Catch-up after Federation Outage (bonus): Catch-up on Synapse Startup (#8322)

Signed-off-by: Olivier Wilkinson (reivilibre) <olivier@librepush.net>
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>

* Fix _set_destination_retry_timings

This came about because the code assumed that retry_interval
could not be NULL — which has been challenged by catch-up.
pull/8354/head
reivilibre 2020-09-18 14:59:13 +01:00 committed by GitHub
parent 8a4a4186de
commit 36efbcaf51
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
10 changed files with 218 additions and 5 deletions

1
changelog.d/8230.bugfix Normal file
View File

@ -0,0 +1 @@
Fix messages over federation being lost until an event is sent into the same room.

View File

@ -1 +0,0 @@
Track the latest event for every destination and room for catch-up after federation outage.

1
changelog.d/8247.bugfix Normal file
View File

@ -0,0 +1 @@
Fix messages over federation being lost until an event is sent into the same room.

View File

@ -1 +0,0 @@
Track the `stream_ordering` of the last successfully-sent event to every destination, so we can use this information to 'catch up' a remote server after an outage.

1
changelog.d/8258.bugfix Normal file
View File

@ -0,0 +1 @@
Fix messages over federation being lost until an event is sent into the same room.

View File

@ -1 +0,0 @@
Track the `stream_ordering` of the last successfully-sent event to every destination, so we can use this information to 'catch up' a remote server after an outage.

1
changelog.d/8322.bugfix Normal file
View File

@ -0,0 +1 @@
Fix messages over federation being lost until an event is sent into the same room.

View File

@ -55,6 +55,15 @@ sent_pdus_destination_dist_total = Counter(
"Total number of PDUs queued for sending across all destinations", "Total number of PDUs queued for sending across all destinations",
) )
# Time (in s) after Synapse's startup that we will begin to wake up destinations
# that have catch-up outstanding.
CATCH_UP_STARTUP_DELAY_SEC = 15
# Time (in s) to wait in between waking up each destination, i.e. one destination
# will be woken up every <x> seconds after Synapse's startup until we have woken
# every destination has outstanding catch-up.
CATCH_UP_STARTUP_INTERVAL_SEC = 5
class FederationSender: class FederationSender:
def __init__(self, hs: "synapse.server.HomeServer"): def __init__(self, hs: "synapse.server.HomeServer"):
@ -125,6 +134,14 @@ class FederationSender:
1000.0 / hs.config.federation_rr_transactions_per_room_per_second 1000.0 / hs.config.federation_rr_transactions_per_room_per_second
) )
# wake up destinations that have outstanding PDUs to be caught up
self._catchup_after_startup_timer = self.clock.call_later(
CATCH_UP_STARTUP_DELAY_SEC,
run_as_background_process,
"wake_destinations_needing_catchup",
self._wake_destinations_needing_catchup,
)
def _get_per_destination_queue(self, destination: str) -> PerDestinationQueue: def _get_per_destination_queue(self, destination: str) -> PerDestinationQueue:
"""Get or create a PerDestinationQueue for the given destination """Get or create a PerDestinationQueue for the given destination
@ -560,3 +577,37 @@ class FederationSender:
# Dummy implementation for case where federation sender isn't offloaded # Dummy implementation for case where federation sender isn't offloaded
# to a worker. # to a worker.
return [], 0, False return [], 0, False
async def _wake_destinations_needing_catchup(self):
"""
Wakes up destinations that need catch-up and are not currently being
backed off from.
In order to reduce load spikes, adds a delay between each destination.
"""
last_processed = None # type: Optional[str]
while True:
destinations_to_wake = await self.store.get_catch_up_outstanding_destinations(
last_processed
)
if not destinations_to_wake:
# finished waking all destinations!
self._catchup_after_startup_timer = None
break
destinations_to_wake = [
d
for d in destinations_to_wake
if self._federation_shard_config.should_handle(self._instance_name, d)
]
for last_processed in destinations_to_wake:
logger.info(
"Destination %s has outstanding catch-up, waking up.",
last_processed,
)
self.wake_destination(last_processed)
await self.clock.sleep(CATCH_UP_STARTUP_INTERVAL_SEC)

View File

@ -218,6 +218,7 @@ class TransactionStore(SQLBaseStore):
retry_interval = EXCLUDED.retry_interval retry_interval = EXCLUDED.retry_interval
WHERE WHERE
EXCLUDED.retry_interval = 0 EXCLUDED.retry_interval = 0
OR destinations.retry_interval IS NULL
OR destinations.retry_interval < EXCLUDED.retry_interval OR destinations.retry_interval < EXCLUDED.retry_interval
""" """
@ -249,7 +250,11 @@ class TransactionStore(SQLBaseStore):
"retry_interval": retry_interval, "retry_interval": retry_interval,
}, },
) )
elif retry_interval == 0 or prev_row["retry_interval"] < retry_interval: elif (
retry_interval == 0
or prev_row["retry_interval"] is None
or prev_row["retry_interval"] < retry_interval
):
self.db_pool.simple_update_one_txn( self.db_pool.simple_update_one_txn(
txn, txn,
"destinations", "destinations",
@ -397,7 +402,7 @@ class TransactionStore(SQLBaseStore):
@staticmethod @staticmethod
def _get_catch_up_room_event_ids_txn( def _get_catch_up_room_event_ids_txn(
txn, destination: str, last_successful_stream_ordering: int, txn: LoggingTransaction, destination: str, last_successful_stream_ordering: int,
) -> List[str]: ) -> List[str]:
q = """ q = """
SELECT event_id FROM destination_rooms SELECT event_id FROM destination_rooms
@ -412,3 +417,60 @@ class TransactionStore(SQLBaseStore):
) )
event_ids = [row[0] for row in txn] event_ids = [row[0] for row in txn]
return event_ids return event_ids
async def get_catch_up_outstanding_destinations(
self, after_destination: Optional[str]
) -> List[str]:
"""
Gets at most 25 destinations which have outstanding PDUs to be caught up,
and are not being backed off from
Args:
after_destination:
If provided, all destinations must be lexicographically greater
than this one.
Returns:
list of up to 25 destinations with outstanding catch-up.
These are the lexicographically first destinations which are
lexicographically greater than after_destination (if provided).
"""
time = self.hs.get_clock().time_msec()
return await self.db_pool.runInteraction(
"get_catch_up_outstanding_destinations",
self._get_catch_up_outstanding_destinations_txn,
time,
after_destination,
)
@staticmethod
def _get_catch_up_outstanding_destinations_txn(
txn: LoggingTransaction, now_time_ms: int, after_destination: Optional[str]
) -> List[str]:
q = """
SELECT destination FROM destinations
WHERE destination IN (
SELECT destination FROM destination_rooms
WHERE destination_rooms.stream_ordering >
destinations.last_successful_stream_ordering
)
AND destination > ?
AND (
retry_last_ts IS NULL OR
retry_last_ts + retry_interval < ?
)
ORDER BY destination
LIMIT 25
"""
txn.execute(
q,
(
# everything is lexicographically greater than "" so this gives
# us the first batch of up to 25.
after_destination or "",
now_time_ms,
),
)
destinations = [row[0] for row in txn]
return destinations

View File

@ -321,3 +321,102 @@ class FederationCatchUpTestCases(FederatingHomeserverTestCase):
per_dest_queue._last_successful_stream_ordering, per_dest_queue._last_successful_stream_ordering,
event_5.internal_metadata.stream_ordering, event_5.internal_metadata.stream_ordering,
) )
@override_config({"send_federation": True})
def test_catch_up_on_synapse_startup(self):
"""
Tests the behaviour of get_catch_up_outstanding_destinations and
_wake_destinations_needing_catchup.
"""
# list of sorted server names (note that there are more servers than the batch
# size used in get_catch_up_outstanding_destinations).
server_names = ["server%02d" % number for number in range(42)] + ["zzzerver"]
# ARRANGE:
# - a local user (u1)
# - a room which u1 is joined to (and remote users @user:serverXX are
# joined to)
# mark the remotes as online
self.is_online = True
self.register_user("u1", "you the one")
u1_token = self.login("u1", "you the one")
room_id = self.helper.create_room_as("u1", tok=u1_token)
for server_name in server_names:
self.get_success(
event_injection.inject_member_event(
self.hs, room_id, "@user:%s" % server_name, "join"
)
)
# create an event
self.helper.send(room_id, "deary me!", tok=u1_token)
# ASSERT:
# - All servers are up to date so none should have outstanding catch-up
outstanding_when_successful = self.get_success(
self.hs.get_datastore().get_catch_up_outstanding_destinations(None)
)
self.assertEqual(outstanding_when_successful, [])
# ACT:
# - Make the remote servers unreachable
self.is_online = False
# - Mark zzzerver as being backed-off from
now = self.clock.time_msec()
self.get_success(
self.hs.get_datastore().set_destination_retry_timings(
"zzzerver", now, now, 24 * 60 * 60 * 1000 # retry in 1 day
)
)
# - Send an event
self.helper.send(room_id, "can anyone hear me?", tok=u1_token)
# ASSERT (get_catch_up_outstanding_destinations):
# - all remotes are outstanding
# - they are returned in batches of 25, in order
outstanding_1 = self.get_success(
self.hs.get_datastore().get_catch_up_outstanding_destinations(None)
)
self.assertEqual(len(outstanding_1), 25)
self.assertEqual(outstanding_1, server_names[0:25])
outstanding_2 = self.get_success(
self.hs.get_datastore().get_catch_up_outstanding_destinations(
outstanding_1[-1]
)
)
self.assertNotIn("zzzerver", outstanding_2)
self.assertEqual(len(outstanding_2), 17)
self.assertEqual(outstanding_2, server_names[25:-1])
# ACT: call _wake_destinations_needing_catchup
# patch wake_destination to just count the destinations instead
woken = []
def wake_destination_track(destination):
woken.append(destination)
self.hs.get_federation_sender().wake_destination = wake_destination_track
# cancel the pre-existing timer for _wake_destinations_needing_catchup
# this is because we are calling it manually rather than waiting for it
# to be called automatically
self.hs.get_federation_sender()._catchup_after_startup_timer.cancel()
self.get_success(
self.hs.get_federation_sender()._wake_destinations_needing_catchup(), by=5.0
)
# ASSERT (_wake_destinations_needing_catchup):
# - all remotes are woken up, save for zzzerver
self.assertNotIn("zzzerver", woken)
# - all destinations are woken exactly once; they appear once in woken.
self.assertCountEqual(woken, server_names[:-1])