This updates the device list updater so that it has a context
per-request, rather than a global 30 seconds for the entire server. This
could mean that talking to a slow remote server or requesting a lot of
user IDs was pretty much guaranteed to fail.
It also uses the process context to allow correct cancellation when
Dendrite wants to shut down cleanly.
`If a device list update goes missing, the server resyncs on the next
one` was failing because a previous test would receive a `waitTime` of
1h, resulting in the test timing out.
This now tries to handle the returned errors differently, e.g. by using
the default `waitTime` of 2s. Also doesn't try further users in the
list, if one of the errors would cause a longer `waitTime`.
This ensures that if the device list updater is already backing off a node, we don't try to call processServer again anyway for server just because the server name arrived in the channel. Otherwise we can keep trying to hit a remote server that is offline or not behaving every second and that spams the logs too.
* Generic-based internal HTTP API (tested out on a few endpoints in the federation API)
* Add `PerformInvite`
* More tweaks
* Fix metric name
* Fix LookupStateIDs
* Lots of changes to clients
* Some serverside stuff
* Some error handling
* Use paths as metric names
* Revert "Use paths as metric names"
This reverts commit a9323a6a343f5ce6461a2e5bd570fe06465f1b15.
* Namespace metric names
* Remove duplicate entry
* Remove another duplicate entry
* Tweak error handling
* Some more tweaks
* Update error behaviour
* Some more error tweaking
* Fix API path for `PerformDeleteKeys`
* Fix another path
* Tweak federation client proxying
* Fix another path
* Don't return typed nils
* Some more tweaks, not that it makes any difference
* Tweak federation client proxying
* Maybe fix the key backup test
* Fix query issue, only add "changed" users if we actually share a room
* Avoid log spam if context is done
* Undo changes to filterSharedUsers
* Add logging again..
* Fix SQLite shared users query
* Change query to include invited users
* Fix flakey sytest 'Local device key changes get to remote servers'
* Debug logs
* Remove internal/test and use /test only
Remove a lot of ancient code too.
* Use FederationRoomserverAPI in more places
* Use more interfaces in federationapi; begin adding regression test
* Linting
* Add regression test
* Unbreak tests
* ALL THE LOGS
* Fix a race condition which could cause events to not be sent to servers
If a new room event which rewrites state arrives, we remove all joined hosts
then re-calculate them. This wasn't done in a transaction so for a brief period
we would have no joined hosts. During this interim, key change events which arrive
would not be sent to destination servers. This would sporadically fail on sytest.
* Unbreak new tests
* Linting
* tidy up interfaces
* remove unused GetCreatorIDForAlias
* Add RoomserverUserAPI interface
* Define more interfaces
* Use AppServiceInternalAPI for consistent naming
* clean up federationapi constructor a bit
* Fix monolith in -http mode
* Find the complete key ID when uploading signatures
* Try that again
* Try splitting the right thing
* Don't do it for device keys
* Refactor `QuerySignatures`
* Revert "Refactor `QuerySignatures`"
This reverts commit c02832a3e92569f64f180dec1555056dc8f8c3e3.
* Both requested key IDs and master/self/user keys
* Fix uniqueness
* Try tweaking GMSL
* Update GMSL again
* Revert "Update GMSL again"
This reverts commit bd6916cc379dd8d9e3f38d979c6550bd658938aa.
* Revert "Try tweaking GMSL"
This reverts commit 2a054524da9d64c6a2a5228262fbba5fde28798c.
* Database migrations
* Fix retrieving cross-signing signatures in `/user/devices/{userId}`
We need to know the target device IDs in order to get the signatures and we weren't populating those.
* Fix up signature retrieval
* Fix SQLite
* Always include the target's own signatures as well as the requesting user
* Move receipt sending to own JetStream producer
* Move SendToDevice to producer
* Remove most parts of the EDU server
* Fix SendToDevice & copyrights
* Move structs, cleanup EDU Server traces
* Use HeadersOnly subscription
* Missing file
* Fix linter issues
* Move consumers to own files
* Rename durable consumer; Consumer cleanup
* Docs/config cleanup
* Add unit test for device list update debouncing
* bugfix: actually return stale device lists in the test...
Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
* Send device_list update to satisfy sytest
* Fix build issue from merged in change
Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
* Merge user API databases into one
* Remove DeviceDatabase from config
* Fix tests
* Try that again
* Clean up keyserver device keys when the devices no longer exist in the user API
* Tweak ordering
* Fix UserExists flag, device check
* Allow including empty entries so we can clean them up
* Remove logging
* Remove unneeded logging
* Add MasterKey & SelfSigningKey to update
Avoid panic if signatures are not present
* Add passing test
* Revert "Add MasterKey & SelfSigningKey to update"
This reverts commit 2c81b34884be8b5b875a33420c0f985b578d3fb8.
* Send MasterKey & SelfSigningKey with update
* Debugging
* Remove delete() so we also query signingkeys
* Remodel how device list change IDs are created
Previously we made them using the offset Kafka supplied.
We don't run Kafka anymore, so now we make the SQL table assign
the change ID via an AUTOINCREMENTing ID. Redesign the
`keyserver_key_changes` table to have `UNIQUE(user_id)` so we
don't accumulate key changes forevermore, we now have at most 1
row per user which contains the highest change ID.
This needs a SQL migration.
* Ensure we bump the change ID on sqlite
* Actually read the DeviceChangeID not the Offset in synapi
* Add SQL migrations
* Prepare after migration; fixup dendrite-upgrade-test logging
* Use higher version numbers; fix sqlite query to increment better
* Default 0 on postgres
* fixup postgres migration on fresh dendrite instances
