Commit Graph

98 Commits

Author SHA1 Message Date
David Robertson
9cd7610f86
Revert "Add event_stream_ordering column to membership state tables (#14979)"
This reverts commit 5fdc12f482.
2023-02-07 15:26:55 +00:00
Nick Mills-Barrett
5fdc12f482
Add event_stream_ordering column to membership state tables (#14979)
This adds an `event_stream_ordering` column to `current_state_events`,
`local_current_membership` and `room_memberships`. Each of these tables
is regularly joined with the `events` table to get the stream ordering
and denormalising this into each table will yield significant query
performance improvements once used. Includes a background job to
populate these values from the `events` table.

Same idea as https://github.com/matrix-org/synapse/pull/13703.

Signed off by Nick @ Beeper (@fizzadar).
2023-02-07 00:10:54 +00:00
Dirk Klimpel
8d5325ec0c
Drop unused table presence (#14825) 2023-01-13 14:17:03 +00:00
reivilibre
d6bda5addd
Add index to improve performance of the /timestamp_to_event endpoint used for jumping to a specific date in the timeline of a room. (#14799) 2023-01-11 12:29:13 +00:00
reivilibre
fb60cb16fe
Faster remote room joins: stream the un-partial-stating of events over replication. [rei:frrj/streams/unpsr] (#14545) 2022-12-14 14:47:11 +00:00
Patrick Cloke
24a97b3e71
Delete event_push_summary_unique_index again. (#14669)
if a Synapse deployment upgraded (from < 1.62.0 to >= 1.70.0) then it
is possible for schema deltas to run before background updates causing
drift in the database schema due to:

1. A delta registered a background update to create an index.
2. A delta dropped the above index if it exists (but it yet exist won't since
  the background job hasn't run).
3. The code assumed the index was dropped.

To fix this we:

1. Cancel the background update which could create the index.
2. Drop the index again.
3. Drop a related index which is dropped by the background update.
2022-12-14 09:25:33 -05:00
Patrick Cloke
c369e95691
Rebuild the user directory and stats tables. (#14643)
Due to the various fixes to the StreamChangeCache it is not
safe to trust the information in the user directory or room/user
stats tables. Rebuild them as background jobs.

In particular see da77720752 (#14639),
and 6a8310f3df (#14435).

Maybe also be related to fac8a38525
(#14592).
2022-12-08 11:40:20 -05:00
reivilibre
501f62d1a6
Faster remote room joins: stream the un-partial-stating of rooms over replication. [rei:frrj/streams/unpsr] (#14473) 2022-12-05 13:07:55 +00:00
Erik Johnston
f38d7d79c8
Add another index to device_lists_changes_in_room (#14534)
This helps avoid reading unnecessarily large amounts of data from the
table when querying with a set of room IDs.
2022-11-23 14:09:00 +00:00
Sean Quah
9cae44f49e
Track unconverted device list outbound pokes using a position instead (#14516)
When a local device list change is added to
`device_lists_changes_in_room`, the `converted_to_destinations` flag is
set to `FALSE` and the `_handle_new_device_update_async` background
process is started. This background process looks for unconverted rows
in `device_lists_changes_in_room`, copies them to
`device_lists_outbound_pokes` and updates the flag.

To update the `converted_to_destinations` flag, the database performs a
`DELETE` and `INSERT` internally, which fragments the table. To avoid
this, track unconverted rows using a `(stream ID, room ID)` position
instead of the flag.

From now on, the `converted_to_destinations` column indicates rows that
need converting to outbound pokes, but does not indicate whether the
conversion has already taken place.

Closes #14037.

Signed-off-by: Sean Quah <seanq@matrix.org>
2022-11-22 16:46:52 +00:00
Sean Quah
b2c2b03079
Fix PostgreSQL sometimes using table scans for event_search (#14409)
PostgreSQL may underestimate the number of distinct `room_id`s in
`event_search`, which can cause it to use table scans for queries for
multiple rooms.

Fix this by setting `n_distinct` on the column.

Resolves #14402.

Signed-off-by: Sean Quah <seanq@matrix.org>
2022-11-10 19:02:27 +00:00
Quentin Gliech
8756d5c87e
Save login tokens in database (#13844)
* Save login tokens in database

Signed-off-by: Quentin Gliech <quenting@element.io>

* Add upgrade notes

* Track login token reuse in a Prometheus metric

Signed-off-by: Quentin Gliech <quenting@element.io>
2022-10-26 11:45:41 +01:00
James Salter
d902181de9
Unified search query syntax using the full-text search capabilities of the underlying DB. (#11635)
Support a unified search query syntax which leverages more of the full-text
search of each database supported by Synapse.

Supports, with the same syntax across Postgresql 11+ and Sqlite:

- quoted "search terms"
- `AND`, `OR`, `-` (negation) operators
- Matching words based on their stem, e.g. searches for "dog" matches
  documents containing "dogs". 

This is achieved by 

- If on postgresql 11+, pass the user input to `websearch_to_tsquery`
- If on sqlite, manually parse the query and transform it into the sqlite-specific
  query syntax.

Note that postgresql 10, which is close to end-of-life, falls back to using
`phraseto_tsquery`, which only supports a subset of the features.

Multiple terms separated by a space are implicitly ANDed.

Note that:

1. There is no escaping of full-text syntax that might be supported by the database;
  e.g. `NOT`, `NEAR`, `*` in sqlite. This runs the risk that people might discover this
  as accidental functionality and depend on something we don't guarantee.
2. English text is assumed for stemming. To support other languages, either the target
  language needs to be known at the time of indexing the message (via room metadata,
  or otherwise), or a separate index for each language supported could be created.

Sqlite docs: https://www.sqlite.org/fts3.html#full_text_index_queries
Postgres docs: https://www.postgresql.org/docs/11/textsearch-controls.html
2022-10-25 14:05:22 -04:00
Patrick Cloke
dbf18f514e
Update the thread_id right before use (in case the bg update hasn't finished) (#14222)
This avoids running a forced-update of a null thread_id rows.

An index is added (in the background) to hopefully make this
easier in the future.
2022-10-18 14:55:41 +00:00
David Robertson
c3a4780080
When restarting a partial join resync, prioritise the server which actioned a partial join (#14126) 2022-10-18 12:33:18 +01:00
Patrick Cloke
3bbe532abb
Add an API for listing threads in a room. (#13394)
Implement the /threads endpoint from MSC3856.

This is currently unstable and behind an experimental configuration
flag.

It includes a background update to backfill data, results from
the /threads endpoint will be partial until that finishes.
2022-10-13 08:02:11 -04:00
Patrick Cloke
a7ba457b2b
Mark events as read using threaded read receipts from MSC3771. (#13877)
Applies the proper logic for unthreaded and threaded receipts to either
apply to all events in the room or only events in the same thread, respectively.
2022-10-04 10:46:42 -04:00
Patrick Cloke
b4ec4f5e71
Track notification counts per thread (implement MSC3773). (#13776)
When retrieving counts of notifications segment the results based on the
thread ID, but choose whether to return them as individual threads or as
a single summed field by letting the client opt-in via a sync flag.

The summarization code is also updated to be per thread, instead of per
room.
2022-10-04 09:47:04 -04:00
Erik Johnston
5a6d025246
Clear out old rows from event_push_actions_staging (#14020)
On matrix.org we have ~5 million stale rows in `event_push_actions_staging`, let's add a background job to make sure we clear them out.
2022-10-03 18:44:44 +01:00
Brendan Abolivier
be76cd8200
Allow admins to require a manual approval process before new accounts can be used (using MSC3866) (#13556) 2022-09-29 15:23:24 +02:00
Erik Johnston
4b17a5ace8
Handle remote device list updates during partial join (#13913)
c.f. #12993 (comment), point 3

This stores all device list updates that we receive while partial joins are ongoing, and processes them once we have the full state.

Note: We don't actually process the device lists in the same ways as if we weren't partially joined. Instead of updating the device list remote cache, we simply notify local users that a change in the remote user's devices has happened. I think this is safe as if the local user requests the keys for the remote user and we don't have them we'll simply fetch them as normal.
2022-09-28 13:42:43 +00:00
David Robertson
f5aaa55e27
Add new columns tracking when we partial-joined (#13892) 2022-09-27 17:26:35 +01:00
Brendan Abolivier
ccca14140a
Track device IDs for pushers (#13831)
Second half of the MSC3881 implementation
2022-09-21 15:31:53 +00:00
Brendan Abolivier
8ae42ab8fa
Support enabling/disabling pushers (from MSC3881) (#13799)
Partial implementation of MSC3881
2022-09-21 14:39:01 +00:00
Eric Eastwood
44be42338e
Add support to purge rows from MSC2716 and other tables when purging a room (#13825)
`event_failed_pull_attempts` added in https://github.com/matrix-org/synapse/pull/13589

MSC2716 related tables added in:

 - https://github.com/matrix-org/synapse/pull/10245/files#diff-3d42dfb44d02f7de3aada105e0bdc1cc9dd7f953cbf0f36c5d0f50827bf0320aR1
    - Renamed in https://github.com/matrix-org/synapse/pull/10838/files#diff-2730bfbe9e688b55e46f9371aefe67dac2bd2b2b7d9d6b92774eea1fcfae156dR1
 - https://github.com/matrix-org/synapse/pull/10498/files#diff-c52bbfbb5921a3f6f023b24343668479d966fac164f13b7c39d2197ce3afa7a5R1
2022-09-16 10:56:56 -05:00
Eric Eastwood
957e3d74fc
Keep track when we try and fail to process a pulled event (#13589)
We can follow-up this PR with:

 1. Only try to backfill from an event if we haven't tried recently -> https://github.com/matrix-org/synapse/issues/13622
 1. When we decide to backfill that event again, process it in the background so it doesn't block and make `/messages` slow when we know it will probably fail again -> https://github.com/matrix-org/synapse/issues/13623
 1. Generally track failures everywhere we try and fail to pull an event over federation -> https://github.com/matrix-org/synapse/issues/13700

Fix https://github.com/matrix-org/synapse/issues/13621

Part of https://github.com/matrix-org/synapse/issues/13356

Mentioned in [internal doc](https://docs.google.com/document/d/1lvUoVfYUiy6UaHB6Rb4HicjaJAU40-APue9Q4vzuW3c/edit#bookmark=id.qv7cj51sv9i5)
2022-09-14 13:57:50 -05:00
Patrick Cloke
666ae87729
Update event push action and receipt tables to support threads. (#13753)
Adds a `thread_id` column to the `event_push_actions`, `event_push_actions_staging`,
and `event_push_summary` tables. This will notifications to be segmented by the thread
in a future pull request. The `thread_id` column stores the root event ID or the special
value `"main"`.

The `thread_id` column for `event_push_actions` and `event_push_summary` is
backfilled with `"main"` for all existing rows. New entries into `event_push_actions`
and `event_push_actions_staging` will get the proper thread ID.

`receipts_linearized` and `receipts_graph` also gain a `thread_id` column, which is similar,
except `NULL` is a special value meaning the receipt is "unthreaded".

See MSC3771 and MSC3773 for where this data will be useful.
2022-09-14 17:11:16 +00:00
Patrick Cloke
f2d12ccabe
Use partial indices on SQLIte. (#13802)
Partial indices have been supported since SQLite 3.8, but Synapse
now requires >= 3.27, so we can enable support for them.

This requires rebuilding previous indices which were partial on
PostgreSQL, but not on SQLite.
2022-09-14 12:01:42 -04:00
Mathieu Velten
12dacecabd
Make sequence cache_invalidation_stream_seq begin at 2 (#13766)
Signed-off-by: Mathieu Velten <mathieuv@matrix.org>
Co-authored-by: Sean Quah <8349537+squahtx@users.noreply.github.com>
2022-09-13 16:14:28 +02:00
Nick Mills-Barrett
cdbb641232
Add receipts event stream ordering (#13703) 2022-09-13 08:16:37 +01:00
Nick Mills-Barrett
da41a7cd61
Remove check current state membership up to date (#13745)
* Remove checks for membership column in current_state_events
* Add schema script to force through the
  `current_state_events_membership` background job

Contributed by Nick @ Beeper (@fizzadar).
2022-09-12 12:58:33 +01:00
Dirk Klimpel
f799eac7ea
Add timestamp to user's consent (#13741)
Co-authored-by: reivilibre <olivier@librepush.net>
2022-09-08 15:41:48 +00:00
Patrick Cloke
48a5c47a9f
Add a schema delta to drop unstable private read receipts. (#13692)
Otherwise they'll be leaked due to the filtering code only respecting
the stable identifiers for private read receipts.
2022-09-01 14:57:47 -04:00
Shay
20c76cecb9
Drop unused column application_services_state.last_txn (#13627) 2022-08-30 10:29:16 -07:00
Richard van der Hoff
b116d3ce00
Bg update to populate new events table columns (#13215)
These columns were added back in Synapse 1.52, and have been populated for new
events since then. It's now (beyond) time to back-populate them for existing
events.
2022-07-15 12:47:26 +01:00
Patrick Cloke
4db7862e0f
Drop unused tables from groups/communities. (#12967)
These tables have been unused since Synapse v1.61.0, although schema version 72
was added in Synapse v1.62.0.
2022-07-13 09:55:14 -04:00
Richard van der Hoff
a366b75b72
Drop unused table event_reference_hashes (#13218)
This is unused since Synapse 1.60.0 (#12679). It's time for it to go.
2022-07-12 18:52:06 +00:00
Patrick Cloke
b0366853ca Merge remote-tracking branch 'origin/release-v1.62' into develop 2022-06-30 13:27:24 -04:00
Erik Johnston
a3a05c812d
Add index to help delete old push actions (#13141) 2022-06-30 14:05:49 +00:00
Šimon Brandner
13e359aec8
Implement MSC3827: Filtering of /publicRooms by room type (#13031)
Signed-off-by: Šimon Brandner <simon.bra.ag@gmail.com>
2022-06-29 17:12:45 +00:00
Erik Johnston
7469824d58
Fix serialization errors when rotating notifications (#13118) 2022-06-28 13:13:44 +01:00
Erik Johnston
0d1d3e0708
Speed up get_unread_event_push_actions_by_room (#13005)
Fixes #11887 hopefully.

The core change here is that `event_push_summary` now holds a summary of counts up until a much more recent point, meaning that the range of rows we need to count in `event_push_actions` is much smaller.

This needs two major changes:
1. When we get a receipt we need to recalculate `event_push_summary` rather than just delete it
2. The logic for deleting `event_push_actions` is now divorced from calculating `event_push_summary`.

In future it would be good to calculate `event_push_summary` while we persist a new event (it should just be a case of adding one to the relevant rows in `event_push_summary`), as that will further simplify the get counts logic and remove the need for us to periodically update `event_push_summary` in a background job.
2022-06-15 15:17:14 +00:00
Richard van der Hoff
75fb10ee45
Clean up schema for event_edges (#12893)
* Remove redundant references to `event_edges.room_id`

We don't need to care about the room_id here, because we are already checking
the event id.

* Clean up the event_edges table

We make a number of changes to `event_edges`:

 * We give the `room_id` and `is_state` columns defaults (null and false
   respectively) so that we can stop populating them.
 * We drop any rows that have `is_state` set true - they should no longer
   exist.
 * We drop any rows that do not exist in `events` - these should not exist
   either.
 * We drop the old unique constraint on all the colums, which wasn't much use.
 * We create a new unique index on `(event_id, prev_event_id)`.
 * We add a foreign key constraint to `events`.

These happen rather differently depending on whether we are on Postgres or
SQLite. For SQLite, we just rebuild the whole table, copying only the rows we
want to keep. For Postgres, we try to do things in the background as much as
possible.

* Stop populating `event_edges.room_id` and `is_state`

We can just rely on the defaults.
2022-06-15 12:29:42 +01:00
Patrick Cloke
5f4ecf759d
Rename delta to apply in the proper schema version. (#13050) 2022-06-14 14:34:04 +00:00
Patrick Cloke
53b77b203a
Replace noop background updates with DELETE. (#12954)
Removes the `register_noop_background_update` and deletes the background
updates directly in a delta file.
2022-06-13 14:06:27 -04:00
Mathieu Velten
7f92ac4c1c
Add a migration step to cleanup potential leftovers of bug 11833 (#12784)
Co-authored-by: Sean Quah <8349537+squahtx@users.noreply.github.com>
2022-05-30 16:51:37 +02:00
Erik Johnston
32ef24fbd7
Add index to cache invalidations (#12747)
For workers that rarely write to the cache the `get_all_updated_caches`
query can become expensive if the worker falls behind when reading the
cache.
2022-05-17 09:34:59 +00:00
Erik Johnston
5c9e39e619
Track device list updates per room. (#12321)
This is a first step in dealing with #7721.

The idea is basically that rather than calculating the full set of users a device list update needs to be sent to up front, we instead simply record the rooms the user was in at the time of the change. This will allow a few things:

1. we can defer calculating the set of remote servers that need to be poked about the change; and
2. during `/sync` and `/keys/changes` we can avoid also avoid calculating users who share rooms with other users, and instead just look at the rooms that have changed.

However, care needs to be taken to correctly handle server downgrades. As such this PR writes to both `device_lists_changes_in_room` and the `device_lists_outbound_pokes` table synchronously. In a future release we can then bump the database schema compat version to `69` and then we can assume that the new `device_lists_changes_in_room` exists and is handled.

There is a temporary option to disable writing to `device_lists_outbound_pokes` synchronously, allowing us to test the new code path does work (and by implication upgrading to a future release and downgrading to this one will work correctly).

Note: Ideally we'd do the calculation of room to servers on a worker (e.g. the background worker), but currently only master can write to the `device_list_outbound_pokes` table.
2022-04-04 15:25:20 +01:00
Nick Mills-Barrett
993d90f82b
Use a sequence to generate AS transaction IDs, drop last_txn AS state (#12209)
Switching to a sequence means there's no need to track `last_txn` on the
AS state table to generate new TXN IDs. This also means that there is
no longer contention between the AS scheduler and AS handler on updates
to the `application_services_state` table, which will prevent serialization
errors during the complete AS txn transaction.
2022-04-01 13:33:25 +00:00
Andrew Morgan
d8d0271977
Send device list updates to application services (MSC3202) - part 1 (#11881)
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
2022-03-30 14:39:27 +01:00