71b625d808
Stop backpaginating when events not visible
2019-02-20 18:14:12 +00:00
a06614bd2a
UPSERT many functionality ( #4644 )
2019-02-20 23:03:30 +11:00
c003450057
Merge pull request #4671 from matrix-org/erikj/state_cache_invalidation
...
Batch cache invalidation over replication
2019-02-19 13:14:30 +00:00
62175a20e5
Docs
2019-02-19 11:38:40 +00:00
1bb35e3a83
Use itertools
2019-02-19 11:34:40 +00:00
bc8fa1509d
Documentation
2019-02-19 11:24:59 +00:00
a9b5ea6fc1
Batch cache invalidation over replication
...
Currently whenever the current state changes in a room invalidate a lot
of caches, which cause *a lot* of traffic over replication. Instead,
lets batch up all those invalidations and send a single poke down
the replication streams.
Hopefully this will reduce load on the master process by substantially
reducing traffic.
2019-02-18 17:53:31 +00:00
4cc4400b4d
Split /login into client_reader
2019-02-18 17:19:01 +00:00
eb2b8523ae
Split out registration to worker
...
This allows registration to be handled by a worker, though the actual
write to the database still happens on master.
Note: due to the in-memory session map all registration requests must be
handled by the same worker.
2019-02-18 12:12:57 +00:00
eaf4d11af9
Add configurable room list publishing rules
...
This allows specifying who and what is allowed to be published onto the
public room list
2019-02-14 16:02:23 +00:00
bb4fd8f927
Run `black` on user directory code ( #4635 )
2019-02-13 23:05:32 +11:00
cf82338930
Merge pull request #4627 from matrix-org/erikj/user_ips_analyze
...
Analyze user_ips before running deduplication
2019-02-12 13:05:09 +00:00
495ea92350
Fix pep8
2019-02-12 12:40:42 +00:00
483ba85c7a
Analyze user_ips before running deduplication
...
Due to the table locks taken out by the naive upsert, the table
statistics may be out of date. During deduplication it is important that
the correct index is used as otherwise a full table scan may be
incorrectly used, which can end up thrashing the database badly.
2019-02-12 11:55:27 +00:00
362d80b770
Reduce user_ips bloat during dedupe background update
...
The background update to remove duplicate rows naively deleted and
reinserted the duplicates. For large tables with a large number of
duplicates this causes a lot of bloat (with postgres), as the inserted
rows are appended to the table, since deleted rows will not be
overwritten until a VACUUM has happened.
This should hopefully also help ensure that the query in the last batch
uses the correct index, as inserting a large number of new rows without
analyzing will upset the query planner.
2019-02-12 11:39:34 +00:00
719e073f00
Merge pull request #4580 from matrix-org/uhoreg/e2e_backup_add_updating
...
add updating of backup versions
2019-02-11 13:45:49 +00:00
56710c7df5
Fix 'no unique or exclusion constraint' error ( #4591 )
...
Add more tables to the list of tables which need a background update to
complete before we can upsert into them, which fixes a race against the
background updates.
2019-02-08 18:30:46 +00:00
8248637173
add new endpoint to update backup versions
2019-02-06 17:57:10 -05:00
627ecd358e
Filter user directory state query to a subset of state events ( #4462 )
...
* Filter user directory state query to a subset of state events
* Add changelog
2019-02-05 12:16:28 +00:00
c45fd0dda0
Merge pull request #4530 from matrix-org/anoa/room_upgrade_federatable
...
Copy over non-federatable trait on room upgrade
2019-01-31 20:28:37 +00:00
d239f67c25
Raise an exception instead of returning None
2019-01-31 18:34:15 +00:00
3ed3cb4339
New function for getting room's create event
2019-01-31 18:21:39 +00:00
e707e7b38d
Fix infinite loop when an event is redacted in a v3 room ( #4535 )
2019-01-31 15:34:17 +00:00
3f189c902e
Fix flake8 ( #4519 )
2019-01-30 10:53:17 +00:00
a696c48133
Merge branch 'develop' of github.com:matrix-org/synapse into erikj/redactions_eiah
2019-01-29 22:00:33 +00:00
c21b7cbc09
Update synapse/storage/events_worker.py
2019-01-29 21:53:48 +00:00
4db252c073
Check redaction state when event is pulled out of the database
2019-01-29 21:48:36 +00:00
6d23ec2111
Fix typo
2019-01-29 21:45:53 +00:00
b5d510ad64
Remove unused arg
2019-01-29 21:45:28 +00:00
82165eeb05
Update synapse/storage/events_worker.py
...
Co-Authored-By: erikjohnston <erikj@jki.re>
2019-01-29 21:14:39 +00:00
7709d2bd16
Implement rechecking of redactions
2019-01-29 11:56:20 +00:00
be47cfa9c9
Refactor event building into EventBuilder
...
This is so that everything is done in one place, making it easier to
change the event format based on room version
2019-01-29 11:13:00 +00:00
94fb63e44f
Fix typo in upserts code ( #4505 )
...
* fix obvious problem :|
* changelog
2019-01-29 10:04:23 +00:00
7072fe3084
Fix UPSERTs on SQLite 3.24+ ( #4477 )
2019-01-28 15:43:32 +00:00
4026d555fa
Merge branch 'develop' of github.com:matrix-org/synapse into anoa/dm_room_upgrade
2019-01-28 14:08:24 +00:00
b6dce9b9fd
Merge pull request #4470 from matrix-org/erikj/require_format_version
...
Require event format version to parse or create events
2019-01-25 15:59:36 +00:00
431e485914
Merge branch 'develop' of github.com:matrix-org/synapse into erikj/sqlite_native_upsert
2019-01-25 14:11:17 +00:00
0b3fd1401f
Don't require sqlite3 when using postgres ( #4466 )
2019-01-25 22:25:02 +11:00
821b65aeb5
Merge branch 'develop' of github.com:matrix-org/synapse into anoa/dm_room_upgrade
2019-01-25 11:09:53 +00:00
b1b6dba2d2
Merge pull request #4415 from matrix-org/anoa/full_search_upgraded_rooms
...
Ability to search entire room history after upgrading room
2019-01-25 10:49:58 +00:00
1953067136
Disable native upserts for sqlite, as they don't work
2019-01-25 10:46:49 +00:00
678a92cb56
Replace missed usages of FrozenEvent
2019-01-25 10:32:30 +00:00
be6a7e47fa
Revert "Require event format version to parse or create events"
2019-01-25 10:23:51 +00:00
efb8ed1d45
Merge branch 'develop' of github.com:matrix-org/synapse into erikj/require_format_version
2019-01-24 18:52:34 +00:00
80bcca659e
Merge pull request #4405 from matrix-org/erikj/fixup_rejecting_invites
...
Store rejected remote invite events as outliers
2019-01-24 18:33:27 +00:00
664b7a2920
Merge pull request #4437 from matrix-org/erikj/event_format_version_v2
...
Add support for persisting event format versions
2019-01-24 18:32:30 +00:00
b8082a5445
Use term 'out of band membership' instead
2019-01-24 17:33:19 +00:00
a383289b0d
Merge branch 'anoa/full_search_upgraded_rooms' of github.com:matrix-org/synapse into anoa/full_search_upgraded_rooms
2019-01-24 17:23:51 +00:00
e1781b043b
Remove redundant create event None check
2019-01-24 17:23:39 +00:00
03c85335d1
Apply suggestions from code review
...
Co-Authored-By: anoadragon453 <1342360+anoadragon453@users.noreply.github.com>
2019-01-24 17:22:09 +00:00
0e27501ee5
Fix UPSERT check ( #4459 )
2019-01-24 22:57:41 +11:00
e8c9f15397
Replace missed usages of FrozenEvent
2019-01-24 11:14:07 +00:00
58f6c48183
Use native UPSERTs where possible ( #4306 )
2019-01-24 21:31:54 +11:00
17898a5ba6
Merge branch 'erikj/fixup_rejecting_invites' of github.com:matrix-org/synapse into erikj/require_format_version
2019-01-23 20:13:07 +00:00
7c288c2250
Clarify the invite flows
2019-01-23 20:07:47 +00:00
886e5acc76
Store rejected remote invite events as outliers
...
Currently they're stored as non-outliers even though the server isn't in
the room, which can be problematic in places where the code assumes it
has the state for all non outlier events.
In particular, there is an edge case where persisting the leave event
triggers a state resolution, which requires looking up the room version
from state. Since the server doesn't have the state, this causes an
exception to be thrown.
2019-01-23 20:04:57 +00:00
2a8edbaf74
Merge branch 'erikj/get_pdu_versions' into erikj/require_format_version
2019-01-23 17:27:49 +00:00
be1065af59
isort
2019-01-23 11:48:16 +00:00
c5a296b10c
Add support for persisting event format versions
...
Currently we only have the one event format version defined, but this
adds the necessary infrastructure to persist and fetch the format
versions alongside the events.
We specify the format version rather than the room version as:
1. We don't necessarily know the room version, existing events may be
either v1 or v2.
2. We'd need to be careful to prevent/handle correctly if different
events in the same room reported to be of different versions, which
sounds annoying.
2019-01-23 11:30:01 +00:00
90743c9d89
Fixup removal of duplicate `user_ips` rows ( #4432 )
...
* Remove unnecessary ORDER BY clause
* Add logging
* Newsfile
2019-01-23 19:45:18 +11:00
7f503f83b9
Refactor to rewrite the SQL instead
2019-01-22 16:31:05 +00:00
1c9704f8ab
Don't shadow params
2019-01-22 16:20:33 +00:00
2557531f0f
Fix bug when removing duplicate rows from user_ips
...
This was caused by accidentally overwritting a `last_seen` variable
in a for loop, causing the wrong value to be written to the progress
table. The result of which was that we didn't scan sections of the table
when searching for duplicates, and so some duplicates did not get
deleted.
2019-01-22 13:33:46 +00:00
277e50462d
Do not return in a deferred function
2019-01-22 12:40:26 +00:00
c433f61091
Ensure new filter is actually created
2019-01-22 12:06:36 +00:00
c9bfb058d8
Fix a bug with single-room search searching all rooms
...
* Create a new method for getting predecessor rooms
* Remove formatting change
2019-01-22 12:00:41 +00:00
25d64a846a
Fix typos
2019-01-22 11:00:04 +00:00
25dd56ace3
Fix race when persisting create event ( #4404 )
...
* Fix race when persisting create event
When persisting a chunk of DAG it is sometimes requried to do a state
resolution, which requires knowledge of the room version. If this
happens while we're persisting the create event then we need to use that
event rather than attempting to look it up in the database.
2019-01-18 23:17:04 +11:00
df3a661e4a
Search for messages across predecessor rooms
...
Signed-off-by: Andrew Morgan <andrew@amorgan.xyz>
2019-01-18 11:19:20 +00:00
a35c66a00b
Remove duplicates in the user_ips table and add an index ( #4370 )
2019-01-12 06:21:50 +11:00
7960c26fda
Fix adding new rows instead of updating them if one of the key values is a NULL in upserts. ( #4369 )
2019-01-09 22:26:25 +11:00
84b6fae1f5
Ensure synchrotrons can access is_support_user in the storage layer
2019-01-02 10:19:59 +00:00
0708f437cc
Log roomid along with Unknown room ( #4297 )
2018-12-24 10:49:42 +01:00
d2f7c4e6b1
create support user ( #4141 )
...
Allow for the creation of a support user.
A support user can access the server, join rooms, interact with other users, but does not appear in the user directory nor does it contribute to monthly active user limits.
2018-12-14 18:20:59 +00:00
e93a0ebf50
Settings Fix deleting e2e room keys on xenial ( #4295 )
2018-12-14 05:10:31 +11:00
30da50a5b8
Initialise user displayname from SAML2 data ( #4272 )
...
When we register a new user from SAML2 data, initialise their displayname
correctly.
2018-12-07 14:44:46 +01:00
9a3e24a13d
drop undocumented dependency on dateutil ( #4266 )
...
It turns out we were relying on dateutil being pulled in transitively by
pysaml2. There's no need for that bloat.
2018-12-06 04:52:42 +11:00
158ffb92f1
Add an option to disable search for homeservers which may not be interested in it ( #4230 )
...
This is useful for homeservers not intended for users, such as bot-only homeservers or ones that only process IoT data.
2018-12-04 12:01:02 +01:00
b5ac0ffa0a
add more detail to logging regarding "More than one row matched" error ( #4234 )
2018-12-04 11:57:39 +01:00
ecc23188f4
Fix UnicodeDecodeError when postgres is not configured in english ( #4253 )
...
This is a bit of a half-assed effort at fixing https://github.com/matrix-org/synapse/issues/4252 . Fundamentally the right answer is to drop support for Python 2.
2018-12-04 11:55:52 +01:00
704c5298f0
Drop sent_transactions
...
Signed-off-by: Aaron Raimist <aaron@raim.ist>
2018-12-01 23:07:35 -06:00
f9b136a886
Neilj/fix mau initial reserved users ( #4211 )
...
* fix transaction wrapping bug that caused get_user_id_by_threepid_txn to fail
* towncrier
* white space
2018-11-28 20:33:41 +11:00
6c18cc4b50
Ignore __pycache__ directories in schema delta dir
...
Now that we use py3, compiled python ends up in __pycache__ rather than *.pyc.
2018-11-20 22:52:34 +00:00
835779f7fb
Add option to track MAU stats (but not limit people) ( #3830 )
2018-11-15 18:08:27 +00:00
bca3b91c2d
Merge remote-tracking branch 'origin/develop' into dbkr/e2e_backup_versions_are_numbers
2018-11-09 18:35:02 +00:00
d44dea0223
pep8
2018-11-09 14:38:31 +00:00
4f93abd62d
add docs
2018-11-09 13:25:38 +00:00
5ebed18692
Lets convert bytes to unicode instead
2018-11-08 12:33:13 +00:00
b1a22b24ab
Fix noop checks when updating device keys
...
Clients often reupload their device keys (for some reason) so its
important for the server to check for no-ops before sending out device
list update notifications.
The check is broken in python 3 due to the fact comparing bytes and
unicode always fails, and that we write bytes to the DB but get unicode
when we read.
2018-11-08 12:18:38 +00:00
bc80b3f454
Add helpers for getting prev and auth events ( #4139 )
...
* Add helpers for getting prev and auth events
This is in preparation for allowing the event format to change between
room versions.
2018-11-06 00:35:15 +11:00
350f654e7b
Add unique indexes to a couple of tables
...
The indexes on device_lists_remote_extremeties can be unique, and they
therefore should, to ensure that the db remains consistent.
2018-11-02 10:36:13 +00:00
50e328d1e7
Remove redundant database locks for device list updates
...
We can rely on the application-level per-user linearizer.
2018-11-02 10:36:13 +00:00
0794504bce
Merge pull request #4006 from matrix-org/erikj/purge_state_groups
...
Delete unreferenced state groups during purge
2018-10-30 16:58:22 +00:00
e0934acdbb
Cast to int here too
2018-10-30 11:12:23 +00:00
12941f5f8b
Cast bacjup version to int when querying
2018-10-30 11:01:07 +00:00
2f0f911c52
Convert version back to a string
2018-10-30 10:35:18 +00:00
64fa557f80
Try & make it work on postgres
2018-10-30 09:51:04 +00:00
563f9b61b1
Make e2e backup versions numeric in the DB
...
We were doing max(version) which does not do what we wanted
on a column of type TEXT.
2018-10-29 21:01:22 +00:00
88e5ffe6fe
Deduplicate device updates sent over replication
...
We currently send several kHz of device list updates over replication
occisonally, which often causes the replications streams to lag and then
get dropped.
A lot of those updates will actually be duplicates, since we don't send
e.g. device_ids across replication, so let's deduplicate it when we pull
them out of the database.
2018-10-29 17:34:34 +00:00
ad88460e0d
Move _find_unreferenced_groups
2018-10-29 14:24:19 +00:00
Erik Johnston