Files
patroni/patroni/api.py
T
Ants AasmaandOleksii Kliukin 7e53a604d4 Add synchronous replication support. (#314)
Adds a new configuration variable synchronous_mode. When enabled Patroni will manage synchronous_standby_names to enable synchronous replication whenever there are healthy standbys available. With synchronous mode enabled Patroni will automatically fail over only to a standby that was synchronously replicating at the time of master failure. This effectively means zero lost user visible transactions.

To enforce the synchronous failover guarantee Patroni stores current synchronous replication state in the DCS, using strict ordering, first enable synchronous replication, then publish the information. Standby can use this to verify that it was indeed a synchronous standby before master failed and is allowed to fail over.

We can't enable multiple standbys as synchronous, allowing PostreSQL to pick one because we can't know which one was actually set to be synchronous on the master when it failed. This means that on standby failure commits will be blocked on the master until next run_cycle iteration. TODO: figure out a way to poke Patroni to run sooner or allow for PostgreSQL to pick one without the possibility of lost transactions.

On graceful shutdown standbys will disable themselves by setting a nosync tag for themselves and waiting for the master to notice and pick another standby. This adds a new mechanism for Ha to publish dynamic tags to the DCS.

When the synchronous standby goes away or disconnects a new one is picked and Patroni switches master over to the new one. If no synchronous standby exists Patroni disables synchronous replication (synchronous_standby_names=''), but not synchronous_mode. In this case, only the node that was previously master is allowed to acquire the leader lock.

Added acceptance tests and documentation.

Implementation by @ants with extensive review by @CyberDem0n.
2016-10-19 16:12:51 +02:00

487 lines
21 KiB
Python

import base64
import fcntl
import json
import logging
import psycopg2
import time
import dateutil.parser
import datetime
import pytz
from patroni.exceptions import PostgresConnectionException
from patroni.utils import deep_compare, patch_config, Retry, RetryFailedError, is_valid_pg_version
from six.moves.BaseHTTPServer import BaseHTTPRequestHandler, HTTPServer
from six.moves.socketserver import ThreadingMixIn
from threading import Thread
logger = logging.getLogger(__name__)
def check_auth(func):
"""Decorator function to check authorization header.
Usage example:
@check_auth
def do_PUT_foo():
pass
"""
def wrapper(handler):
if handler.check_auth_header():
return func(handler)
return wrapper
class RestApiHandler(BaseHTTPRequestHandler):
def _write_response(self, status_code, body, content_type='text/html', headers=None):
self.send_response(status_code)
headers = headers or {}
if content_type:
headers['Content-Type'] = content_type
for name, value in headers.items():
self.send_header(name, value)
self.end_headers()
self.wfile.write(body.encode('utf-8'))
def _write_json_response(self, status_code, response):
self._write_response(status_code, json.dumps(response), content_type='application/json')
def send_auth_request(self, body):
headers = {'WWW-Authenticate': 'Basic realm="' + self.server.patroni.__class__.__name__ + '"'}
self._write_response(401, body, headers=headers)
def check_auth_header(self):
auth_header = self.headers.get('Authorization')
status = self.server.check_auth_header(auth_header)
return not status or self.send_auth_request(status)
def _write_status_response(self, status_code, response):
patroni = self.server.patroni
tags = patroni.ha.get_effective_tags()
if tags:
response['tags'] = tags
if patroni.postgresql.sysid:
response['database_system_identifier'] = patroni.postgresql.sysid
if patroni.postgresql.pending_restart:
response['pending_restart'] = True
response['patroni'] = {'version': patroni.version, 'scope': patroni.postgresql.scope}
if patroni.scheduled_restart and isinstance(patroni.scheduled_restart, dict):
response['scheduled_restart'] = patroni.scheduled_restart.copy()
del response['scheduled_restart']['postmaster_start_time']
response['scheduled_restart']['schedule'] = (response['scheduled_restart']['schedule']).isoformat()
self._write_json_response(status_code, response)
def do_GET(self, write_status_code_only=False):
"""Default method for processing all GET requests which can not be routed to other methods"""
path = '/master' if self.path == '/' else self.path
response = self.get_postgresql_status()
patroni = self.server.patroni
cluster = patroni.dcs.cluster
if cluster: # dcs available
if cluster.leader and cluster.leader.name == patroni.postgresql.name: # is_leader
status_code = 200 if 'master' in path else 503
elif 'role' not in response:
status_code = 503
elif response['role'] == 'master': # running as master but without leader lock!!!!
status_code = 503
elif response['role'] in path: # response['role'] != 'master'
status_code = 503 if patroni.noloadbalance else 200
else:
status_code = 503
elif 'role' in response and response['role'] in path:
status_code = 503 if response['role'] != 'master' and patroni.noloadbalance else 200
elif patroni.ha.restart_scheduled() and patroni.postgresql.role == 'master' and 'master' in path:
# exceptional case for master node when the postgres is being restarted via API
status_code = 200
else:
status_code = 503
if write_status_code_only: # when haproxy sends OPTIONS request it reads only status code and nothing more
message = self.responses[status_code][0]
self.wfile.write('{0} {1} {2}\r\n'.format(self.protocol_version, status_code, message).encode('utf-8'))
else:
self._write_status_response(status_code, response)
def do_OPTIONS(self):
self.do_GET(write_status_code_only=True)
def do_GET_patroni(self):
response = self.get_postgresql_status(True)
self._write_status_response(200, response)
def do_GET_config(self):
cluster = self.server.patroni.dcs.cluster or self.server.patroni.dcs.get_cluster()
if cluster.config:
self._write_json_response(200, cluster.config.data)
else:
self.send_error(502)
def _read_json_content(self, body_is_optional=False):
if 'content-length' not in self.headers:
return self.send_error(411) if not body_is_optional else {}
try:
content_length = int(self.headers.get('content-length'))
if content_length == 0 and body_is_optional:
return {}
request = json.loads(self.rfile.read(content_length).decode('utf-8'))
if isinstance(request, dict) and (request or body_is_optional):
return request
except Exception:
logger.exception('Bad request')
self.send_error(400)
@check_auth
def do_PATCH_config(self):
request = self._read_json_content()
if request:
cluster = self.server.patroni.dcs.get_cluster()
data = cluster.config.data.copy()
if patch_config(data, request):
value = json.dumps(data, separators=(',', ':'))
if not self.server.patroni.dcs.set_config_value(value, cluster.config.index):
return self.send_error(409)
self._write_json_response(200, data)
@check_auth
def do_PUT_config(self):
request = self._read_json_content()
if request:
cluster = self.server.patroni.dcs.get_cluster()
if not deep_compare(request, cluster.config.data):
value = json.dumps(request, separators=(',', ':'))
if not self.server.patroni.dcs.set_config_value(value):
return self.send_error(502)
self._write_json_response(200, request)
@check_auth
def do_POST_reload(self):
try:
if self.server.patroni.config.reload_local_configuration(True):
status_code = 202
response = 'reload scheduled'
self.server.patroni.sighup_handler()
else:
status_code = 200
response = 'nothing changed'
except Exception as e:
status_code = 500
response = str(e)
self._write_response(status_code, response)
@staticmethod
def parse_schedule(schedule, action):
""" parses the given schedule and validates at """
error = None
scheduled_at = None
try:
scheduled_at = dateutil.parser.parse(schedule)
if scheduled_at.tzinfo is None:
error = 'Timezone information is mandatory for the scheduled {0}'.format(action)
status_code = 400
elif scheduled_at < datetime.datetime.now(pytz.utc):
error = 'Cannot schedule {0} in the past'.format(action)
status_code = 422
else:
status_code = None
except (ValueError, TypeError):
logger.exception('Invalid scheduled %s time: %s', action, schedule)
error = 'Unable to parse scheduled timestamp. It should be in an unambiguous format, e.g. ISO 8601'
status_code = 422
return (status_code, error, scheduled_at)
@check_auth
def do_POST_restart(self):
status_code = 500
data = 'restart failed'
request = self._read_json_content(body_is_optional=True)
cluster = self.server.patroni.dcs.get_cluster()
if request is None:
# failed to parse the json
return
if request:
logger.debug("received restart request: {0}".format(request))
if cluster.is_paused() and 'schedule' in request:
self._write_response(status_code, "Can't schedule restart in the paused state")
return
for k in request:
if k == 'schedule':
(_, data, request[k]) = self.parse_schedule(request[k], "restart")
if _:
status_code = _
break
elif k == 'role':
if request[k] not in ('master', 'replica'):
status_code = 400
data = "PostgreSQL role should be either master or replica"
break
elif k == 'postgres_version':
if not is_valid_pg_version(request[k]):
status_code = 400
data = "PostgreSQL version should be in the first.major.minor format"
break
elif k != 'restart_pending':
status_code = 400
data = "Unknown filter for the scheduled restart: {0}".format(k)
break
else:
if 'schedule' not in request:
try:
status, data = self.server.patroni.ha.restart(request)
status_code = 200 if status else 503
except Exception:
logger.exception('Exception during restart')
status_code = 400
else:
request['postmaster_start_time'] = self.server.patroni.ha.state_handler.postmaster_start_time()
if self.server.patroni.ha.schedule_future_restart(request):
data = "Restart scheduled"
status_code = 202
else:
data = "Another restart is already scheduled"
status_code = 409
self._write_response(status_code, data)
@check_auth
def do_DELETE_restart(self):
if self.server.patroni.ha.delete_future_restart():
data = "scheduled restart deleted"
code = 200
else:
data = "no restarts are scheduled"
code = 404
self._write_response(code, data)
@check_auth
def do_POST_reinitialize(self):
data = self.server.patroni.ha.reinitialize()
if data is None:
status_code = 200
data = 'reinitialize started'
else:
status_code = 503
self._write_response(status_code, data)
def poll_failover_result(self, leader, candidate):
timeout = max(10, self.server.patroni.dcs.loop_wait)
for _ in range(0, timeout*2):
time.sleep(1)
try:
cluster = self.server.patroni.dcs.get_cluster()
if cluster.leader and cluster.leader.name != leader:
if not candidate or candidate == cluster.leader.name:
return 200, 'Successfully failed over to "{0}"'.format(cluster.leader.name)
else:
return 200, 'Failed over to "{0}" instead of "{1}"'.format(cluster.leader.name, candidate)
if not cluster.failover:
return 503, 'Failover failed'
except Exception as e:
logger.debug('Exception occured during polling failover result: %s', e)
return 503, 'Failover status unknown'
def is_failover_possible(self, cluster, leader, candidate):
if leader and (not cluster.leader or cluster.leader.name != leader):
return 'leader name does not match'
if candidate:
members = [m for m in cluster.members if m.name == candidate]
if not members:
return 'candidate does not exists'
else:
members = [m for m in cluster.members if m.name != cluster.leader.name and m.api_url]
if not members:
return 'failover is not possible: cluster does not have members except leader'
for _, reachable, _, _, tags in self.server.patroni.ha.fetch_nodes_statuses(members):
if reachable and not tags.get('nofailover', False):
return None
return 'failover is not possible: no good candidates have been found'
@check_auth
def do_POST_failover(self):
request = self._read_json_content()
status_code = 500
if not request:
return
leader = request.get('leader')
candidate = request.get('candidate') or request.get('member')
scheduled_at = request.get('scheduled_at')
cluster = self.server.patroni.dcs.get_cluster()
if scheduled_at and cluster.is_paused():
self._write_response(status_code, "Can't schedule failover in the paused state")
logger.info("received failover request with leader=%s candidate=%s scheduled_at=%s",
leader, candidate, scheduled_at)
data = ''
if leader or candidate:
if scheduled_at:
(_, data, scheduled_at) = self.parse_schedule(scheduled_at, "failover")
if _:
status_code = _
elif self.server.patroni.dcs.manual_failover(leader, candidate, scheduled_at=scheduled_at):
self.server.patroni.dcs.event.set()
data = 'Failover scheduled'
status_code = 202
else:
data = 'failed to write failover key into DCS'
status_code = 503
else:
data = self.is_failover_possible(cluster, leader, candidate)
if not data:
if self.server.patroni.dcs.manual_failover(leader, candidate):
self.server.patroni.dcs.event.set()
status_code, data = self.poll_failover_result(cluster.leader and cluster.leader.name, candidate)
else:
data = 'failed to write failover key into DCS'
status_code = 503
else:
status_code = 400
data = 'No values given for required parameters leader and candidate'
self._write_response(status_code, data)
def parse_request(self):
"""Override parse_request method to enrich basic functionality of `BaseHTTPRequestHandler` class
Original class can only invoke do_GET, do_POST, do_PUT, etc method implementations if they are defined.
But we would like to have at least some simple routing mechanism, i.e.:
GET /uri1/part2 request should invoke `do_GET_uri1()`
POST /other should invoke `do_POST_other()`
If the `do_<REQUEST_METHOD>_<first_part_url>` method does not exists we'll fallback to original behavior."""
ret = BaseHTTPRequestHandler.parse_request(self)
if ret:
mname = self.path.lstrip('/').split('/')[0]
mname = self.command + ('_' + mname if mname else '')
if hasattr(self, 'do_' + mname):
self.command = mname
return ret
def query(self, sql, *params, **kwargs):
if not kwargs.get('retry', False):
return self.server.query(sql, *params)
retry = Retry(delay=1, retry_exceptions=PostgresConnectionException)
return retry(self.server.query, sql, *params)
def get_postgresql_status(self, retry=False):
try:
row = self.query("""WITH replication_info AS (
SELECT usename, application_name, client_addr, state, sync_state, sync_priority
FROM pg_stat_replication
)
SELECT to_char(pg_postmaster_start_time(), 'YYYY-MM-DD HH24:MI:SS.MS TZ'),
pg_is_in_recovery(),
CASE WHEN pg_is_in_recovery()
THEN 0
ELSE pg_xlog_location_diff(pg_current_xlog_location(), '0/0')::bigint
END,
pg_xlog_location_diff(pg_last_xlog_receive_location(), '0/0')::bigint,
pg_xlog_location_diff(pg_last_xlog_replay_location(), '0/0')::bigint,
to_char(pg_last_xact_replay_timestamp(), 'YYYY-MM-DD HH24:MI:SS.MS TZ'),
pg_is_in_recovery() AND pg_is_xlog_replay_paused(),
(SELECT array_to_json(array_agg(row_to_json(ri))) FROM replication_info ri)""",
retry=retry)[0]
result = {
'state': self.server.patroni.postgresql.state,
'postmaster_start_time': row[0],
'role': 'replica' if row[1] else 'master',
'server_version': self.server.patroni.postgresql.server_version,
'xlog': ({
'received_location': row[3],
'replayed_location': row[4],
'replayed_timestamp': row[5],
'paused': row[6]} if row[1] else {
'location': row[2]
})
}
if row[7]:
result['replication'] = row[7]
return result
except (psycopg2.Error, RetryFailedError, PostgresConnectionException):
state = self.server.patroni.postgresql.state
if state == 'running':
logger.exception('get_postgresql_status')
state = 'unknown'
return {'state': state, 'role': self.server.patroni.postgresql.role}
def log_message(self, fmt, *args):
logger.debug("API thread: %s - - [%s] %s", self.client_address[0], self.log_date_time_string(), fmt % args)
class RestApiServer(ThreadingMixIn, HTTPServer, Thread):
def __init__(self, patroni, config):
self.patroni = patroni
self.__initialize(config)
self.__set_config_parameters(config)
self.daemon = True
def query(self, sql, *params):
cursor = None
try:
with self.patroni.postgresql.connection().cursor() as cursor:
cursor.execute(sql, params)
return [r for r in cursor]
except psycopg2.Error as e:
if cursor and cursor.connection.closed == 0:
raise e
raise PostgresConnectionException('connection problems')
@staticmethod
def _set_fd_cloexec(fd):
flags = fcntl.fcntl(fd, fcntl.F_GETFD)
fcntl.fcntl(fd, fcntl.F_SETFD, flags | fcntl.FD_CLOEXEC)
def check_basic_auth_key(self, key):
return self.__auth_key == key
def check_auth_header(self, auth_header):
if self.__auth_key:
if auth_header is None:
return 'no auth header received'
if not auth_header.startswith('Basic ') or not self.check_basic_auth_key(auth_header[6:]):
return 'not authenticated'
@staticmethod
def __get_ssl_options(config):
return {option: config[option] for option in ['certfile', 'keyfile'] if option in config}
def __set_connection_string(self, connect_address):
self.connection_string = '{0}://{1}/patroni'.format(self.__protocol, connect_address or self.__listen)
def __set_config_parameters(self, config):
self.__auth_key = base64.b64encode(config['auth'].encode('utf-8')).decode('utf-8') if 'auth' in config else None
self.__set_connection_string(config.get('connect_address'))
def __initialize(self, config):
self.__ssl_options = self.__get_ssl_options(config)
self.__listen = config['listen']
host, port = config['listen'].split(':')
HTTPServer.__init__(self, (host, int(port)), RestApiHandler)
Thread.__init__(self, target=self.serve_forever)
self._set_fd_cloexec(self.socket)
self.__protocol = 'http'
# wrap socket with ssl if 'certfile' is defined in a config.yaml
# Sometime it's also needed to pass reference to a 'keyfile'.
if self.__ssl_options.get('certfile'):
import ssl
self.socket = ssl.wrap_socket(self.socket, server_side=True, **self.__ssl_options)
self.__protocol = 'https'
self.__set_connection_string(config.get('connect_address'))
def reload_config(self, config):
self.__set_config_parameters(config)
if self.__listen != config['listen'] or self.__ssl_options != self.__get_ssl_options(config):
self.shutdown()
self.__initialize(config)
self.start()