prevIndex=index or 0)
+        return self.retry(self._client.write, self.sync_path, value, prevIndex=index or 0)
 
     @catch_etcd_errors
     def delete_sync_state(self, index=None):
@@ -536,13 +782,14 @@ class Etcd(AbstractDCS):
 
             while timeout >= 1:  # when timeout is too small urllib3 doesn't have enough time to connect
                 try:
-                    self._client.watch(self.leader_path, index=leader_index, timeout=timeout + 0.5)
+                    result = self._client.watch(self.leader_path, index=leader_index, timeout=timeout + 0.5)
                     self._has_failed = False
+                    if result.action == 'compareAndSwap':
+                        time.sleep(0.01)
                     # Synchronous work of all cluster members with etcd is less expensive
                     # than reestablishing http connection every time from every replica.
                     return True
                 except etcd.EtcdWatchTimedOut:
-                    self._client.http.clear()
                     self._has_failed = False
                     return False
                 except (etcd.EtcdEventIndexCleared, etcd.EtcdWatcherCleared):  # Watch failed
@@ -557,3 +804,6 @@ class Etcd(AbstractDCS):
             return super(Etcd, self).watch(None, timeout)
         finally:
             self.event.clear()
+
+
+etcd.EtcdError.error_exceptions[300] = EtcdRaftInternal

patroni/dcs/etcd3.py

@@ -0,0 +1,891 @@
+from __future__ import absolute_import
+import base64
+import etcd
+import json
+import logging
+import os
+import six
+import socket
+import sys
+import time
+import urllib3
+
+from collections import defaultdict
+from threading import Condition, Lock, Thread
+from urllib3.exceptions import ReadTimeoutError, ProtocolError
+
+from . import ClusterConfig, Cluster, Failover, Leader, Member, SyncState,\
+        TimelineHistory, ReturnFalseException, catch_return_false_exception, citus_group_re
+from .etcd import AbstractEtcdClientWithFailover, AbstractEtcd, catch_etcd_errors
+from ..exceptions import DCSError, PatroniException
+from ..utils import deep_compare, enable_keepalive, iter_response_objects, RetryFailedError, USER_AGENT
+
+logger = logging.getLogger(__name__)
+
+
+class Etcd3Error(DCSError):
+    pass
+
+
+class UnsupportedEtcdVersion(PatroniException):
+    pass
+
+
+# google.golang.org/grpc/codes
+GRPCCode = type('Enum', (), {'OK': 0, 'Canceled': 1, 'Unknown': 2, 'InvalidArgument': 3, 'DeadlineExceeded': 4,
+                             'NotFound': 5, 'AlreadyExists': 6, 'PermissionDenied': 7, 'ResourceExhausted': 8,
+                             'FailedPrecondition': 9, 'Aborted': 10, 'OutOfRange': 11, 'Unimplemented': 12,
+                             'Internal': 13, 'Unavailable': 14, 'DataLoss': 15, 'Unauthenticated': 16})
+GRPCcodeToText = {v: k for k, v in GRPCCode.__dict__.items() if not k.startswith('__') and isinstance(v, int)}
+
+
+class Etcd3Exception(etcd.EtcdException):
+    pass
+
+
+class Etcd3ClientError(Etcd3Exception):
+
+    def __init__(self, code=None, error=None, status=None):
+        if not hasattr(self, 'error'):
+            self.error = error and error.strip()
+        self.codeText = GRPCcodeToText.get(code)
+        self.status = status
+
+    def __repr__(self):
+        return "<{0} error: '{1}', code: {2}>".format(self.__class__.__name__, self.error, self.code)
+
+    __str__ = __repr__
+
+    def as_dict(self):
+        return {'error': self.error, 'code': self.code, 'codeText': self.codeText, 'status': self.status}
+
+    @classmethod
+    def get_subclasses(cls):
+        for subclass in cls.__subclasses__():
+            for subsubclass in subclass.get_subclasses():
+                yield subsubclass
+            yield subclass
+
+
+class Unknown(Etcd3ClientError):
+    code = GRPCCode.Unknown
+
+
+class InvalidArgument(Etcd3ClientError):
+    code = GRPCCode.InvalidArgument
+
+
+class DeadlineExceeded(Etcd3ClientError):
+    code = GRPCCode.DeadlineExceeded
+    error = "context deadline exceeded"
+
+
+class NotFound(Etcd3ClientError):
+    code = GRPCCode.NotFound
+
+
+class FailedPrecondition(Etcd3ClientError):
+    code = GRPCCode.FailedPrecondition
+
+
+class Unavailable(Etcd3ClientError):
+    code = GRPCCode.Unavailable
+
+
+# https://github.com/etcd-io/etcd/commits/main/api/v3rpc/rpctypes/error.go
+class LeaseNotFound(NotFound):
+    error = "etcdserver: requested lease not found"
+
+
+class UserEmpty(InvalidArgument):
+    error = "etcdserver: user name is empty"
+
+
+class AuthFailed(InvalidArgument):
+    error = "etcdserver: authentication failed, invalid user ID or password"
+
+
+class PermissionDenied(Etcd3ClientError):
+    code = GRPCCode.PermissionDenied
+    error = "etcdserver: permission denied"
+
+
+class AuthNotEnabled(FailedPrecondition):
+    error = "etcdserver: authentication is not enabled"
+
+
+class InvalidAuthToken(Etcd3ClientError):
+    code = GRPCCode.Unauthenticated
+    error = "etcdserver: invalid auth token"
+
+
+errStringToClientError = {s.error: s for s in Etcd3ClientError.get_subclasses() if hasattr(s, 'error')}
+errCodeToClientError = {s.code: s for s in Etcd3ClientError.__subclasses__()}
+
+
+def _raise_for_data(data, status_code=None):
+    try:
+        error = data.get('error') or data.get('Error')
+        if isinstance(error, dict):  # streaming response
+            status_code = error.get('http_code')
+            code = error['grpc_code']
+            error = error['message']
+        else:
+            code = data.get('code') or data.get('Code')
+    except Exception:
+        error = str(data)
+        code = GRPCCode.Unknown
+    err = errStringToClientError.get(error) or errCodeToClientError.get(code) or Unknown
+    raise err(code, error, status_code)
+
+
+def to_bytes(v):
+    return v if isinstance(v, bytes) else v.encode('utf-8')
+
+
+def prefix_range_end(v):
+    v = bytearray(to_bytes(v))
+    for i in range(len(v) - 1, -1, -1):
+        if v[i] < 0xff:
+            v[i] += 1
+            break
+    return bytes(v)
+
+
+def base64_encode(v):
+    return base64.b64encode(to_bytes(v)).decode('utf-8')
+
+
+def base64_decode(v):
+    return base64.b64decode(v).decode('utf-8')
+
+
+def build_range_request(key, range_end=None):
+    fields = {'key': base64_encode(key)}
+    if range_end:
+        fields['range_end'] = base64_encode(range_end)
+    return fields
+
+
+class Etcd3Client(AbstractEtcdClientWithFailover):
+
+    ERROR_CLS = Etcd3Error
+
+    def __init__(self, config, dns_resolver, cache_ttl=300):
+        self._token = None
+        self._cluster_version = None
+        self.version_prefix = '/v3beta'
+        super(Etcd3Client, self).__init__(config, dns_resolver, cache_ttl)
+
+        if six.PY2:  # pragma: no cover
+            # Old grpc-gateway sometimes sends double 'transfer-encoding: chunked' headers,
+            # what breaks the old (python2.7) httplib.HTTPConnection (it closes the socket).
+            def dedup_addheader(httpm, key, value):
+                prev = httpm.dict.get(key)
+                if prev is None:
+                    httpm.dict[key] = value
+                elif key != 'transfer-encoding' or prev != value:
+                    combined = ", ".join((prev, value))
+                    httpm.dict[key] = combined
+
+            import httplib
+            httplib.HTTPMessage.addheader = dedup_addheader
+
+        try:
+            self.authenticate()
+        except AuthFailed as e:
+            logger.fatal('Etcd3 authentication failed: %r', e)
+            sys.exit(1)
+
+    def _get_headers(self):
+        headers = urllib3.make_headers(user_agent=USER_AGENT)
+        if self._token and self._cluster_version >= (3, 3, 0):
+            headers['authorization'] = self._token
+        return headers
+
+    def _prepare_request(self, kwargs, params=None, method=None):
+        if params is not None:
+            kwargs['body'] = json.dumps(params)
+            kwargs['headers']['Content-Type'] = 'application/json'
+        return self.http.urlopen
+
+    @staticmethod
+    def _handle_server_response(response):
+        data = response.data
+        try:
+            data = data.decode('utf-8')
+            data = json.loads(data)
+        except (TypeError, ValueError, UnicodeError) as e:
+            if response.status < 400:
+                raise etcd.EtcdException('Server response was not valid JSON: %r' % e)
+        if response.status < 400:
+            return data
+        _raise_for_data(data, response.status)
+
+    def _ensure_version_prefix(self, base_uri, **kwargs):
+        if self.version_prefix != '/v3':
+            response = self.http.urlopen(self._MGET, base_uri + '/version', **kwargs)
+            response = self._handle_server_response(response)
+
+            server_version_str = response['etcdserver']
+            server_version = tuple(int(x) for x in server_version_str.split('.'))
+            cluster_version_str = response['etcdcluster']
+            self._cluster_version = tuple(int(x) for x in cluster_version_str.split('.'))
+
+            if self._cluster_version < (3, 0) or server_version < (3, 0, 4):
+                raise UnsupportedEtcdVersion('Detected Etcd version {0} is lower than 3.0.4'.format(server_version_str))
+
+            if self._cluster_version < (3, 3):
+                if self.version_prefix != '/v3alpha':
+                    if self._cluster_version < (3, 1):
+                        logger.warning('Detected Etcd version %s is lower than 3.1.0, watches are not supported',
+                                       cluster_version_str)
+                    if self.username and self.password:
+                        logger.warning('Detected Etcd version %s is lower than 3.3.0, authentication is not supported',
+                                       cluster_version_str)
+                    self.version_prefix = '/v3alpha'
+            elif self._cluster_version < (3, 4):
+                self.version_prefix = '/v3beta'
+            else:
+                self.version_prefix = '/v3'
+
+    def _prepare_get_members(self, etcd_nodes):
+        kwargs = self._prepare_common_parameters(etcd_nodes)
+        self._prepare_request(kwargs, {})
+        return kwargs
+
+    def _get_members(self, base_uri, **kwargs):
+        self._ensure_version_prefix(base_uri, **kwargs)
+        resp = self.http.urlopen(self._MPOST, base_uri + self.version_prefix + '/cluster/member/list', **kwargs)
+        members = self._handle_server_response(resp)['members']
+        return set(url for member in members for url in member.get('clientURLs', []))
+
+    def call_rpc(self, method, fields, retry=None):
+        fields['retry'] = retry
+        return self.api_execute(self.version_prefix + method, self._MPOST, fields)
+
+    def authenticate(self):
+        if self._use_proxies and self._cluster_version is None:
+            kwargs = self._prepare_common_parameters(1)
+            self._ensure_version_prefix(self._base_uri, **kwargs)
+        if self._cluster_version >= (3, 3) and self.username and self.password:
+            logger.info('Trying to authenticate on Etcd...')
+            old_token, self._token = self._token, None
+            try:
+                response = self.call_rpc('/auth/authenticate', {'name': self.username, 'password': self.password})
+            except AuthNotEnabled:
+                logger.info('Etcd authentication is not enabled')
+                self._token = None
+            except Exception:
+                self._token = old_token
+                raise
+            else:
+                self._token = response.get('token')
+            return old_token != self._token
+
+    def _handle_auth_errors(func):
+        def wrapper(self, *args, **kwargs):
+            def retry(ex):
+                if self.username and self.password:
+                    self.authenticate()
+                    return func(self, *args, **kwargs)
+                else:
+                    logger.fatal('Username or password not set, authentication is not possible')
+                    raise ex
+
+            try:
+                return func(self, *args, **kwargs)
+            except (UserEmpty, PermissionDenied) as e:  # no token provided
+                # PermissionDenied is raised on 3.0 and 3.1
+                if self._cluster_version < (3, 3) and (not isinstance(e, PermissionDenied)
+                                                       or self._cluster_version < (3, 2)):
+                    raise UnsupportedEtcdVersion('Authentication is required by Etcd cluster but not '
+                                                 'supported on version lower than 3.3.0. Cluster version: '
+                                                 '{0}'.format('.'.join(map(str, self._cluster_version))))
+                return retry(e)
+            except InvalidAuthToken as e:
+                logger.error('Invalid auth token: %s', self._token)
+                return retry(e)
+
+        return wrapper
+
+    @_handle_auth_errors
+    def range(self, key, range_end=None, retry=None):
+        params = build_range_request(key, range_end)
+        params['serializable'] = True  # For better performance. We can tolerate stale reads.
+        return self.call_rpc('/kv/range', params, retry)
+
+    def prefix(self, key, retry=None):
+        return self.range(key, prefix_range_end(key), retry)
+
+    @_handle_auth_errors
+    def lease_grant(self, ttl, retry=None):
+        return self.call_rpc('/lease/grant', {'TTL': ttl}, retry)['ID']
+
+    def lease_keepalive(self, ID, retry=None):
+        return self.call_rpc('/lease/keepalive', {'ID': ID}, retry).get('result', {}).get('TTL')
+
+    def txn(self, compare, success, retry=None):
+        return self.call_rpc('/kv/txn', {'compare': [compare], 'success': [success]}, retry).get('succeeded')
+
+    @_handle_auth_errors
+    def put(self, key, value, lease=None, create_revision=None, mod_revision=None, retry=None):
+        fields = {'key': base64_encode(key), 'value': base64_encode(value)}
+        if lease:
+            fields['lease'] = lease
+        if create_revision is not None:
+            compare = {'target': 'CREATE', 'create_revision': create_revision}
+        elif mod_revision is not None:
+            compare = {'target': 'MOD', 'mod_revision': mod_revision}
+        else:
+            return self.call_rpc('/kv/put', fields, retry)
+        compare['key'] = fields['key']
+        return self.txn(compare, {'request_put': fields}, retry)
+
+    @_handle_auth_errors
+    def deleterange(self, key, range_end=None, mod_revision=None, retry=None):
+        fields = build_range_request(key, range_end)
+        if mod_revision is None:
+            return self.call_rpc('/kv/deleterange', fields, retry)
+        compare = {'target': 'MOD', 'mod_revision': mod_revision, 'key': fields['key']}
+        return self.txn(compare, {'request_delete_range': fields}, retry)
+
+    def deleteprefix(self, key, retry=None):
+        return self.deleterange(key, prefix_range_end(key), retry=retry)
+
+    def watchrange(self, key, range_end=None, start_revision=None, filters=None, read_timeout=None):
+        """returns: response object"""
+        params = build_range_request(key, range_end)
+        if start_revision is not None:
+            params['start_revision'] = start_revision
+        params['filters'] = filters or []
+        kwargs = self._prepare_common_parameters(1, self.read_timeout)
+        request_executor = self._prepare_request(kwargs, {'create_request': params})
+        kwargs.update(timeout=urllib3.Timeout(connect=kwargs['timeout'], read=read_timeout), retries=0)
+        return request_executor(self._MPOST, self._base_uri + self.version_prefix + '/watch', **kwargs)
+
+    def watchprefix(self, key, start_revision=None, filters=None, read_timeout=None):
+        return self.watchrange(key, prefix_range_end(key), start_revision, filters, read_timeout)
+
+
+class KVCache(Thread):
+
+    def __init__(self, dcs, client):
+        Thread.__init__(self)
+        self.daemon = True
+        self._dcs = dcs
+        self._client = client
+        self.condition = Condition()
+        self._config_key = base64_encode(dcs.config_path)
+        self._leader_key = base64_encode(dcs.leader_path)
+        self._optime_key = base64_encode(dcs.leader_optime_path)
+        self._status_key = base64_encode(dcs.status_path)
+        self._name = base64_encode(dcs._name)
+        self._is_ready = False
+        self._response = None
+        self._response_lock = Lock()
+        self._object_cache = {}
+        self._object_cache_lock = Lock()
+        self.start()
+
+    def set(self, value, overwrite=False):
+        with self._object_cache_lock:
+            name = value['key']
+            old_value = self._object_cache.get(name)
+            ret = not old_value or int(old_value['mod_revision']) < int(value['mod_revision'])
+            if ret or overwrite and old_value['mod_revision'] == value['mod_revision']:
+                self._object_cache[name] = value
+        return ret, old_value
+
+    def delete(self, name, mod_revision):
+        with self._object_cache_lock:
+            old_value = self._object_cache.get(name)
+            ret = old_value and int(old_value['mod_revision']) < int(mod_revision)
+            if ret:
+                del self._object_cache[name]
+        return not old_value or ret, old_value
+
+    def copy(self):
+        with self._object_cache_lock:
+            return [v.copy() for v in self._object_cache.values()]
+
+    def get(self, name):
+        with self._object_cache_lock:
+            return self._object_cache.get(name)
+
+    def _process_event(self, event):
+        kv = event['kv']
+        key = kv['key']
+        if event.get('type') == 'DELETE':
+            success, old_value = self.delete(key, kv['mod_revision'])
+        else:
+            success, old_value = self.set(kv, True)
+
+        if success:
+            old_value = old_value and old_value.get('value')
+            new_value = kv.get('value')
+
+            value_changed = old_value != new_value and \
+                (key == self._leader_key or key in (self._optime_key, self._status_key) and new_value is not None or
+                 key == self._config_key and old_value is not None and new_value is not None)
+
+            if value_changed:
+                logger.debug('%s changed from %s to %s', key, old_value, new_value)
+
+            # We also want to wake up HA loop on replicas if leader optime (or status key) was updated
+            if value_changed and (key not in (self._optime_key, self._status_key) or
+                                  (self.get(self._leader_key) or {}).get('value') != self._name):
+                self._dcs.event.set()
+
+    def _process_message(self, message):
+        logger.debug('Received message: %s', message)
+        if 'error' in message:
+            _raise_for_data(message)
+        for event in message.get('result', {}).get('events', []):
+            self._process_event(event)
+
+    @staticmethod
+    def _finish_response(response):
+        try:
+            response.close()
+        finally:
+            response.release_conn()
+
+    def _do_watch(self, revision):
+        with self._response_lock:
+            self._response = None
+        # We do most of requests with timeouts. The only exception /watch requests to Etcd v3.
+        # In order to interrupt the /watch request we do socket.shutdown() from the main thread,
+        # which doesn't work on Windows. Therefore we want to use the last resort, `read_timeout`.
+        # Setting it to TTL will help to partially mitigate the problem.
+        # Setting it to lower value is not nice because for idling clusters it will increase
+        # the numbers of interrupts and reconnects.
+        read_timeout = self._dcs.ttl if os.name == 'nt' else None
+        response = self._client.watchprefix(self._dcs.cluster_prefix, revision, read_timeout=read_timeout)
+        with self._response_lock:
+            if self._response is None:
+                self._response = response
+
+        if not self._response:
+            return self._finish_response(response)
+
+        for message in iter_response_objects(response):
+            self._process_message(message)
+
+    def _build_cache(self):
+        result = self._dcs.retry(self._client.prefix, self._dcs.cluster_prefix)
+        with self._object_cache_lock:
+            self._object_cache = {node['key']: node for node in result.get('kvs', [])}
+        with self.condition:
+            self._is_ready = True
+            self.condition.notify()
+
+        try:
+            self._do_watch(result['header']['revision'])
+        except Exception as e:
+            # Following exceptions are expected on Windows because the /watch request  is done with `read_timeout`
+            if not (os.name == 'nt' and isinstance(e, (ReadTimeoutError, ProtocolError))):
+                logger.error('watchprefix failed: %r', e)
+        finally:
+            with self.condition:
+                self._is_ready = False
+            with self._response_lock:
+                response, self._response = self._response, None
+            if response:
+                self._finish_response(response)
+
+    def run(self):
+        while True:
+            try:
+                self._build_cache()
+            except Exception as e:
+                logger.error('KVCache.run %r', e)
+                time.sleep(1)
+
+    def kill_stream(self):
+        sock = None
+        with self._response_lock:
+            if self._response:
+                try:
+                    sock = self._response.connection.sock
+                except Exception:
+                    sock = None
+            else:
+                self._response = False
+        if sock:
+            try:
+                sock.shutdown(socket.SHUT_RDWR)
+                sock.close()
+            except Exception as e:
+                logger.debug('Error on socket.shutdown: %r', e)
+
+    def is_ready(self):
+        """Must be called only when holding the lock on `condition`"""
+        return self._is_ready
+
+
+class PatroniEtcd3Client(Etcd3Client):
+
+    def __init__(self, *args, **kwargs):
+        self._kv_cache = None
+        super(PatroniEtcd3Client, self).__init__(*args, **kwargs)
+
+    def configure(self, etcd3):
+        self._etcd3 = etcd3
+
+    def start_watcher(self):
+        if self._cluster_version >= (3, 1):
+            self._kv_cache = KVCache(self._etcd3, self)
+
+    def _restart_watcher(self):
+        if self._kv_cache:
+            self._kv_cache.kill_stream()
+
+    def set_base_uri(self, value):
+        super(PatroniEtcd3Client, self).set_base_uri(value)
+        self._restart_watcher()
+
+    def authenticate(self):
+        ret = super(PatroniEtcd3Client, self).authenticate()
+        if ret:
+            self._restart_watcher()
+        return ret
+
+    def _wait_cache(self, timeout):
+        stop_time = time.time() + timeout
+        while not self._kv_cache.is_ready():
+            timeout = stop_time - time.time()
+            if timeout <= 0:
+                raise RetryFailedError('Exceeded retry deadline')
+            self._kv_cache.condition.wait(timeout)
+
+    def get_cluster(self, path):
+        if self._kv_cache and path.startswith(self._etcd3.cluster_prefix):
+            with self._kv_cache.condition:
+                self._wait_cache(self._etcd3._retry.deadline)
+                ret = self._kv_cache.copy()
+        else:
+            ret = self._etcd3.retry(self.prefix, path).get('kvs', [])
+        for node in ret:
+            node.update({'key': base64_decode(node['key']),
+                         'value': base64_decode(node.get('value', '')),
+                         'lease': node.get('lease')})
+        return ret
+
+    def call_rpc(self, method, fields, retry=None):
+        ret = super(PatroniEtcd3Client, self).call_rpc(method, fields, retry)
+
+        if self._kv_cache:
+            value = delete = None
+            if method == '/kv/txn' and ret.get('succeeded'):
+                on_success = fields['success'][0]
+                value = on_success.get('request_put')
+                delete = on_success.get('request_delete_range')
+            elif method == '/kv/put' and ret:
+                value = fields
+            elif method == '/kv/deleterange' and ret:
+                delete = fields
+
+            if value:
+                value['mod_revision'] = ret['header']['revision']
+                self._kv_cache.set(value)
+            elif delete and 'range_end' not in delete:
+                self._kv_cache.delete(delete['key'], ret['header']['revision'])
+
+        return ret
+
+
+class Etcd3(AbstractEtcd):
+
+    def __init__(self, config):
+        super(Etcd3, self).__init__(config, PatroniEtcd3Client, (DeadlineExceeded, Unavailable, FailedPrecondition))
+        self.__do_not_watch = False
+        self._lease = None
+        self._last_lease_refresh = 0
+
+        self._client.configure(self)
+        if not self._ctl:
+            self._client.start_watcher()
+            self.create_lease()
+
+    def set_socket_options(self, sock, socket_options):
+        enable_keepalive(sock, self.ttl, int(self.loop_wait + self._retry.deadline))
+
+    def set_ttl(self, ttl):
+        self.__do_not_watch = super(Etcd3, self).set_ttl(ttl)
+        if self.__do_not_watch:
+            self._lease = None
+
+    def _do_refresh_lease(self, force=False, retry=None):
+        if not force and self._lease and self._last_lease_refresh + self._loop_wait > time.time():
+            return False
+
+        if self._lease and not self._client.lease_keepalive(self._lease, retry):
+            self._lease = None
+
+        ret = not self._lease
+        if ret:
+            self._lease = self._client.lease_grant(self._ttl, retry)
+
+        self._last_lease_refresh = time.time()
+        return ret
+
+    def refresh_lease(self):
+        try:
+            return self.retry(self._do_refresh_lease)
+        except (Etcd3ClientError, RetryFailedError):
+            logger.exception('refresh_lease')
+        raise Etcd3Error('Failed to keepalive/grant lease')
+
+    def create_lease(self):
+        while not self._lease:
+            try:
+                self.refresh_lease()
+            except Etcd3Error:
+                logger.info('waiting on etcd')
+                time.sleep(5)
+
+    @property
+    def cluster_prefix(self):
+        return self._base_path + '/' if self.is_citus_coordinator() else self.client_path('')
+
+    @staticmethod
+    def member(node):
+        return Member.from_node(node['mod_revision'], os.path.basename(node['key']), node['lease'], node['value'])
+
+    def _cluster_from_nodes(self, nodes):
+        # get initialize flag
+        initialize = nodes.get(self._INITIALIZE)
+        initialize = initialize and initialize['value']
+
+        # get global dynamic configuration
+        config = nodes.get(self._CONFIG)
+        config = config and ClusterConfig.from_node(config['mod_revision'], config['value'])
+
+        # get timeline history
+        history = nodes.get(self._HISTORY)
+        history = history and TimelineHistory.from_node(history['mod_revision'], history['value'])
+
+        # get last know leader lsn and slots
+        status = nodes.get(self._STATUS)
+        if status:
+            try:
+                status = json.loads(status['value'])
+                last_lsn = status.get(self._OPTIME)
+                slots = status.get('slots')
+            except Exception:
+                slots = last_lsn = None
+        else:
+            last_lsn = nodes.get(self._LEADER_OPTIME)
+            last_lsn = last_lsn and last_lsn['value']
+            slots = None
+
+        try:
+            last_lsn = int(last_lsn)
+        except Exception:
+            last_lsn = 0
+
+        # get list of members
+        members = [self.member(n) for k, n in nodes.items() if k.startswith(self._MEMBERS) and k.count('/') == 1]
+
+        # get leader
+        leader = nodes.get(self._LEADER)
+        if not self._ctl and leader and leader['value'] == self._name and self._lease != leader.get('lease'):
+            logger.warning('I am the leader but not owner of the lease')
+
+        if leader:
+            member = Member(-1, leader['value'], None, {})
+            member = ([m for m in members if m.name == leader['value']] or [member])[0]
+            leader = Leader(leader['mod_revision'], leader['lease'], member)
+
+        # failover key
+        failover = nodes.get(self._FAILOVER)
+        if failover:
+            failover = Failover.from_node(failover['mod_revision'], failover['value'])
+
+        # get synchronization state
+        sync = nodes.get(self._SYNC)
+        sync = SyncState.from_node(sync and sync['mod_revision'], sync and sync['value'])
+
+        # get failsafe topology
+        failsafe = nodes.get(self._FAILSAFE)
+        try:
+            failsafe = json.loads(failsafe['value']) if failsafe else None
+        except Exception:
+            failsafe = None
+
+        return Cluster(initialize, config, leader, last_lsn, members, failover, sync, history, slots, failsafe)
+
+    def _cluster_loader(self, path):
+        nodes = {node['key'][len(path):]: node
+                 for node in self._client.get_cluster(path)
+                 if node['key'].startswith(path)}
+        return self._cluster_from_nodes(nodes)
+
+    def _citus_cluster_loader(self, path):
+        clusters = defaultdict(dict)
+        path = self._base_path + '/'
+        for node in self._client.get_cluster(path):
+            key = node['key'][len(path):].split('/', 1)
+            if len(key) == 2 and citus_group_re.match(key[0]):
+                clusters[int(key[0])][key[1]] = node
+        return {group: self._cluster_from_nodes(nodes) for group, nodes in clusters.items()}
+
+    def _load_cluster(self, path, loader):
+        cluster = None
+        try:
+            cluster = loader(path)
+        except UnsupportedEtcdVersion:
+            raise
+        except Exception as e:
+            self._handle_exception(e, 'get_cluster', raise_ex=Etcd3Error('Etcd is not responding properly'))
+        self._has_failed = False
+        return cluster
+
+    @catch_etcd_errors
+    def touch_member(self, data):
+        try:
+            self.refresh_lease()
+        except Etcd3Error:
+            return False
+
+        cluster = self.cluster
+        member = cluster and cluster.get_member(self._name, fallback_to_leader=False)
+
+        if member and member.session == self._lease and deep_compare(data, member.data):
+            return True
+
+        data = json.dumps(data, separators=(',', ':'))
+        try:
+            return self._client.put(self.member_path, data, self._lease)
+        except LeaseNotFound:
+            self._lease = None
+            logger.error('Our lease disappeared from Etcd, can not "touch_member"')
+
+    @catch_etcd_errors
+    def take_leader(self):
+        return self.retry(self._client.put, self.leader_path, self._name, self._lease)
+
+    def _do_attempt_to_acquire_leader(self, retry):
+        def _retry(*args, **kwargs):
+            kwargs['retry'] = retry
+            return retry(*args, **kwargs)
+
+        try:
+            return _retry(self._client.put, self.leader_path, self._name, self._lease, 0)
+        except LeaseNotFound:
+            logger.error('Our lease disappeared from Etcd. Will try to get a new one and retry attempt')
+            self._lease = None
+            retry.deadline = retry.stoptime - time.time()
+
+            _retry(self._do_refresh_lease)
+
+            retry.deadline = retry.stoptime - time.time()
+            if retry.deadline < 1:
+                raise Etcd3Error('_do_attempt_to_acquire_leader timeout')
+
+            return _retry(self._client.put, self.leader_path, self._name, self._lease, 0)
+
+    @catch_return_false_exception
+    def attempt_to_acquire_leader(self):
+        retry = self._retry.copy()
+
+        def _retry(*args, **kwargs):
+            kwargs['retry'] = retry
+            return retry(*args, **kwargs)
+
+        self._run_and_handle_exceptions(self._do_refresh_lease, retry=_retry)
+
+        retry.deadline = retry.stoptime - time.time()
+        if retry.deadline < 1:
+            raise Etcd3Error('attempt_to_acquire_leader timeout')
+
+        ret = self._run_and_handle_exceptions(self._do_attempt_to_acquire_leader, retry, retry=None)
+        if not ret:
+            logger.info('Could not take out TTL lock')
+        return ret
+
+    @catch_etcd_errors
+    def set_failover_value(self, value, index=None):
+        return self._client.put(self.failover_path, value, mod_revision=index)
+
+    @catch_etcd_errors
+    def set_config_value(self, value, index=None):
+        return self._client.put(self.config_path, value, mod_revision=index)
+
+    @catch_etcd_errors
+    def _write_leader_optime(self, last_lsn):
+        return self._client.put(self.leader_optime_path, last_lsn)
+
+    @catch_etcd_errors
+    def _write_status(self, value):
+        return self._client.put(self.status_path, value)
+
+    @catch_etcd_errors
+    def _write_failsafe(self, value):
+        return self._client.put(self.failsafe_path, value)
+
+    @catch_return_false_exception
+    def _update_leader(self):
+        retry = self._retry.copy()
+
+        def _retry(*args, **kwargs):
+            kwargs['retry'] = retry
+            return retry(*args, **kwargs)
+
+        self._run_and_handle_exceptions(self._do_refresh_lease, True, retry=_retry)
+
+        if self._lease:
+            cluster = self.cluster
+            leader_lease = cluster and isinstance(cluster.leader, Leader) and cluster.leader.session
+            if leader_lease != self._lease:
+                retry.deadline = retry.stoptime - time.time()
+                if retry.deadline < 1:
+                    raise Etcd3Error('update_leader timeout')
+
+                try:
+                    self._run_and_handle_exceptions(self._client.put, self.leader_path,
+                                                    self._name, self._lease, retry=_retry)
+                except ReturnFalseException:
+                    pass
+        return bool(self._lease)
+
+    @catch_etcd_errors
+    def initialize(self, create_new=True, sysid=""):
+        return self.retry(self._client.put, self.initialize_path, sysid, None, 0 if create_new else None)
+
+    @catch_etcd_errors
+    def _delete_leader(self):
+        cluster = self.cluster
+        if cluster and isinstance(cluster.leader, Leader) and cluster.leader.name == self._name:
+            return self._client.deleterange(self.leader_path, mod_revision=cluster.leader.index)
+
+    @catch_etcd_errors
+    def cancel_initialization(self):
+        return self.retry(self._client.deleterange, self.initialize_path)
+
+    @catch_etcd_errors
+    def delete_cluster(self):
+        return self.retry(self._client.deleteprefix, self.client_path(''))
+
+    @catch_etcd_errors
+    def set_history_value(self, value):
+        return self._client.put(self.history_path, value)
+
+    @catch_etcd_errors
+    def set_sync_state_value(self, value, index=None):
+        return self.retry(self._client.put, self.sync_path, value, mod_revision=index)
+
+    @catch_etcd_errors
+    def delete_sync_state(self, index=None):
+        return self.retry(self._client.deleterange, self.sync_path, mod_revision=index)
+
+    def watch(self, leader_index, timeout):
+        if self.__do_not_watch:
+            self.__do_not_watch = False
+            return True
+
+        try:
+            return super(Etcd3, self).watch(None, timeout)
+        finally:
+            self.event.clear()

patroni/dcs/exhibitor.py

@@ -1,10 +1,11 @@
+import json
 import logging
 import random
-import requests
 import time
 
 from patroni.dcs.zookeeper import ZooKeeper
-from requests.exceptions import RequestException
+from patroni.request import get as requests_get
+from patroni.utils import uri
 
 logger = logging.getLogger(__name__)
 
@@ -18,7 +19,7 @@ class ExhibitorEnsembleProvider(object):
         self._uri_path = uri_path
         self._poll_interval = poll_interval
         self._exhibitors = hosts
-        self._master_exhibitors = hosts
+        self._boot_exhibitors = hosts
         self._zookeeper_hosts = ''
         self._next_poll = None
         while not self.poll():
@@ -31,7 +32,7 @@ class ExhibitorEnsembleProvider(object):
 
         json = self._query_exhibitors(self._exhibitors)
         if not json:
-            json = self._query_exhibitors(self._master_exhibitors)
+            json = self._query_exhibitors(self._boot_exhibitors)
 
         if isinstance(json, dict) and 'servers' in json and 'port' in json:
             self._next_poll = time.time() + self._poll_interval
@@ -46,12 +47,11 @@ class ExhibitorEnsembleProvider(object):
     def _query_exhibitors(self, exhibitors):
         random.shuffle(exhibitors)
         for host in exhibitors:
-            uri = 'http://{0}:{1}{2}'.format(host, self._exhibitor_port, self._uri_path)
             try:
-                response = requests.get(uri, timeout=self.TIMEOUT)
-                return response.json()
-            except RequestException:
-                pass
+                response = requests_get(uri('http', (host, self._exhibitor_port), self._uri_path), timeout=self.TIMEOUT)
+                return json.loads(response.data.decode('utf-8'))
+            except Exception:
+                logging.debug('Request to %s failed', host)
         return None
 
     @property
@@ -68,7 +68,7 @@ class Exhibitor(ZooKeeper):
         config['hosts'] = self._ensemble_provider.zookeeper_hosts
         super(Exhibitor, self).__init__(config)
 
-    def _load_cluster(self):
+    def _load_cluster(self, path, loader):
         if self._ensemble_provider.poll():
             self._client.set_hosts(self._ensemble_provider.zookeeper_hosts)
-        return super(Exhibitor, self)._load_cluster()
+        return super(Exhibitor, self)._load_cluster(path, loader)