Acceptance tests for patronictl

Call patronictl.py when it's possible instead of doing REST API calls.

features/environment.py

@@ -118,6 +118,7 @@ class PatroniController(AbstractController):
 
         with open(patroni_config_name) as f:
             config = yaml.safe_load(f)
+            config.pop('etcd')
 
         host = config['postgresql']['listen'].split(':')[0]
 
@@ -139,16 +140,6 @@ class PatroniController(AbstractController):
         if tags:
             config['tags'] = tags
 
-        if dcs != 'etcd':
-            dcs_config = config.pop('etcd')
-            dcs_config.pop('host')
-
-            if dcs == 'exhibitor':
-                dcs_config.update({'hosts': ['127.0.0.1'], 'port': 8181})
-            elif dcs == 'zookeeper':
-                dcs_config['hosts'] = ['127.0.0.1:2181']
-            config[dcs] = dcs_config
-
         with open(patroni_config_path, 'w') as f:
             yaml.safe_dump(config, f, default_flow_style=False)
 
@@ -220,6 +211,7 @@ class ConsulController(AbstractDcsController):
 
     def __init__(self, output_dir):
         super(ConsulController, self).__init__('consul', tempfile.mkdtemp(), output_dir)
+        os.environ['PATRONI_CONSUL_HOST'] = 'localhost:8500'
         self._client = consul.Consul()
 
     def _start(self):
@@ -252,6 +244,7 @@ class EtcdController(AbstractDcsController):
 
     def __init__(self, output_dir):
         super(EtcdController, self).__init__('etcd', tempfile.mkdtemp(), output_dir)
+        os.environ['PATRONI_ETCD_HOST'] = 'localhost:4001'
         self._client = etcd.Client()
 
     def _start(self):
@@ -287,8 +280,10 @@ class ZooKeeperController(AbstractDcsController):
 
     """ handles all zookeeper related tasks, used for the tests setup and cleanup """
 
-    def __init__(self, output_dir):
+    def __init__(self, output_dir, export_env=True):
         super(ZooKeeperController, self).__init__('zookeeper', None, output_dir)
+        if export_env:
+            os.environ['PATRONI_ZOOKEEPER_HOSTS'] = 'localhost:2181'
         self._client = kazoo.client.KazooClient()
 
     def _start(self):
@@ -321,10 +316,17 @@ class ZooKeeperController(AbstractDcsController):
             return False
 
 
+class ExhibitorController(ZooKeeperController):
+
+    def __init__(self, output_dir):
+        super(ExhibitorController, self).__init__(output_dir, False)
+        os.environ.update({'PATRONI_EXHIBITOR_HOSTS': 'localhost', 'PATRONI_EXHIBITOR_PORT': '8181'})
+
+
 class PatroniPoolController(object):
 
     KNOWN_DCS = {'consul': ConsulController, 'etcd': EtcdController,
-                 'zookeeper': ZooKeeperController, 'exhibitor': ZooKeeperController}
+                 'zookeeper': ZooKeeperController, 'exhibitor': ExhibitorController}
 
     def __init__(self):
         self._dcs = None
@@ -376,8 +378,8 @@ class PatroniPoolController(object):
     @property
     def dcs(self):
         if self._dcs is None:
-            self._dcs = os.environ.get('DCS', 'etcd')
-            assert self._dcs in self.KNOWN_DCS, 'Unsupported dcs: ' + self.dcs
+            self._dcs = os.environ.pop('DCS', 'etcd')
+            assert self._dcs in self.KNOWN_DCS, 'Unsupported dcs: ' + self._dcs
         return self._dcs
 
 

features/patroni_api.feature

@@ -10,9 +10,12 @@ Scenario: check API requests on a stand-alone server
 	And I receive a response role master
 	When I issue a GET request to http://127.0.0.1:8008/replica
 	Then I receive a response code 503
-	When I issue an empty POST request to http://127.0.0.1:8008/reinitialize
-	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 reinit batman postgres0 --force
+	Then I receive a response returncode 0
+	And I receive a response output "reinitialize failed for member postgres0, status code=503, (I am the leader, can not reinitialize)"
+	When I run patronictl.py failover 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
@@ -23,24 +26,24 @@ Scenario: check API requests on a stand-alone server
 	And I receive a response text "No values given for required parameters leader and 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
-        And I issue an empty POST request to http://127.0.0.1:8008/reload
-        Then I receive a response code 202
+	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
+	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": 20, "loop_wait": 1, "postgresql": {"parameters": {"max_connections": 101}}}
-        Then I receive a response code 200
-        And I receive a response loop_wait 1
-        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 1
-        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'}
+	Given I issue a PATCH request to http://127.0.0.1:8008/config with {"ttl": 20, "loop_wait": 1, "postgresql": {"parameters": {"max_connections": 101}}}
+	Then I receive a response code 200
+	And I receive a response loop_wait 1
+	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 1
+	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'}
 
 Scenario: check API requests for the primary-replica pair
 	Given I start postgres1
@@ -49,26 +52,28 @@ Scenario: check API requests for the primary-replica pair
 	Then I receive a response code 200
 	And I receive a response state running
 	And I receive a response role replica
-	When I issue an empty POST request to http://127.0.0.1:8009/reinitialize
-	Then I receive a response code 200
-	When I issue an empty POST request to http://127.0.0.1:8008/restart
-	Then I receive a response code 200
-        And postgres0 role is the primary after 5 seconds
-        When I sleep for 10 seconds
-        Then postgres1 role is the secondary after 15 seconds
+	When I run patronictl.py reinit batman postgres1 --force
+	Then I receive a response returncode 0
+	And I receive a response output "Succesful reinitialize on member postgres1"
+	When I run patronictl.py restart batman postgres0 --force
+	Then I receive a response returncode 0
+	And I receive a response output "Succesful restart on member postgres0"
+	And postgres0 role is the primary after 5 seconds
+	When I sleep for 10 seconds
+	Then postgres1 role is the secondary after 15 seconds
 
 Scenario: check the failover via the API
-	Given I issue a POST request to http://127.0.0.1:8008/failover with {"leader": "postgres0", "candidate": "postgres1"}
-	Then I receive a response code 200
+	Given I run patronictl.py failover batman --master postgres0 --candidate postgres1 --force
+	Then I receive a response returncode 0
 	And postgres1 is a leader after 5 seconds
-        And postgres1 role is the primary after 5 seconds
-        And postgres0 role is the secondary after 10 seconds
+	And postgres1 role is the primary after 5 seconds
+	And postgres0 role is the secondary after 10 seconds
 	And replication works from postgres1 to postgres0 after 20 seconds
 
 Scenario: check the scheduled failover
-	Given I issue a scheduled failover at http://127.0.0.1:8009 from postgres1 to postgres0 in 1 seconds
-	Then I receive a response code 202
+	Given I issue a scheduled failover from postgres1 to postgres0 in 1 seconds
+	Then I receive a response returncode 0
 	And postgres0 is a leader after 20 seconds
-        And postgres0 role is the primary after 5 seconds
-        And postgres1 role is the secondary after 10 seconds
+	And postgres0 role is the primary after 5 seconds
+	And postgres1 role is the secondary after 10 seconds
 	And replication works from postgres0 to postgres1 after 25 seconds

features/steps/patroni_api.py

@@ -2,6 +2,8 @@ import json
 import parse
 import pytz
 import requests
+import shlex
+import subprocess
 import time
 import yaml
 
@@ -84,23 +86,39 @@ def do_request(context, request_method, url, data):
         _set_response(context, r)
 
 
+@step('I run {cmd}')
+def do_run(context, cmd):
+    cmd = ['coverage', 'run', '--source=patroni', '-p'] + shlex.split(cmd)
+    try:
+        response = subprocess.check_output(cmd, stderr=subprocess.STDOUT)
+        context.status_code = 0
+    except subprocess.CalledProcessError as e:
+        response = e.output
+        context.status_code = e.returncode
+    context.response = response.strip()
+
+
 @then('I receive a response {component:w} {data}')
 def check_response(context, component, data):
     if component == 'code':
         assert context.status_code == int(data),\
-            "status code {0} != {1}, response: {2}".format(context.status_code, int(data), context.response)
+            "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)
     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)
         assert str(context.response[component]) == str(data), "{0} does not contain {1}".format(component, data)
 
 
-@step('I issue a scheduled failover at {at_url:url} from {from_host:w} to {to_host:w} in {in_seconds:d} seconds')
-def scheduled_failover(context, at_url, from_host, to_host, in_seconds):
+@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):
     context.execute_steps(u"""
-        Given I issue a POST request to {0}/failover with {{"leader": "{1}", "candidate": "{2}", "scheduled_at": "{3}"}}
-    """.format(at_url, from_host, to_host, datetime.now(pytz.utc) + timedelta(seconds=int(in_seconds))))
+        Given I run patronictl.py failover batman --master {0} --candidate {1} --scheduled "{2}" --force
+    """.format(from_host, to_host, datetime.now(pytz.utc) + timedelta(seconds=int(in_seconds))))
 
 
 @step('I add tag {tag:w} {value:w} to {pg_name:w} config')