Revert the deletion of stale devices due to performance issues. (#14662)

pull/14666/head
reivilibre 2022-12-12 13:55:23 +00:00 committed by GitHub
parent 527366f962
commit 74b89c2761
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
7 changed files with 5 additions and 152 deletions

View File

@ -1 +0,0 @@
Prune user's old devices on login if they have too many.

View File

@ -1 +0,0 @@
Prune user's old devices on login if they have too many.

View File

@ -0,0 +1 @@
(remove from changelog: unreleased) Revert the deletion of stale devices due to performance issues.

View File

@ -52,7 +52,6 @@ from synapse.util import stringutils
from synapse.util.async_helpers import Linearizer from synapse.util.async_helpers import Linearizer
from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.caches.expiringcache import ExpiringCache
from synapse.util.cancellation import cancellable from synapse.util.cancellation import cancellable
from synapse.util.iterutils import batch_iter
from synapse.util.metrics import measure_func from synapse.util.metrics import measure_func
from synapse.util.retryutils import NotRetryingDestination from synapse.util.retryutils import NotRetryingDestination
@ -422,9 +421,6 @@ class DeviceHandler(DeviceWorkerHandler):
self._check_device_name_length(initial_device_display_name) self._check_device_name_length(initial_device_display_name)
# Prune the user's device list if they already have a lot of devices.
await self._prune_too_many_devices(user_id)
if device_id is not None: if device_id is not None:
new_device = await self.store.store_device( new_device = await self.store.store_device(
user_id=user_id, user_id=user_id,
@ -456,33 +452,6 @@ class DeviceHandler(DeviceWorkerHandler):
raise errors.StoreError(500, "Couldn't generate a device ID.") raise errors.StoreError(500, "Couldn't generate a device ID.")
async def _prune_too_many_devices(self, user_id: str) -> None:
"""Delete any excess old devices this user may have."""
device_ids = await self.store.check_too_many_devices_for_user(user_id, 100)
if not device_ids:
return
logger.info("Pruning %d old devices for user %s", len(device_ids), user_id)
# We don't want to block and try and delete tonnes of devices at once,
# so we cap the number of devices we delete synchronously.
first_batch, remaining_device_ids = device_ids[:10], device_ids[10:]
await self.delete_devices(user_id, first_batch)
if not remaining_device_ids:
return
# Now spawn a background loop that deletes the rest.
async def _prune_too_many_devices_loop() -> None:
for batch in batch_iter(remaining_device_ids, 10):
await self.delete_devices(user_id, batch)
await self.clock.sleep(1)
run_as_background_process(
"_prune_too_many_devices_loop", _prune_too_many_devices_loop
)
async def _delete_stale_devices(self) -> None: async def _delete_stale_devices(self) -> None:
"""Background task that deletes devices which haven't been accessed for more than """Background task that deletes devices which haven't been accessed for more than
a configured time period. a configured time period.
@ -512,7 +481,7 @@ class DeviceHandler(DeviceWorkerHandler):
device_ids = [d for d in device_ids if d != except_device_id] device_ids = [d for d in device_ids if d != except_device_id]
await self.delete_devices(user_id, device_ids) await self.delete_devices(user_id, device_ids)
async def delete_devices(self, user_id: str, device_ids: Collection[str]) -> None: async def delete_devices(self, user_id: str, device_ids: List[str]) -> None:
"""Delete several devices """Delete several devices
Args: Args:

View File

@ -1569,77 +1569,6 @@ class DeviceBackgroundUpdateStore(SQLBaseStore):
return rows return rows
async def check_too_many_devices_for_user(
self, user_id: str, limit: int
) -> List[str]:
"""Check if the user has a lot of devices, and if so return the set of
devices we can prune.
This does *not* return hidden devices or devices with E2E keys.
Returns at most `limit` number of devices, ordered by last seen.
"""
num_devices = await self.db_pool.simple_select_one_onecol(
table="devices",
keyvalues={"user_id": user_id, "hidden": False},
retcol="COALESCE(COUNT(*), 0)",
desc="count_devices",
)
# We let users have up to ten devices without pruning.
if num_devices <= 10:
return []
# We prune everything older than N days.
max_last_seen = self._clock.time_msec() - 14 * 24 * 60 * 60 * 1000
if num_devices > 50:
# If the user has more than 50 devices, then we chose a last seen
# that ensures we keep at most 50 devices.
sql = """
SELECT last_seen FROM devices
LEFT JOIN e2e_device_keys_json USING (user_id, device_id)
WHERE
user_id = ?
AND NOT hidden
AND last_seen IS NOT NULL
AND key_json IS NULL
ORDER BY last_seen DESC
LIMIT 1
OFFSET 50
"""
rows = await self.db_pool.execute(
"check_too_many_devices_for_user_last_seen", None, sql, (user_id,)
)
if rows:
max_last_seen = max(rows[0][0], max_last_seen)
# Now fetch the devices to delete.
sql = """
SELECT device_id FROM devices
LEFT JOIN e2e_device_keys_json USING (user_id, device_id)
WHERE
user_id = ?
AND NOT hidden
AND last_seen < ?
AND key_json IS NULL
ORDER BY last_seen
LIMIT ?
"""
def check_too_many_devices_for_user_txn(
txn: LoggingTransaction,
) -> List[str]:
txn.execute(sql, (user_id, max_last_seen, limit))
return [device_id for device_id, in txn]
return await self.db_pool.runInteraction(
"check_too_many_devices_for_user",
check_too_many_devices_for_user_txn,
)
class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
# Because we have write access, this will be a StreamIdGenerator # Because we have write access, this will be a StreamIdGenerator
@ -1698,7 +1627,6 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
values={}, values={},
insertion_values={ insertion_values={
"display_name": initial_device_display_name, "display_name": initial_device_display_name,
"last_seen": self._clock.time_msec(),
"hidden": False, "hidden": False,
}, },
desc="store_device", desc="store_device",
@ -1744,15 +1672,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
) )
raise StoreError(500, "Problem storing device.") raise StoreError(500, "Problem storing device.")
@cached(max_entries=0) async def delete_devices(self, user_id: str, device_ids: List[str]) -> None:
async def delete_device(self, user_id: str, device_id: str) -> None:
raise NotImplementedError()
# Note: sometimes deleting rows out of `device_inbox` can take a long time,
# so we use a cache so that we deduplicate in flight requests to delete
# devices.
@cachedList(cached_method_name="delete_device", list_name="device_ids")
async def delete_devices(self, user_id: str, device_ids: Collection[str]) -> dict:
"""Deletes several devices. """Deletes several devices.
Args: Args:
@ -1789,8 +1709,6 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore):
for device_id in device_ids: for device_id in device_ids:
self.device_id_exists_cache.invalidate((user_id, device_id)) self.device_id_exists_cache.invalidate((user_id, device_id))
return {}
async def update_device( async def update_device(
self, user_id: str, device_id: str, new_display_name: Optional[str] = None self, user_id: str, device_id: str, new_display_name: Optional[str] = None
) -> None: ) -> None:

View File

@ -20,8 +20,6 @@ from twisted.test.proto_helpers import MemoryReactor
from synapse.api.errors import NotFoundError, SynapseError from synapse.api.errors import NotFoundError, SynapseError
from synapse.handlers.device import MAX_DEVICE_DISPLAY_NAME_LEN, DeviceHandler from synapse.handlers.device import MAX_DEVICE_DISPLAY_NAME_LEN, DeviceHandler
from synapse.rest import admin
from synapse.rest.client import account, login
from synapse.server import HomeServer from synapse.server import HomeServer
from synapse.util import Clock from synapse.util import Clock
@ -32,12 +30,6 @@ user2 = "@theresa:bbb"
class DeviceTestCase(unittest.HomeserverTestCase): class DeviceTestCase(unittest.HomeserverTestCase):
servlets = [
login.register_servlets,
admin.register_servlets,
account.register_servlets,
]
def make_homeserver(self, reactor: MemoryReactor, clock: Clock) -> HomeServer: def make_homeserver(self, reactor: MemoryReactor, clock: Clock) -> HomeServer:
hs = self.setup_test_homeserver("server", federation_http_client=None) hs = self.setup_test_homeserver("server", federation_http_client=None)
handler = hs.get_device_handler() handler = hs.get_device_handler()
@ -123,7 +115,7 @@ class DeviceTestCase(unittest.HomeserverTestCase):
"device_id": "xyz", "device_id": "xyz",
"display_name": "display 0", "display_name": "display 0",
"last_seen_ip": None, "last_seen_ip": None,
"last_seen_ts": 1000000, "last_seen_ts": None,
}, },
device_map["xyz"], device_map["xyz"],
) )
@ -237,29 +229,6 @@ class DeviceTestCase(unittest.HomeserverTestCase):
NotFoundError, NotFoundError,
) )
def test_login_delete_old_devices(self) -> None:
"""Delete old devices if the user already has too many."""
user_id = self.register_user("user", "pass")
# Create a bunch of devices
for _ in range(50):
self.login("user", "pass")
self.reactor.advance(1)
# Advance the clock for ages (as we only delete old devices)
self.reactor.advance(60 * 60 * 24 * 300)
# Log in again to start the pruning
self.login("user", "pass")
# Give the background job time to do its thing
self.reactor.pump([1.0] * 100)
# We should now only have the most recent device.
devices = self.get_success(self.handler.get_devices_by_user(user_id))
self.assertEqual(len(devices), 1)
def _record_users(self) -> None: def _record_users(self) -> None:
# check this works for both devices which have a recorded client_ip, # check this works for both devices which have a recorded client_ip,
# and those which don't. # and those which don't.

View File

@ -170,8 +170,6 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase):
) )
) )
last_seen = self.clock.time_msec()
if after_persisting: if after_persisting:
# Trigger the storage loop # Trigger the storage loop
self.reactor.advance(10) self.reactor.advance(10)
@ -192,7 +190,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase):
"device_id": device_id, "device_id": device_id,
"ip": None, "ip": None,
"user_agent": None, "user_agent": None,
"last_seen": last_seen, "last_seen": None,
}, },
], ],
) )