summary refs log tree commit diff
path: root/synapse/app/synchrotron.py (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Factor out common application startRichard van der Hoff2017-08-151-49/+20
| | | | | We have 10 copies of this code, and I don't really want to update each one separately.
* Make workers report to master for user ip updatesErik Johnston2017-06-271-2/+2
|
* Move get_presence_list_* to SlaveStoreErik Johnston2017-04-111-11/+1
|
* Move get_interested_partiesErik Johnston2017-04-111-3/+2
|
* Reduce federation presence replication trafficErik Johnston2017-04-101-4/+2
| | | | | | | | This is mainly done by moving the calculation of where to send presence updates from the presence handler to the transaction queue, so we only need to send the presence event (and not the destinations) across the replication connection. Before we were duplicating by sending the full state across once per destination.
* Merge pull request #2103 from matrix-org/erikj/no-double-encodeErik Johnston2017-04-071-3/+1
|\ | | | | Don't double encode replication data
| * Don't double json encode typing replication dataErik Johnston2017-04-051-3/+1
| |
* | CommentErik Johnston2017-04-061-1/+1
| |
* | DocsErik Johnston2017-04-061-0/+7
| |
* | Don't immediately notify the master about users whose syncs have gone awayErik Johnston2017-04-051-4/+36
|/
* Add the appropriate amount of preserve_fnErik Johnston2017-04-041-7/+6
|
* Shuffle and comment synchrotron presenceErik Johnston2017-04-041-9/+11
|
* Update all the workers and master to use TCP replicationErik Johnston2017-04-031-196/+110
|
* Merge pull request #2033 from matrix-org/erikj/repl_speedErik Johnston2017-03-211-4/+8
|\ | | | | Don't send the full event json over replication
| * PEP8Erik Johnston2017-03-201-1/+0
| |
| * Don't send the full event json over replicationErik Johnston2017-03-171-3/+8
| |
* | Run the reactor with the sentinel logcontextRichard van der Hoff2017-03-181-2/+7
|/ | | | | | This fixes a class of 'Unexpected logcontext' messages, which were happening because the logcontext was somewhat arbitrarily swapping between the sentinel and the `run` logcontext.
* Don't recreate so many setsErik Johnston2017-03-161-2/+1
|
* Refactor logger config for workersRichard van der Hoff2017-03-101-1/+1
| | | | - to make it easier to add more config options.
* Add db functions needed for room initial sync to slaveMark Haines2017-02-131-0/+4
|
* Remove debug loggingErik Johnston2017-01-301-1/+0
|
* Hook device list updates to replicationErik Johnston2017-01-271-1/+26
|
* Restore default bind addressErik Johnston2017-01-101-10/+2
|
* Fix check for bind_addressJohannes Löthberg2016-12-201-2/+2
| | | | | | | The empty string is a valid setting for the bind_address option, so explicitly check for None here instead. Signed-off-by: Johannes Löthberg <johannes@kyriasis.com>
* Add support for specifying multiple bind addressesJohannes Löthberg2016-12-181-20/+34
| | | | Signed-off-by: Johannes Löthberg <johannes@kyriasis.com>
* Respect use_frozen_dicts option in workersErik Johnston2016-11-081-0/+2
|
* Support /initialSync in synchrotron workerErik Johnston2016-09-211-0/+6
|
* Enable state caches on workersErik Johnston2016-09-151-0/+1
|
* CommentErik Johnston2016-09-091-0/+3
|
* Fix up the calls to the notifier for device messagesMark Haines2016-09-011-0/+3
|
* Add new direct message storage functions to slave storeErik Johnston2016-08-301-0/+2
|
* Remove broken cache stuffErik Johnston2016-08-151-5/+0
|
* Implement cache replication streamErik Johnston2016-08-151-13/+0
|
* Make synchrotron accept /eventsErik Johnston2016-08-121-2/+34
|
* Synced up synchrotron set_state with PresenceHandler set_stateWill Hunt2016-08-111-1/+1
|
* Add federation /version APIErik Johnston2016-08-051-1/+1
|
* Fix setting gc thresholds in the workersMark Haines2016-06-171-1/+2
|
* Use worker_ prefixes for worker config, use existing support for multiple ↵Mark Haines2016-06-161-17/+12
| | | | config files
* Access the event_cache_size directly from the server object.Mark Haines2016-06-161-3/+3
| | | | | This means that the workers can override the event_cache_size directly without clobbering the value in the main synapse config.
* Access replication_url from the worker config directlyMark Haines2016-06-161-2/+3
|
* Inline the synchrotron and pusher configs into the main configMark Haines2016-06-161-103/+34
|
* Add gc_threshold to pusher and synchrotronErik Johnston2016-06-071-0/+14
|
* Allow setting of gc.set_thresholdsErik Johnston2016-06-071-5/+10
|
* Merge pull request #845 from matrix-org/markjh/synchrotron_presenceMark Haines2016-06-061-1/+4
|\ | | | | Fix a KeyError in the synchrotron presence
| * Fix a KeyError in the synchrotron presenceMark Haines2016-06-061-1/+4
| |
* | Yield on the sleeps intended to backoff replicationMark Haines2016-06-061-1/+1
|/
* Fix spellingMark Haines2016-06-031-1/+1
|
* Add get_presence_list_accepted to the broken caches in synchrotronMark Haines2016-06-031-4/+9
|
* Clear the list of ongoing syncs on shutdownMark Haines2016-06-031-0/+9
|
* Appease flake8Mark Haines2016-06-031-1/+1
|
* Use ClientIpStore to record client ipsMark Haines2016-06-031-3/+4
|
* Send updates to the syncing users every ten seconds or immediately if ↵Mark Haines2016-06-031-10/+43
| | | | they've just come online
* Add a separate process that can handle /sync requestsMark Haines2016-06-031-0/+467