Commit Graph

914 Commits

Author SHA1 Message Date
Erik Johnston
c71176858b Newline, remove debug logging 2015-05-18 10:11:14 +01:00
Erik Johnston
c3b37abdfd PEP8 2015-05-15 16:59:58 +01:00
Erik Johnston
6c74fd62a0 Revert limiting of fetching, it didn't help perf. 2015-05-15 16:45:35 +01:00
Erik Johnston
9ff7f66a2b init j 2015-05-15 16:36:03 +01:00
Erik Johnston
70f272f71c Don't completely drain the list 2015-05-15 16:34:17 +01:00
Erik Johnston
8763dd80ef Don't fetch prev_content for current_state 2015-05-15 15:33:01 +01:00
Erik Johnston
acb12cc811 Make store.get_current_state fetch events asyncly 2015-05-15 15:20:05 +01:00
Erik Johnston
d62dee7eae Remove more debug logging 2015-05-15 15:06:37 +01:00
Erik Johnston
0f29cfabc3 Remove debug logging 2015-05-15 14:06:42 +01:00
Erik Johnston
e275a9c0d9 preserve log context 2015-05-15 11:54:51 +01:00
Erik Johnston
aa32bd38e4 Add a wait 2015-05-15 11:35:04 +01:00
Erik Johnston
372d4c6d7b Srsly. Don't use closures. Baaaaaad 2015-05-15 11:26:00 +01:00
Erik Johnston
575ec91d82 Correctly pass through params 2015-05-15 11:15:10 +01:00
Erik Johnston
de01438a57 Sort out error handling 2015-05-15 11:00:50 +01:00
Erik Johnston
a2c4f3f150 Fix daedlock 2015-05-15 10:54:04 +01:00
Erik Johnston
1d566edb81 Remove race condition 2015-05-14 16:54:35 +01:00
Erik Johnston
ef3d8754f5 Call from right thread 2015-05-14 15:41:55 +01:00
Erik Johnston
142934084a Count and loop 2015-05-14 15:40:21 +01:00
Erik Johnston
96c5b9f87c Don't start up more fetch_events 2015-05-14 15:36:04 +01:00
Erik Johnston
7cd6a6f6cf Awful idea for speeding up fetching of events 2015-05-14 15:34:02 +01:00
Erik Johnston
7f4105a5c9 Turn off preemptive transactions 2015-05-14 14:51:06 +01:00
Erik Johnston
f4d58deba1 PEP8 2015-05-14 14:45:42 +01:00
Erik Johnston
386b7330d2 Move from _base to events 2015-05-14 14:45:22 +01:00
Erik Johnston
7d6a1dae31 Jump out early 2015-05-14 14:27:58 +01:00
Erik Johnston
656223fbd3 Actually, we probably want to run this in a transaction 2015-05-14 14:26:35 +01:00
Erik Johnston
2f7f8e1c2b Preemptively jump into a transaction if we ask for get_prev_content 2015-05-14 14:17:36 +01:00
Erik Johnston
e1e9f0c5b2 loop -> gatherResults 2015-05-14 13:58:49 +01:00
Erik Johnston
ab78a8926e Err, we probably want a bigger limit 2015-05-14 13:47:16 +01:00
Erik Johnston
f6f902d459 Move fetching of events into their own transactions 2015-05-14 13:45:48 +01:00
Erik Johnston
cdb3757942 Refactor _get_events 2015-05-14 13:31:55 +01:00
Erik Johnston
36ea26c5c0 Merge branch 'develop' of github.com:matrix-org/synapse into erikj/perf 2015-05-14 12:01:38 +01:00
David Baker
7c549dd557 Add ID generator for push_rules_enable to #resolve SYN-378 2015-05-14 11:44:03 +01:00
Mark Haines
899d4675dd Merge branch 'notifier_unify' into notifier_performance 2015-05-14 11:36:44 +01:00
Erik Johnston
4f1d984e56 Add index on events 2015-05-13 17:22:26 +01:00
Erik Johnston
968b01a91a Actually use async method 2015-05-13 17:02:46 +01:00
Erik Johnston
4071f29653 Fetch events from events_id in their own transactions 2015-05-13 16:59:41 +01:00
Erik Johnston
a988361aea Typo 2015-05-13 15:44:15 +01:00
Erik Johnston
8888982db3 Don't insert None 2015-05-13 15:43:32 +01:00
Erik Johnston
cf706cc6ef Don't return None 2015-05-13 15:31:25 +01:00
Erik Johnston
5971d240d4 Limit batch size 2015-05-13 15:26:49 +01:00
Erik Johnston
ca4f458787 Fetch events in bulk 2015-05-13 15:13:42 +01:00
Erik Johnston
6edff11a88 Don't fetch redaction and rejection stuff for each event, so we can use index only scan 2015-05-13 14:39:05 +01:00
Mark Haines
63878c0379 Don't bother checking for updates if the stream token hasn't advanced for a user 2015-05-13 13:42:21 +01:00
Erik Johnston
02590c3e1d Temp turn off checking for rejections and redactions 2015-05-13 11:31:28 +01:00
Erik Johnston
619a21812b defer.gatherResults loop 2015-05-13 11:29:03 +01:00
Erik Johnston
fec4485e28 Batch fetching of events for state groups 2015-05-13 11:22:42 +01:00
Erik Johnston
409bcc76bd Load events for state group seperately 2015-05-13 11:13:31 +01:00
Erik Johnston
e122685978 You need to call contextmanager 2015-05-12 16:12:37 +01:00
Mark Haines
a6fb2aa2a5 Merge pull request #144 from matrix-org/erikj/logging_context
Preserving logging contexts
2015-05-12 15:23:50 +01:00
Mark Haines
1fce36b111 Merge pull request #149 from matrix-org/erikj/backfill
Backfill support
2015-05-12 15:20:32 +01:00
Erik Johnston
8b28209c60 Err, delete the right stuff 2015-05-12 15:02:53 +01:00
Erik Johnston
30c72d377e Newlines 2015-05-12 14:47:40 +01:00
Erik Johnston
e4eddf9b36 We do actually want to delete rows out of event_backward_extremities 2015-05-12 14:47:23 +01:00
Erik Johnston
c1779a79bc Fix up _handle_prev_events to not try to insert duplicate rows 2015-05-12 14:41:50 +01:00
Erik Johnston
74850d7f75 Do state groups persistence /after/ checking if we have already persisted the event 2015-05-12 14:14:58 +01:00
Erik Johnston
da6a7bbdde Merge branch 'develop' of github.com:matrix-org/synapse into erikj/logging_context 2015-05-12 13:10:42 +01:00
Mark Haines
5002056b16 SYN-377: Make sure that the StreamIdGenerator.get_next.__exit__ is called from the main thread after the transaction completes, not from database thread before the transaction completes. 2015-05-12 11:20:40 +01:00
Erik Johnston
4df11b5039 Make get_current_token accept a direction parameter, which tells whether the source whether we want a token for going 'forwards' or 'backwards' 2015-05-12 10:28:10 +01:00
Erik Johnston
84e6b4001f Initial hack at wiring together pagination and backfill 2015-05-11 18:01:31 +01:00
Erik Johnston
17653a5dfe Move storage.stream._StreamToken to types.RoomStreamToken 2015-05-11 18:01:01 +01:00
Erik Johnston
b036596b75 Prefer to use _simple_*. 2015-05-11 11:24:01 +01:00
Erik Johnston
cd525c0f5a push_rules table expects an 'id' field 2015-05-11 11:24:01 +01:00
Erik Johnston
22d2f498fa Fix push rule bug: can't insert bool into small int column 2015-05-10 10:50:51 +01:00
Erik Johnston
476899295f Change the way we do logging contexts so that they survive divergences 2015-05-08 16:32:18 +01:00
Mark Haines
1c7912751e Drop the old table not the new table 2015-05-08 16:04:32 +01:00
Mark Haines
9d36eb4eab Rename unique constraint 2015-05-08 16:01:55 +01:00
Mark Haines
b0f71db3ff Remove unsigned 2015-05-08 15:59:51 +01:00
Mark Haines
84e1cacea4 Bump schema version 2015-05-08 15:58:14 +01:00
Mark Haines
6538d445e8 Make the timestamps in server_keys_json bigints 2015-05-08 15:55:17 +01:00
Erik Johnston
0c4ac271ca Merge branch 'erikj/executemany' of github.com:matrix-org/synapse into erikj/SYN-371 2015-05-05 18:21:19 +01:00
Erik Johnston
0cf7e480b4 And use buffer(...) there as well 2015-05-05 18:20:01 +01:00
Erik Johnston
ed2584050f Merge branch 'develop' of github.com:matrix-org/synapse into erikj/executemany 2015-05-05 18:15:20 +01:00
Erik Johnston
977338a7af Use buffer(...) when inserting into bytea column 2015-05-05 18:12:53 +01:00
Mark Haines
31049c4d72 Merge pull request #139 from matrix-org/bugs/SYN-369
Fix race with cache invalidation. SYN-369
2015-05-05 17:46:13 +01:00
Mark Haines
deb0237166 Add some doc-string 2015-05-05 17:45:11 +01:00
Mark Haines
d18f37e026 Collect the invalidate callbacks on the transaction object rather than passing around a separate list 2015-05-05 17:32:21 +01:00
Erik Johnston
9951542393 Add a comment about the zip(*[zip(sorted(...),...)]) 2015-05-05 17:06:55 +01:00
Mark Haines
041b6cba61 SYN-369: Add comments to the sequence number logic in the cache 2015-05-05 16:32:44 +01:00
Mark Haines
63075118a5 Add debug flag in synapse/storage/_base.py for debugging the cache logic by comparing what is in the cache with what was in the database on every access 2015-05-05 16:24:04 +01:00
Erik Johnston
531d7955fd Don't insert without deduplication. In this case we never actually use this table, so simply remove the insert entirely 2015-05-05 16:12:28 +01:00
Mark Haines
bfa4a7f8b0 Invalidate the room_member cache if the current state events updates 2015-05-05 15:43:49 +01:00
Mark Haines
d0fece8d3c Missing return for when the event was already persisted 2015-05-05 15:39:09 +01:00
Erik Johnston
bdcd7693c8 Fix indentation 2015-05-05 15:14:48 +01:00
Erik Johnston
43c2e8deae Add support for using executemany 2015-05-05 15:13:25 +01:00
Erik Johnston
1692dc019d Don't call 'encode_parameter' no-op 2015-05-05 15:00:30 +01:00
Mark Haines
a9aea68fd5 Invalidate the caches from the correct thread 2015-05-05 14:57:08 +01:00
Mark Haines
261d809a47 Sequence the modifications to the cache so that selects don't race with inserts 2015-05-05 14:13:50 +01:00
Erik Johnston
d9cc5de9e5 Correctly name transaction 2015-05-05 10:24:10 +01:00
Erik Johnston
b8940cd902 Remove some unused indexes 2015-05-01 16:14:25 +01:00
David Baker
eb9bd2d949 user_id now in user_threepids 2015-05-01 15:04:37 +01:00
David Baker
9eeb03c0dd Don't use self.execute: it's designed for fetching stuff 2015-05-01 14:21:25 +01:00
David Baker
7693f24792 No id field on user 2015-05-01 13:55:42 +01:00
David Baker
92b20713d7 More missed get_user_by_id API changes 2015-05-01 13:45:54 +01:00
Erik Johnston
4011cf1c42 Cache latest_event_ids_in_room 2015-05-01 13:06:26 +01:00
Erik Johnston
657298cebd Don't lock user_ips table for upsert. 2015-05-01 10:46:48 +01:00
Erik Johnston
fabb7acd45 Fix bug where we reconnected to the database on every query. 2015-05-01 10:24:24 +01:00
Erik Johnston
23c639ff32 Split a storage function in two so that we don't have to do extra work. 2015-05-01 10:17:19 +01:00
Erik Johnston
8be5284e91 Remove pointless join 2015-04-30 18:49:26 +01:00
Erik Johnston
503e4d3d52 Fix broken SQL 2015-04-30 18:44:47 +01:00
Erik Johnston
00718ae7a9 Need more yield 2015-04-30 18:43:39 +01:00
Erik Johnston
0465560c1a Add missing param 2015-04-30 18:42:44 +01:00
Erik Johnston
61d05daab1 More join conditions 2015-04-30 18:41:05 +01:00
Erik Johnston
6ead27ddda Add more conditions on JOINs to make postgres go a little faster. 2015-04-30 18:32:03 +01:00
Erik Johnston
f383d5a801 Fix up get_current_state and get_room_name_and_aliases queries to parse events in transaction 2015-04-30 17:12:52 +01:00
Erik Johnston
69d4063651 Add get_rooms_for_user cache 2015-04-30 16:47:51 +01:00
Erik Johnston
3c4c229788 Don't use sub queries, it makes postgres sad 2015-04-30 16:16:53 +01:00
Mark Haines
29400b45b9 SYN-367: Use upsert rather than insert_or_replace 2015-04-30 15:21:31 +01:00
Erik Johnston
2366d28780 Don't needlessly join on state_events 2015-04-30 14:02:06 +01:00
Erik Johnston
1d7702833d Make simple query rather than long one and then throw away half the results 2015-04-30 10:16:12 +01:00
Erik Johnston
b1ca784aca Correctly decode, for sqlite and postgres, rows from pushers table 2015-04-29 19:41:14 +01:00
Erik Johnston
4a9dc5b2f5 pushkey' are also bytes. 2015-04-29 19:27:02 +01:00
Erik Johnston
0ade2712d1 Typo 2015-04-29 19:17:25 +01:00
Erik Johnston
50f96f256f Also remove yield from within lock in the other generator 2015-04-29 19:17:00 +01:00
Erik Johnston
d2d61a8288 Fix deadlock in id_generators. No idea why this was an actual deadlock. 2015-04-29 19:15:23 +01:00
Erik Johnston
3864b3a8e6 Actually return rows 2015-04-29 18:07:36 +01:00
Erik Johnston
0618978238 Typo, args wrong way round 2015-04-29 18:04:35 +01:00
Erik Johnston
09177f4f2e Decode buffers in same thread 2015-04-29 18:03:42 +01:00
Erik Johnston
472be88674 We store pusher data as bytes 2015-04-29 17:43:46 +01:00
David Baker
12d381bd5d Decode the data json in the storage layer (was moved but this part was missed) 2015-04-29 17:13:51 +01:00
David Baker
f8c30faf25 Oops, update the contraint too 2015-04-29 16:58:42 +01:00
David Baker
61cd5d9045 Be more postgressive 2015-04-29 16:57:14 +01:00
David Baker
fb95035a65 Be postgressive 2015-04-29 16:53:41 +01:00
David Baker
4669def000 Oops, forgot the schema delta file 2015-04-29 16:50:16 +01:00
Erik Johnston
0337eaf321 txn.execute doesn't return cursors 2015-04-29 16:43:39 +01:00
Erik Johnston
884fb88e28 txn.execute doesn't return cursors 2015-04-29 16:35:20 +01:00
Erik Johnston
d76c058eea Fix invalid SQL to work in postgres land 2015-04-29 16:30:25 +01:00
Erik Johnston
c8ed9bd278 pushers table requires an access_token 2015-04-29 15:33:27 +01:00
Erik Johnston
970a9b9d2b We can't use REPLACE when upgrading databases now we have postgres 2015-04-29 13:55:44 +01:00
Erik Johnston
64991b0c8b Merge pull request #129 from matrix-org/key_distribution
Key distribution v2
2015-04-29 13:34:38 +01:00
Mark Haines
e26a3d8d9e bump database schema version 2015-04-29 13:32:32 +01:00
Mark Haines
a9549fdce3 Use bytea rather than BLOB 2015-04-29 13:16:09 +01:00
Mark Haines
4ad8b45155 Merge branch 'develop' into key_distribution
Conflicts:
	synapse/config/homeserver.py
2015-04-29 13:15:14 +01:00
Erik Johnston
cd0864121b Make postgres database error slightly more helpful 2015-04-29 12:12:25 +01:00
Erik Johnston
4932a7e2d9 Use __all__ instead of assert to stop pyflakes from warning about unused import in __init__ 2015-04-29 12:12:25 +01:00
Erik Johnston
a08bf11138 Appease PEP8 2015-04-29 11:44:48 +01:00
Erik Johnston
204132a998 Check that postgres database has correct charset set 2015-04-29 11:42:28 +01:00
Erik Johnston
327ca883ec Merge branch 'develop' of github.com:matrix-org/synapse into postgres 2015-04-28 13:39:42 +01:00
Erik Johnston
18f8247701 Use TEXT instead of VARCHAR(n), since PostgreSQL treats them the same except for a limit 2015-04-28 12:41:33 +01:00
Mark Haines
9182f87664 Merge pull request #126 from matrix-org/csauth
Client / Server Auth Refactor
2015-04-28 11:00:27 +01:00
Mark Haines
55e1bc8920 And don't bump the schema version unnecessarily 2015-04-28 10:54:15 +01:00
Mark Haines
b96c133034 Add server_keys.sql to the current delta rather than creating a new delta 2015-04-28 10:50:00 +01:00
Erik Johnston
8558e1ec73 Make get_max_token into inlineCallbacks so that the lock works. 2015-04-27 15:19:44 +01:00
Erik Johnston
6f8e2d517e Merge branch 'develop' of github.com:matrix-org/synapse into postgres 2015-04-27 14:41:40 +01:00
Erik Johnston
2732be83d9 Shuffle operations so that locking upsert happens last in the txn. This ensures the lock is held for the least amount of time possible. 2015-04-27 13:22:30 +01:00
Erik Johnston
e4c4664d73 Handle the fact that postgres databases can be restarted from under us 2015-04-27 12:40:49 +01:00
David Baker
8a5be236e0 pep8 2015-04-27 11:49:18 +01:00
David Baker
b02e1006b9 Run database check before daemonizing, at the cost of database hygiene. 2015-04-27 11:46:00 +01:00
David Baker
f8152f2708 rename db method to be more informative 2015-04-27 10:16:26 +01:00
David Baker
2f475bd5d5 pep8 2015-04-24 18:15:07 +01:00
David Baker
a7b51f4539 Check users in our table aren't on a different domain to the one we're configured with to try & fix SYN-266 2015-04-24 18:11:21 +01:00