Merge branch 'master' of github.com:zalando/patroni into feature/acceptance_tests_travis

patroni/api.py

@@ -174,38 +174,48 @@ class RestApiHandler(BaseHTTPRequestHandler):
     @check_auth
     def do_POST_failover(self):
         content_length = int(self.headers.get('content-length', 0))
-        request = json.loads(self.rfile.read(content_length).decode('utf-8'))
+        try:
+            request = json.loads(self.rfile.read(content_length).decode('utf-8'))
+        except ValueError:
+            request = {}
         leader = request.get('leader')
         member = request.get('member')
         cluster = self.server.patroni.ha.dcs.get_cluster()
         status_code = 500
 
+        logger.info("received failover request with leader {0} member {1} scheduled_at {2}".
+                    format(leader, member, request.get("scheduled_at")))
+
         data = b''
-        if request.get('scheduled_at'):
-            try:
-                scheduled_at = dateutil.parser.parse(request['scheduled_at'])
-                if scheduled_at.tzinfo is None:
-                    data = b'Timezone information is mandatory for scheduled_at'
-                    status_code = 400
-                elif scheduled_at < datetime.datetime.now(pytz.utc):
-                    data = b'Cannot schedule failover in the past'
+        if leader or member:
+            if request.get('scheduled_at'):
+                try:
+                    scheduled_at = dateutil.parser.parse(request['scheduled_at'])
+                    if scheduled_at.tzinfo is None:
+                        data = b'Timezone information is mandatory for scheduled_at'
+                        status_code = 400
+                    elif scheduled_at < datetime.datetime.now(pytz.utc):
+                        data = b'Cannot schedule failover in the past'
+                        status_code = 422
+                    elif self.server.patroni.dcs.manual_failover(leader, member, scheduled_at):
+                        data = b'Failover scheduled'
+                        status_code = 200
+                except (ValueError, TypeError):
+                    logger.exception('Invalid scheduled failover time: {}'.format(request['scheduled_at']))
+                    data = b'Unable to parse scheduled timestamp. It should be in an unambiguous format, e.g. ISO 8601'
                     status_code = 422
-                elif self.server.patroni.dcs.manual_failover(leader, member, scheduled_at):
-                    data = b'Failover scheduled'
-                    status_code = 200
-            except (ValueError, TypeError):
-                logger.exception('Invalid scheduled failover time: {}'.format(request['scheduled_at']))
-                data = b'Unable to parse scheduled timestamp. It should be in an unambiguous format, e.g. ISO 8601'
-                status_code = 422
+            else:
+                data = self.is_failover_possible(cluster, leader, member)
+                if not data:
+                    if not self.server.patroni.dcs.manual_failover(leader, member):
+                        data = b'failed to write failover key into DCS'
+                        status_code = 503
+                    else:
+                        self.server.patroni.dcs.event.set()
+                        status_code, data = self.poll_failover_result(cluster.leader and cluster.leader.name, member)
         else:
-            data = self.is_failover_possible(cluster, leader, member)
-            if not data:
-                if not self.server.patroni.dcs.manual_failover(leader, member):
-                    data = b'failed to write failover key into DCS'
-                    status_code = 503
-                else:
-                    self.server.patroni.dcs.event.set()
-                    status_code, data = self.poll_failover_result(cluster.leader and cluster.leader.name, member)
+            status_code = 400
+            data = b'No values given for required parameters leader and member'
 
         self.send_response(status_code)
         self.send_header('Content-Type', 'text/html')

patroni/ha.py

@@ -287,10 +287,10 @@ class Ha(object):
             try:
                 delta = (failover.scheduled_at - now).total_seconds()
 
-                if delta > 10:
+                if delta > self.patroni.nap_time:
                     logging.info('Awaiting failover at %s (in %.0f seconds)', failover.scheduled_at.isoformat(), delta)
                     return
-                elif delta < -15:
+                elif delta < - int(self.patroni.nap_time * 1.5):
                     logger.warning('Found a stale failover value, cleaning up: %s', failover.scheduled_at)
                     self.dcs.manual_failover('', '', self.cluster.failover.index)
                     return

tests/test_api.py

@@ -153,6 +153,10 @@ class TestRestApiHandler(unittest.TestCase):
     @patch.object(MockHa, 'dcs')
     def test_do_POST_failover(self, dcs):
         cluster = dcs.get_cluster.return_value
+
+        request = b'POST /failover HTTP/1.0\nAuthorization: Basic dGVzdDp0ZXN0\n' +\
+                  b'Content-Length: 0\n\n'
+        MockRestApiServer(RestApiHandler, request)
         request = b'POST /failover HTTP/1.0\nAuthorization: Basic dGVzdDp0ZXN0\n' +\
                   b'Content-Length: 25\n\n{"leader": "postgresql1"}'
         MockRestApiServer(RestApiHandler, request)

tests/test_ha.py

@@ -54,6 +54,7 @@ class MockPatroni(object):
         self.api = Mock()
         self.tags = {}
         self.nofailover = None
+        self.nap_time = 10
         self.replicatefrom = None
         self.api.connection_string = 'http://127.0.0.1:8008'