Add 'failure_ts' column to 'destinations' table

changelog.d/6016.misc

@@ -0,0 +1 @@
+Add a 'failure_ts' column to the 'destinations' database table.

changelog.d/6025.bugfix

@@ -0,0 +1 @@
+Fix bug in calculating the federation retry backoff period.

synapse/storage/schema/delta/56/destinations_failure_ts.sql

@@ -0,0 +1,25 @@
+/* Copyright 2019 The Matrix.org Foundation C.I.C
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/*
+ * Record the timestamp when a given server started failing
+ */
+ALTER TABLE destinations ADD failure_ts BIGINT;
+
+/* as a rough approximation, we assume that the server started failing at
+ * retry_interval before the last retry
+ */
+UPDATE destinations SET failure_ts = retry_last_ts - retry_interval
+    WHERE retry_last_ts > 0;

synapse/storage/transactions.py

@@ -165,7 +165,7 @@ def _get_destination_retry_timings(self, txn, destination):
             txn,
             table="destinations",
             keyvalues={"destination": destination},
-            retcols=("destination", "retry_last_ts", "retry_interval"),
+            retcols=("destination", "failure_ts", "retry_last_ts", "retry_interval"),
             allow_none=True,
         )
 
@@ -174,12 +174,15 @@ def _get_destination_retry_timings(self, txn, destination):
         else:
             return None
 
-    def set_destination_retry_timings(self, destination, retry_last_ts, retry_interval):
+    def set_destination_retry_timings(
+        self, destination, failure_ts, retry_last_ts, retry_interval
+    ):
         """Sets the current retry timings for a given destination.
         Both timings should be zero if retrying is no longer occuring.
 
         Args:
             destination (str)
+            failure_ts (int|None) - when the server started failing (ms since epoch)
             retry_last_ts (int) - time of last retry attempt in unix epoch ms
             retry_interval (int) - how long until next retry in ms
         """
@@ -189,30 +192,34 @@ def set_destination_retry_timings(self, destination, retry_last_ts, retry_interv
             "set_destination_retry_timings",
             self._set_destination_retry_timings,
             destination,
+            failure_ts,
             retry_last_ts,
             retry_interval,
         )
 
     def _set_destination_retry_timings(
-        self, txn, destination, retry_last_ts, retry_interval
+        self, txn, destination, failure_ts, retry_last_ts, retry_interval
     ):
 
         if self.database_engine.can_native_upsert:
             # Upsert retry time interval if retry_interval is zero (i.e. we're
             # resetting it) or greater than the existing retry interval.
 
             sql = """
-                INSERT INTO destinations (destination, retry_last_ts, retry_interval)
-                    VALUES (?, ?, ?)
+                INSERT INTO destinations (
+                    destination, failure_ts, retry_last_ts, retry_interval
+                )
+                    VALUES (?, ?, ?, ?)
                 ON CONFLICT (destination) DO UPDATE SET
+                        failure_ts = EXCLUDED.failure_ts,
                         retry_last_ts = EXCLUDED.retry_last_ts,
                         retry_interval = EXCLUDED.retry_interval
                     WHERE
                         EXCLUDED.retry_interval = 0
                         OR destinations.retry_interval < EXCLUDED.retry_interval
             """
 
-            txn.execute(sql, (destination, retry_last_ts, retry_interval))
+            txn.execute(sql, (destination, failure_ts, retry_last_ts, retry_interval))
 
             return
 
@@ -225,7 +232,7 @@ def _set_destination_retry_timings(
             txn,
             table="destinations",
             keyvalues={"destination": destination},
-            retcols=("retry_last_ts", "retry_interval"),
+            retcols=("failure_ts", "retry_last_ts", "retry_interval"),
             allow_none=True,
         )
 
@@ -235,6 +242,7 @@ def _set_destination_retry_timings(
                 table="destinations",
                 values={
                     "destination": destination,
+                    "failure_ts": failure_ts,
                     "retry_last_ts": retry_last_ts,
                     "retry_interval": retry_interval,
                 },
@@ -245,6 +253,7 @@ def _set_destination_retry_timings(
                 "destinations",
                 keyvalues={"destination": destination},
                 updatevalues={
+                    "failure_ts": failure_ts,
                     "retry_last_ts": retry_last_ts,
                     "retry_interval": retry_interval,
                 },

synapse/util/retryutils.py

@@ -80,11 +80,13 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs)
             # We aren't ready to retry that destination.
             raise
     """
+    failure_ts = None
     retry_last_ts, retry_interval = (0, 0)
 
     retry_timings = yield store.get_destination_retry_timings(destination)
 
     if retry_timings:
+        failure_ts = retry_timings["failure_ts"]
         retry_last_ts, retry_interval = (
             retry_timings["retry_last_ts"],
             retry_timings["retry_interval"],
@@ -108,6 +110,7 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs)
         destination,
         clock,
         store,
+        failure_ts,
         retry_interval,
         backoff_on_failure=backoff_on_failure,
         **kwargs
@@ -120,6 +123,7 @@ def __init__(
         destination,
         clock,
         store,
+        failure_ts,
         retry_interval,
         backoff_on_404=False,
         backoff_on_failure=True,
@@ -133,6 +137,8 @@ def __init__(
             destination (str)
             clock (Clock)
             store (DataStore)
+            failure_ts (int|None): when this destination started failing (in ms since
+                the epoch), or zero if the last request was successful
             retry_interval (int): The next retry interval taken from the
                 database in milliseconds, or zero if the last request was
                 successful.
@@ -145,6 +151,7 @@ def __init__(
         self.store = store
         self.destination = destination
 
+        self.failure_ts = failure_ts
         self.retry_interval = retry_interval
         self.backoff_on_404 = backoff_on_404
         self.backoff_on_failure = backoff_on_failure
@@ -186,15 +193,17 @@ def __exit__(self, exc_type, exc_val, exc_tb):
             logger.debug(
                 "Connection to %s was successful; clearing backoff", self.destination
             )
+            self.failure_ts = None
             retry_last_ts = 0
             self.retry_interval = 0
         elif not self.backoff_on_failure:
             return
         else:
             # We couldn't connect.
             if self.retry_interval:
-                self.retry_interval *= RETRY_MULTIPLIER
-                self.retry_interval *= int(random.uniform(0.8, 1.4))
+                self.retry_interval = int(
+                    self.retry_interval * RETRY_MULTIPLIER * random.uniform(0.8, 1.4)
+                )
 
                 if self.retry_interval >= MAX_RETRY_INTERVAL:
                     self.retry_interval = MAX_RETRY_INTERVAL
@@ -210,11 +219,17 @@ def __exit__(self, exc_type, exc_val, exc_tb):
             )
             retry_last_ts = int(self.clock.time_msec())
 
+            if self.failure_ts is None:
+                self.failure_ts = retry_last_ts
+
         @defer.inlineCallbacks
         def store_retry_timings():
             try:
                 yield self.store.set_destination_retry_timings(
-                    self.destination, retry_last_ts, self.retry_interval
+                    self.destination,
+                    self.failure_ts,
+                    retry_last_ts,
+                    self.retry_interval,
                 )
             except Exception:
                 logger.exception("Failed to store destination_retry_timings")

tests/handlers/test_typing.py

@@ -99,7 +99,12 @@ def prepare(self, reactor, clock, hs):
         self.event_source = hs.get_event_sources().sources["typing"]
 
         self.datastore = hs.get_datastore()
-        retry_timings_res = {"destination": "", "retry_last_ts": 0, "retry_interval": 0}
+        retry_timings_res = {
+            "destination": "",
+            "retry_last_ts": 0,
+            "retry_interval": 0,
+            "failure_ts": None,
+        }
         self.datastore.get_destination_retry_timings.return_value = defer.succeed(
             retry_timings_res
         )

tests/storage/test_transactions.py

@@ -29,17 +29,19 @@ def test_get_set_transactions(self):
         r = self.get_success(d)
         self.assertIsNone(r)
 
-        d = self.store.set_destination_retry_timings("example.com", 50, 100)
+        d = self.store.set_destination_retry_timings("example.com", 1000, 50, 100)
         self.get_success(d)
 
         d = self.store.get_destination_retry_timings("example.com")
         r = self.get_success(d)
 
-        self.assert_dict({"retry_last_ts": 50, "retry_interval": 100}, r)
+        self.assert_dict(
+            {"retry_last_ts": 50, "retry_interval": 100, "failure_ts": 1000}, r
+        )
 
     def test_initial_set_transactions(self):
         """Tests that we can successfully set the destination retries (there
         was a bug around invalidating the cache that broke this)
         """
-        d = self.store.set_destination_retry_timings("example.com", 50, 100)
+        d = self.store.set_destination_retry_timings("example.com", 1000, 50, 100)
         self.get_success(d)

tests/util/test_retryutils.py

@@ -0,0 +1,127 @@
+# -*- coding: utf-8 -*-
+# Copyright 2019 The Matrix.org Foundation C.I.C.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from synapse.util.retryutils import (
+    MIN_RETRY_INTERVAL,
+    RETRY_MULTIPLIER,
+    NotRetryingDestination,
+    get_retry_limiter,
+)
+
+from tests.unittest import HomeserverTestCase
+
+
+class RetryLimiterTestCase(HomeserverTestCase):
+    def test_new_destination(self):
+        """A happy-path case with a new destination and a successful operation"""
+        store = self.hs.get_datastore()
+        d = get_retry_limiter("test_dest", self.clock, store)
+        self.pump()
+        limiter = self.successResultOf(d)
+
+        # advance the clock a bit before making the request
+        self.pump(1)
+
+        with limiter:
+            pass
+
+        d = store.get_destination_retry_timings("test_dest")
+        self.pump()
+        new_timings = self.successResultOf(d)
+        self.assertIsNone(new_timings)
+
+    def test_limiter(self):
+        """General test case which walks through the process of a failing request"""
+        store = self.hs.get_datastore()
+
+        d = get_retry_limiter("test_dest", self.clock, store)
+        self.pump()
+        limiter = self.successResultOf(d)
+
+        self.pump(1)
+        try:
+            with limiter:
+                self.pump(1)
+                failure_ts = self.clock.time_msec()
+                raise AssertionError("argh")
+        except AssertionError:
+            pass
+
+        # wait for the update to land
+        self.pump()
+
+        d = store.get_destination_retry_timings("test_dest")
+        self.pump()
+        new_timings = self.successResultOf(d)
+        self.assertEqual(new_timings["failure_ts"], failure_ts)
+        self.assertEqual(new_timings["retry_last_ts"], failure_ts)
+        self.assertEqual(new_timings["retry_interval"], MIN_RETRY_INTERVAL)
+
+        # now if we try again we should get a failure
+        d = get_retry_limiter("test_dest", self.clock, store)
+        self.pump()
+        self.failureResultOf(d, NotRetryingDestination)
+
+        #
+        # advance the clock and try again
+        #
+
+        self.pump(MIN_RETRY_INTERVAL)
+        d = get_retry_limiter("test_dest", self.clock, store)
+        self.pump()
+        limiter = self.successResultOf(d)
+
+        self.pump(1)
+        try:
+            with limiter:
+                self.pump(1)
+                retry_ts = self.clock.time_msec()
+                raise AssertionError("argh")
+        except AssertionError:
+            pass
+
+        # wait for the update to land
+        self.pump()
+
+        d = store.get_destination_retry_timings("test_dest")
+        self.pump()
+        new_timings = self.successResultOf(d)
+        self.assertEqual(new_timings["failure_ts"], failure_ts)
+        self.assertEqual(new_timings["retry_last_ts"], retry_ts)
+        self.assertGreaterEqual(
+            new_timings["retry_interval"], MIN_RETRY_INTERVAL * RETRY_MULTIPLIER * 0.5
+        )
+        self.assertLessEqual(
+            new_timings["retry_interval"], MIN_RETRY_INTERVAL * RETRY_MULTIPLIER * 2.0
+        )
+
+        #
+        # one more go, with success
+        #
+        self.pump(MIN_RETRY_INTERVAL * RETRY_MULTIPLIER * 2.0)
+        d = get_retry_limiter("test_dest", self.clock, store)
+        self.pump()
+        limiter = self.successResultOf(d)
+
+        self.pump(1)
+        with limiter:
+            self.pump(1)
+
+        # wait for the update to land
+        self.pump()
+
+        d = store.get_destination_retry_timings("test_dest")
+        self.pump()
+        new_timings = self.successResultOf(d)
+        self.assertIsNone(new_timings)