c0d6045776
It should be all
2017-02-27 18:45:24 +00:00
49f4bc4709
Don't fetch current state in common case
...
Currently we fetch the list of current state events whenever we send
something in a room. This is overkill for the common case of persisting
a simple chain of non-state events, so lets handle that case specially.
2017-02-27 18:33:41 +00:00
7723b4caa4
Ignore new rejected events when working out forward extremeties.
2017-02-08 14:48:06 +00:00
ebfaff84c9
Merge pull request #1870 from matrix-org/erikj/cache_get_all_new_events
...
Add a small cache get_all_new_events
2017-02-01 13:22:02 +00:00
368c88c487
Add a small cache get_all_new_events
2017-02-01 10:50:44 +00:00
458b6f4733
Only invalidate membership caches based on the cache stream
...
Before we completely invalidated get_users_in_room whenever we updated
any current_state_events table. This was way too aggressive.
2017-01-31 16:09:03 +00:00
4b3403ca9b
Stream cache invalidations for room membership storage functions
2017-01-30 17:28:22 +00:00
fdf2a31a51
Typo
2017-01-23 16:14:14 +00:00
c77b24c092
Refactor to calculate state delta outside transaction
2017-01-23 14:51:33 +00:00
5d2134d485
Comments
2017-01-20 17:13:24 +00:00
a55fa2047f
Insert delta of current_state_events to be more efficient
2017-01-20 17:10:18 +00:00
d0897dead5
Spelling
2017-01-20 15:05:11 +00:00
567aa35b67
Update all call sites after rename
2017-01-20 14:40:31 +00:00
f2f40e64a9
Comments
2017-01-20 14:38:13 +00:00
4c6a31cd6e
Calculate the forward extremeties once
2017-01-20 14:28:53 +00:00
09eb08f910
Derive current_state_events from state groups
2017-01-20 11:52:51 +00:00
4ec1cf49e2
Lower loading events log to DEBUG
2017-01-17 17:28:32 +00:00
dd3df11c55
More logging for the linearizer and for get_events
2017-01-05 12:32:47 +00:00
f9d156d270
New Flake8 fixes
2016-11-15 11:22:29 +00:00
816988baaa
Merge branch 'develop' of github.com:matrix-org/synapse into erikj/remove_auth
2016-10-17 11:10:37 +01:00
2869a29fd7
Drop some unused indices
2016-10-17 11:08:19 +01:00
9040c9ffa1
Fix background reindex of origin_server_ts
...
The storage function `_get_events_txn` was removed everywhere except
from this background reindex. The function was removed due to it being
(almost) completely unused while also being large and complex.
Therefore, instead of resurrecting `_get_events_txn` we manually
reimplement the bits that are needed directly.
2016-09-27 11:23:49 +01:00
0b32bb20bb
Index contains_url for file search queries
2016-09-12 16:57:05 +01:00
70332a12dd
Take value in a better way
2016-09-05 14:57:14 +01:00
485d999c8a
Correctly delete old state groups in purge history API
2016-09-05 14:49:08 +01:00
598317927c
Limit the length of state chains
2016-09-02 10:41:38 +01:00
1bb8ec296d
Generate state group ids in state layer
2016-08-31 10:09:46 +01:00
5dc2a702cf
Make _state_groups_id_gen a normal IdGenerator
2016-08-30 16:55:11 +01:00
bed10f9880
Use state handler instead of get_users_in_room/get_joined_hosts
2016-08-26 14:54:30 +01:00
9219139351
Preserve some logcontexts
2016-08-24 11:58:40 +01:00
37adde32dc
Move defer.returnValue out of Measure
2016-08-19 18:23:44 +01:00
47dd8f02a1
Measure _get_event_from_row
2016-08-19 18:23:44 +01:00
8a57cc3123
Add missing database corruption recovery case
...
Signed-off-by: Benjamin Saunders <ben.e.saunders@gmail.com>
2016-08-14 11:50:22 -07:00
a8b946decb
Raise 404 when couldn't find event
2016-08-05 15:31:02 +01:00
f0fa66f495
Delete more tables
2016-08-05 10:40:08 +01:00
a2b7102eea
Tweak integrity error recovery to work as intended
2016-08-04 20:38:08 -07:00
8ad8490cff
Fix typo
2016-08-04 15:21:29 +01:00
59fa91fe88
Retry event persistence on IntegrityError
...
Due to a bug in the porting script some backfilled events were not
correctly persisted, causing irrecoverable IntegrityErrors on future
attempts to persist those events.
This commit adds a retry mechanism invoked upon IntegrityError,
where when retried the tables are purged for all references to the
events being persisted.
2016-08-04 15:02:15 +01:00
80ad710217
Remove other bit of deduplication
2016-08-03 13:25:59 +01:00
a8a32d2714
Ensure we only persist an event once at a time
2016-08-03 11:23:39 +01:00
a6f06ce3e2
Fix how push_actions are redacted.
2016-07-26 11:05:39 +01:00
efeb6176c1
Don't add rejected events if we've seen them befrore. Add some comments to explain what the code is doing mechanically
2016-07-26 10:49:52 +01:00
1b3c3e6d68
Only update the events and event_json tables for rejected events
2016-07-25 18:44:30 +01:00
8f7f4cb92b
Don't add the events to forward extremities if the event is rejected
2016-07-25 17:13:37 +01:00
2623cec874
Don't add rejections to the state_group, persist all rejections
2016-07-25 16:12:16 +01:00
1e2a7f18a1
Merge pull request #922 from matrix-org/erikj/file_api2
...
Feature: Add filter to /messages. Add 'contains_url' to filter.
2016-07-20 10:40:48 +01:00
f52565de50
Fix /purge_history bug
...
This was caused by trying to insert duplicate backward extremeties
2016-07-15 14:23:15 +01:00
eec9609e96
event_backwards_extremeties may not be empty
2016-07-15 10:22:09 +01:00
b64aa6d687
Add sender and contains_url field to events table
2016-07-14 15:35:43 +01:00
b92e7955be
Comment
2016-07-07 11:42:15 +01:00
c98e1479bd
Return 400 rather than 500
2016-07-07 11:41:07 +01:00
67f2c901ea
Add rest servlet. Fix SQL.
2016-07-06 15:56:59 +01:00
2d21d43c34
Add purge_history API
2016-07-05 10:28:51 +01:00
a67bf0b074
Add storage function to purge history for a room
2016-07-04 16:02:50 +01:00
870c45913e
Use similar naming we use in email notifs for push
...
Fixes https://github.com/vector-im/vector-web/issues/1654
2016-06-24 11:41:11 +01:00
e4134c5e13
Merge pull request #841 from matrix-org/erikj/event_counter
...
Add metric counter for number of persisted events
2016-06-06 14:17:40 +01:00
7aa778fba9
Add metric counter for number of persisted events
2016-06-06 11:58:09 +01:00
70aee0717c
Add events to cache when we persist them
2016-06-06 11:34:53 +01:00
cffe46408f
Don't rely on options when inserting event into cache
2016-06-03 18:25:21 +01:00
10ea3f46ba
Change the way we cache events
2016-06-03 17:57:50 +01:00
05e01f21d7
Remove event fetching from DB threads
2016-06-03 17:22:13 +01:00
c8285564a3
Use state to calculate get_users_in_room
2016-06-01 15:25:25 +01:00
997db04648
Merge remote-tracking branch 'origin/develop' into dbkr/email_notifs
2016-05-10 14:40:19 +02:00
fcb2c3f0db
Remove unused import
2016-05-06 15:47:40 +01:00
fd85b167ec
Pull loop one level up
2016-05-06 15:38:42 +01:00
b6e0be701e
Queue events for persistence
2016-05-06 14:31:44 +01:00
acded821c4
Merge remote-tracking branch 'origin/develop' into dbkr/email_notifs
2016-04-29 10:05:20 +01:00
871357d539
Check that somethign has happend before running the selects
2016-04-27 11:54:13 +01:00
07d765209d
First bits of emailpusher
...
Mostly logic of when to send an email
2016-04-19 14:24:36 +01:00
762ada1e07
Add back backfilled parameter that was removed
2016-04-12 11:58:04 +01:00
cccf86dd05
Check if we've already backfilled events
2016-04-12 11:19:32 +01:00
9c99ab4572
Merge remote-tracking branch 'origin/develop' into dbkr/pushers_use_event_actions
2016-04-07 16:35:22 +01:00
0fd1cd2400
pep8
2016-04-06 16:50:47 +01:00
7e2c89a37f
Make pushers use the event_push_actions table instead of listening on an event stream & running the rules again. Sytest passes, but remaining to do:
...
* Make badges work again
* Remove old, unused code
2016-04-06 15:42:15 +01:00
75fb9ac1be
Add a slaved events store class
...
Add a test to check that get_room_names_and_aliases does the same
thing on both the master and on the slave data store.
2016-04-06 14:18:35 +01:00
3d76b7cb2b
Store invites in a separate table.
2016-04-04 16:30:15 +01:00
f2b916534b
Merge pull request #684 from matrix-org/markjh/backfill_id_gen
...
Use a stream id generator for backfilled ids
2016-04-01 15:13:14 +01:00
e36bfbab38
Use a stream id generator for backfilled ids
2016-04-01 13:29:05 +01:00
7753fc6570
Fix the invalidation of the names and aliases cache
2016-04-01 10:34:51 +01:00
5d06929169
Move the check for backfilled outside the for loop
2016-03-31 15:09:09 +01:00
76503f95ed
Remove the is_new_state argument to persist event.
...
Move the checks for whether an event is new state inside persist
event itself.
This was harder than expected because there wasn't enough information
passed to persist event to correctly handle invites from remote servers
for new rooms.
2016-03-31 15:00:42 +01:00
62e395f0e3
Merge pull request #676 from matrix-org/markjh/replicate_stateIII
...
Add replication streams for ex outliers and current state resets
2016-03-31 11:20:57 +01:00
2ec5426035
Use a namedtuple rather than tuple unpacking
2016-03-31 10:33:02 +01:00
1fbb094c6f
Add replication streams for ex outliers and current state resets
2016-03-30 17:19:56 +01:00
8b8052909f
return the state_group for backfill
2016-03-30 16:20:07 +01:00
61407986b4
Add a entry to current_state_resets table when the current state is reset
2016-03-30 16:18:46 +01:00
31a9eceda5
Add a replication stream for state groups
2016-03-30 16:01:58 +01:00
1e25f62ee6
Use a stream id generator to assign state group ids
2016-03-30 12:55:02 +01:00
e14baa7a3b
Merge pull request #665 from matrix-org/erikj/dont_cache_events
...
Only cache events in the event cache
2016-03-23 13:19:12 +00:00
8b0dfc9fc4
Don't cache events in get_current_state_for_key
2016-03-23 11:42:17 +00:00
99f929f36b
Make StateHandler._state_cache only store event_ids.
2016-03-22 16:06:04 +00:00
5244c0b48e
Remove unused backfilled parameter from persist_event
2016-03-21 18:06:08 +00:00
67ed8065db
Dedupe requested event list in _get_events
2016-03-18 14:31:31 +00:00
7076082ae6
Fix relative imports so they work in both py3 and py27
2016-03-08 11:45:50 +00:00
a612ce6659
Merge pull request #489 from matrix-org/markjh/replication
...
Add a /replication API for extracting the updates that happened on synapse.
2016-03-01 15:08:24 +00:00
60a0f81c7a
Add a /replication API for extracting the updates that happened on
...
synapse
This is necessary for replicating the data in synapse to be visible to a
separate service because presence and typing notifications aren't stored
in a database so won't be visible to another process.
This API can be used to either get the raw data by requesting the tables
themselves or to just receive notifications for updates by following the
streams meta-stream.
Returns updates for each table requested a JSON array of arrays with a
row for each row in the table.
Each table is prefixed by a header row with the: name of the table,
current stream_id position for the table, number of rows, number of
columns and the names of the columns.
This is followed by the rows that have been added to the server since
the requester last asked.
The API has a timeout and is hooked up to the notifier so that a slave
can long poll for updates.
2016-03-01 14:49:41 +00:00
54172924c8
Load the current id in the IdGenerator constructor
...
Rather than loading them lazily. This allows us to remove all
the yield statements and spurious arguments for the get_next
methods.
It also allows us to replace all instances of get_next_txn with
get_next since get_next no longer needs to access the db.
2016-03-01 14:32:56 +00:00
42109a62a4
Remove unused param from get_max_token
2016-02-18 16:37:28 +00:00
0214745239
Rename functions
2016-02-10 11:09:56 +00:00
7b0d846407
Atomically persit push actions when we persist the event
2016-02-09 16:19:15 +00:00
Erik Johnston