Fix outlier re-persisting causing problems with sliding sync tables
Follow-up to https://github.com/element-hq/synapse/pull/17512
When running on `matrix.org`, we discovered that a remote invite is
first persisted as an `outlier` and then re-persisted again where it is
de-outliered. The first the time, the `outlier` is persisted with one
`stream_ordering` but when persisted again and de-outliered, it is
assigned a different `stream_ordering` that won't end up being used.
Since we call `_calculate_sliding_sync_table_changes()` before
`_update_outliers_txn()` which fixes this discrepancy (always use the
`stream_ordering` from the first time it was persisted), we're working
with an unreliable `stream_ordering` value that will possibly be unused
and not make it into the `events` table.
Pre-populate room data for quick filtering/sorting in the Sliding Sync
API
Spawning from
https://github.com/element-hq/synapse/pull/17450#discussion_r1697335578
This PR is acting as the Synapse version `N+1` step in the gradual
migration being tracked by
https://github.com/element-hq/synapse/issues/17623
Adding two new database tables:
- `sliding_sync_joined_rooms`: A table for storing room meta data that
the local server is still participating in. The info here can be shared
across all `Membership.JOIN`. Keyed on `(room_id)` and updated when the
relevant room current state changes or a new event is sent in the room.
- `sliding_sync_membership_snapshots`: A table for storing a snapshot of
room meta data at the time of the local user's membership. Keyed on
`(room_id, user_id)` and only updated when a user's membership in a room
changes.
Also adds background updates to populate these tables with all of the
existing data.
We want to have the guarantee that if a row exists in the sliding sync
tables, we are able to rely on it (accurate data). And if a row doesn't
exist, we use a fallback to get the same info until the background
updates fill in the rows or a new event comes in triggering it to be
fully inserted. This means we need a couple extra things in place until
we bump `SCHEMA_COMPAT_VERSION` and run the foreground update in the
`N+2` part of the gradual migration. For context on why we can't rely on
the tables without these things see [1].
1. On start-up, block until we clear out any rows for the rooms that
have had events since the max-`stream_ordering` of the
`sliding_sync_joined_rooms` table (compare to max-`stream_ordering` of
the `events` table). For `sliding_sync_membership_snapshots`, we can
compare to the max-`stream_ordering` of `local_current_membership`
- This accounts for when someone downgrades their Synapse version and
then upgrades it again. This will ensure that we don't have any
stale/out-of-date data in the
`sliding_sync_joined_rooms`/`sliding_sync_membership_snapshots` tables
since any new events sent in rooms would have also needed to be written
to the sliding sync tables. For example a new event needs to bump
`event_stream_ordering` in `sliding_sync_joined_rooms` table or some
state in the room changing (like the room name). Or another example of
someone's membership changing in a room affecting
`sliding_sync_membership_snapshots`.
1. Add another background update that will catch-up with any rows that
were just deleted from the sliding sync tables (based on the activity in
the `events`/`local_current_membership`). The rooms that need
recalculating are added to the
`sliding_sync_joined_rooms_to_recalculate` table.
1. Making sure rows are fully inserted. Instead of partially inserting,
we need to check if the row already exists and fully insert all data if
not.
All of this extra functionality can be removed once the
`SCHEMA_COMPAT_VERSION` is bumped with support for the new sliding sync
tables so people can no longer downgrade (the `N+2` part of the gradual
migration).
<details>
<summary><sup>[1]</sup></summary>
For `sliding_sync_joined_rooms`, since we partially insert rows as state
comes in, we can't rely on the existence of the row for a given
`room_id`. We can't even rely on looking at whether the background
update has finished. There could still be partial rows from when someone
reverted their Synapse version after the background update finished, had
some state changes (or new rooms), then upgraded again and more state
changes happen leaving a partial row.
For `sliding_sync_membership_snapshots`, we insert items as a whole
except for the `forgotten` column ~~so we can rely on rows existing and
just need to always use a fallback for the `forgotten` data. We can't
use the `forgotten` column in the table for the same reasons above about
`sliding_sync_joined_rooms`.~~ We could have an out-of-date membership
from when someone reverted their Synapse version. (same problems as
outlined for `sliding_sync_joined_rooms` above)
Discussed in an [internal
meeting](https://docs.google.com/document/d/1MnuvPkaCkT_wviSQZ6YKBjiWciCBFMd-7hxyCO-OCbQ/edit#bookmark=id.dz5x6ef4mxz7)
</details>
### TODO
- [x] Update `stream_ordering`/`bump_stamp`
- [x] Handle remote invites
- [x] Handle state resets
- [x] Consider adding `sender` so we can filter `LEAVE` memberships and
distinguish from kicks.
- [x] We should add it to be able to tell leaves from kicks
- [x] Consider adding `tombstone` state to help address
https://github.com/element-hq/synapse/issues/17540
- [x] We should add it `tombstone_successor_room_id`
- [x] Consider adding `forgotten` status to avoid extra
lookup/table-join on `room_memberships`
- [x] We should add it
- [x] Background update to fill in values for all joined rooms and
non-join membership
- [x] Clean-up tables when room is deleted
- [ ] Make sure tables are useful to our use case
- First explored in
https://github.com/element-hq/synapse/compare/erikj/ss_use_new_tables
- Also explored in
76b5a576eb
- [x] Plan for how can we use this with a fallback
- See plan discussed above in main area of the issue description
- Discussed in an [internal
meeting](https://docs.google.com/document/d/1MnuvPkaCkT_wviSQZ6YKBjiWciCBFMd-7hxyCO-OCbQ/edit#bookmark=id.dz5x6ef4mxz7)
- [x] Plan for how we can rely on this new table without a fallback
- Synapse version `N+1`: (this PR) Bump `SCHEMA_VERSION` to `87`. Add
new tables and background update to backfill all rows. Since this is a
new table, we don't have to add any `NOT VALID` constraints and validate
them when the background update completes. Read from new tables with a
fallback in cases where the rows aren't filled in yet.
- Synapse version `N+2`: Bump `SCHEMA_VERSION` to `88` and bump
`SCHEMA_COMPAT_VERSION` to `87` because we don't want people to
downgrade and miss writes while they are on an older version. Add a
foreground update to finish off the backfill so we can read from new
tables without the fallback. Application code can now rely on the new
tables being populated.
- Discussed in an [internal
meeting](https://docs.google.com/document/d/1MnuvPkaCkT_wviSQZ6YKBjiWciCBFMd-7hxyCO-OCbQ/edit#bookmark=id.hh7shg4cxdhj)
### Dev notes
```
SYNAPSE_TEST_LOG_LEVEL=INFO poetry run trial tests.storage.test_events.SlidingSyncPrePopulatedTablesTestCase
SYNAPSE_POSTGRES=1 SYNAPSE_POSTGRES_USER=postgres SYNAPSE_TEST_LOG_LEVEL=INFO poetry run trial tests.storage.test_events.SlidingSyncPrePopulatedTablesTestCase
```
```
SYNAPSE_TEST_LOG_LEVEL=INFO poetry run trial tests.handlers.test_sliding_sync.FilterRoomsTestCase
```
Reference:
- [Development docs on background updates and worked examples of gradual
migrations
](1dfa59b238/docs/development/database_schema.md (background-updates))
- A real example of a gradual migration:
https://github.com/matrix-org/synapse/pull/15649#discussion_r1213779514
- Adding `rooms.creator` field that needed a background update to
backfill data, https://github.com/matrix-org/synapse/pull/10697
- Adding `rooms.room_version` that needed a background update to
backfill data, https://github.com/matrix-org/synapse/pull/6729
- Adding `room_stats_state.room_type` that needed a background update to
backfill data, https://github.com/matrix-org/synapse/pull/13031
- Tables from MSC2716: `insertion_events`, `insertion_event_edges`,
`insertion_event_extremities`, `batch_events`
- `current_state_events` updated in
`synapse/storage/databases/main/events.py`
---
```
persist_event (adds to queue)
_persist_event_batch
_persist_events_and_state_updates (assigns `stream_ordering` to events)
_persist_events_txn
_store_event_txn
_update_metadata_tables_txn
_store_room_members_txn
_update_current_state_txn
```
---
> Concatenated Indexes [...] (also known as multi-column, composite or
combined index)
>
> [...] key consists of multiple columns.
>
> We can take advantage of the fact that the first index column is
always usable for searching
>
> *--
https://use-the-index-luke.com/sql/where-clause/the-equals-operator/concatenated-keys*
---
Dealing with `portdb` (`synapse/_scripts/synapse_port_db.py`),
https://github.com/element-hq/synapse/pull/17512#discussion_r1725998219
---
<details>
<summary>SQL queries:</summary>
Both of these are equivalent and work in SQLite and Postgres
Options 1:
```sql
WITH data_table (room_id, user_id, membership_event_id, membership, event_stream_ordering, {", ".join(insert_keys)}) AS (
VALUES (
?, ?, ?,
(SELECT membership FROM room_memberships WHERE event_id = ?),
(SELECT stream_ordering FROM events WHERE event_id = ?),
{", ".join("?" for _ in insert_values)}
)
)
INSERT INTO sliding_sync_non_join_memberships
(room_id, user_id, membership_event_id, membership, event_stream_ordering, {", ".join(insert_keys)})
SELECT * FROM data_table
WHERE membership != ?
ON CONFLICT (room_id, user_id)
DO UPDATE SET
membership_event_id = EXCLUDED.membership_event_id,
membership = EXCLUDED.membership,
event_stream_ordering = EXCLUDED.event_stream_ordering,
{", ".join(f"{key} = EXCLUDED.{key}" for key in insert_keys)}
```
Option 2:
```sql
INSERT INTO sliding_sync_non_join_memberships
(room_id, user_id, membership_event_id, membership, event_stream_ordering, {", ".join(insert_keys)})
SELECT
column1 as room_id,
column2 as user_id,
column3 as membership_event_id,
column4 as membership,
column5 as event_stream_ordering,
{", ".join("column" + str(i) for i in range(6, 6 + len(insert_keys)))}
FROM (
VALUES (
?, ?, ?,
(SELECT membership FROM room_memberships WHERE event_id = ?),
(SELECT stream_ordering FROM events WHERE event_id = ?),
{", ".join("?" for _ in insert_values)}
)
) as v
WHERE membership != ?
ON CONFLICT (room_id, user_id)
DO UPDATE SET
membership_event_id = EXCLUDED.membership_event_id,
membership = EXCLUDED.membership,
event_stream_ordering = EXCLUDED.event_stream_ordering,
{", ".join(f"{key} = EXCLUDED.{key}" for key in insert_keys)}
```
If we don't need the `membership` condition, we could use:
```sql
INSERT INTO sliding_sync_non_join_memberships
(room_id, membership_event_id, user_id, membership, event_stream_ordering, {", ".join(insert_keys)})
VALUES (
?, ?, ?,
(SELECT membership FROM room_memberships WHERE event_id = ?),
(SELECT stream_ordering FROM events WHERE event_id = ?),
{", ".join("?" for _ in insert_values)}
)
ON CONFLICT (room_id, user_id)
DO UPDATE SET
membership_event_id = EXCLUDED.membership_event_id,
membership = EXCLUDED.membership,
event_stream_ordering = EXCLUDED.event_stream_ordering,
{", ".join(f"{key} = EXCLUDED.{key}" for key in insert_keys)}
```
</details>
### Pull Request Checklist
<!-- Please read
https://element-hq.github.io/synapse/latest/development/contributing_guide.html
before submitting your pull request -->
* [x] Pull request is based on the develop branch
* [x] Pull request includes a [changelog
file](https://element-hq.github.io/synapse/latest/development/contributing_guide.html#changelog).
The entry should:
- Be a short description of your change which makes sense to users.
"Fixed a bug that prevented receiving messages from other servers."
instead of "Moved X method from `EventStore` to `EventWorkerStore`.".
- Use markdown where necessary, mostly for `code blocks`.
- End with either a period (.) or an exclamation mark (!).
- Start with a capital letter.
- Feel free to credit yourself, by adding a sentence "Contributed by
@github_username." or "Contributed by [Your Name]." to the end of the
entry.
* [x] [Code
style](https://element-hq.github.io/synapse/latest/code_style.html) is
correct
(run the
[linters](https://element-hq.github.io/synapse/latest/development/contributing_guide.html#run-the-linters))
---------
Co-authored-by: Erik Johnston <erik@matrix.org>
`bump_stamp` corresponds to the `stream_ordering` of the latest `DEFAULT_BUMP_EVENT_TYPES` in the room. This helps clients sort more readily without them needing to pull in a bunch of the timeline to determine the last activity. `bump_event_types` is a thing because for example, we don't want display name changes to mark the room as unread and bump it to the top. For encrypted rooms, we just have to consider any activity as a bump because we can't see the content and the client has to figure it out for themselves.
Outside of Synapse, `bump_stamp` is just a free-form counter so other implementations could use `received_ts`or `origin_server_ts` (see the [*Security considerations* section in MSC3575 about the potential pitfalls of using `origin_server_ts`](https://github.com/matrix-org/matrix-spec-proposals/blob/kegan/sync-v3/proposals/3575-sync.md#security-considerations)). It doesn't have any guarantee about always going up. In the Synapse case, it could go down if an event was redacted/removed (or purged in cases of retention policies).
In the future, we could add `bump_event_types` as [MSC3575](https://github.com/matrix-org/matrix-spec-proposals/pull/3575) mentions if people need to customize the event types.
---
In the Sliding Sync proxy, a similar [`timestamp` field was added](https://github.com/matrix-org/sliding-sync/pull/247) for the same purpose but the name is not obvious what it pertains to or what it's for.
The `timestamp` field was also added to Ruma in https://github.com/ruma/ruma/pull/1622
This is #17291 (which got reverted), with some added fixups, and change
so that tests actually pick up the error.
The problem was that we were not calculating any new chain IDs due to a
missing `not` in a condition.
Reduce the replication traffic of device lists, by not sending every
destination that needs to be sent the device list update over
replication. Instead a "hosts to send to have been calculated"
notification over replication, and then federation senders read the
destinations from the DB.
For non federation senders this should heavily reduce the impact of a
user in many large rooms changing a device.
This reverts commit bdf82efea5 (#17291)
This seems to have stopped persisting auth chains for new events, and so
is causing state res to fall back to the slow methods
We calculate the auth chain links outside of the main persist event
transaction to ensure that we do not block other event sending during
the calculation.
Sort is no longer configurable and we always sort rooms by the `stream_ordering` of the last event in the room or the point where the user can see up to in cases of leave/ban/invite/knock.
Add `event.internal_metadata.instance_name` (the worker instance that persisted the event) to go alongside the existing `event.internal_metadata.stream_ordering`.
`instance_name` is useful to properly compare and query for events with a token since you need to compare both the `stream_ordering` and `instance_name` against the vector clock/`instance_map` in the `RoomStreamToken`.
This is pre-requisite work and may be used in https://github.com/element-hq/synapse/pull/17293
Adding `event.internal_metadata.instance_name` was first mentioned in the initial Sliding Sync PR while pairing with @erikjohnston, see 09609cb0db (diff-5cd773fb307aa754bd3948871ba118b1ef0303f4d72d42a2d21e38242bf4e096R405-R410)
PR where this was introduced: https://github.com/matrix-org/synapse/pull/14817
### What does this affect?
`get_last_event_in_room_before_stream_ordering(...)` is used in Sync v2 in a lot of different state calculations.
`get_last_event_in_room_before_stream_ordering(...)` is also used in `/rooms/{roomId}/members`
There is a problem with `StreamIdGenerator` where it can go backwards
over restarts when a stream ID is requested but then not inserted into
the DB. This is problematic if we want to land #17215, and is generally
a potential cause for all sorts of nastiness.
Instead of trying to fix `StreamIdGenerator`, we may as well move to
`MultiWriterIdGenerator` that does not suffer from this problem (the
latest positions are stored in `stream_positions` table). This involves
adding SQLite support to the class.
This only changes id generators that were already using
`MultiWriterIdGenerator` under postgres, a separate PR will move the
rest of the uses of `StreamIdGenerator` over.
Weakness in auth chain indexing allows DoS from remote room members
through disk fill and high CPU usage.
A remote Matrix user with malicious intent, sharing a room with Synapse
instances before 1.104.1, can dispatch specially crafted events to
exploit a weakness in how the auth chain cover index is calculated. This
can induce high CPU consumption and accumulate excessive data in the
database of such instances, resulting in a denial of service.
Servers in private federations, or those that do not federate, are not
affected.
During the migration the automated script to update the copyright
headers accidentally got rid of some of the existing copyright lines.
Reinstate them.
There are a couple of things we need to be careful of here:
1. The current python code does no validation when loading from the DB,
so we need to be careful to ignore such errors (at least on jki.re there
are some old events with internal metadata fields of the wrong type).
2. We want to be memory efficient, as we often have many hundreds of
thousands of events in the cache at a time.
---------
Co-authored-by: Quentin Gliech <quenting@element.io>
* Fix the CI query that did not detect all cases of missing primary keys
* Add more missing REPLICA IDENTITY entries
* Newsfile
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
---------
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
* Add Postgres replica identities to tables that don't have an implicit one
Fixes#16224
* Newsfile
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
* Move the delta to version 83 as we missed the boat for 82
* Add a test that all tables have a REPLICA IDENTITY
* Extend the test to include when indices are deleted
* isort
* black
* Fully qualify `oid` as it is a 'hidden attribute' in Postgres 11
* Update tests/storage/test_database.py
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
* Add missed tables
---------
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
If simple_{insert,upsert,update}_many_txn is called without any data
to modify then return instead of executing the query.
This matches the behavior of simple_{select,delete}_many_txn.
* Fix bug where a new writer advances their token too quickly
When starting a new writer (for e.g. persisting events), the
`MultiWriterIdGenerator` doesn't have a minimum token for it as there
are no rows matching that new writer in the DB.
This results in the the first stream ID it acquired being announced as
persisted *before* it actually finishes persisting, if another writer
gets and persists a subsequent stream ID. This is due to the logic of
setting the minimum persisted position to the minimum known position of
across all writers, and the new writer starts off not being considered.
* Fix sending out POSITIONs when our token advances without update
Broke in #14820
* For replication HTTP requests, only wait for minimal position
* Fix rare bug that broke looping calls
We can't interact with the reactor from the main thread via looping
call.
Introduced in v1.90.0 / #15791.
* Newsfile
If we don't have all the auth events in a room then not all state events will have a chain cover index. Even so, we can still use the chain cover index on the events that do have it, rather than bailing and using the slower functions.
This situation should not arise for newly persisted rooms, as we check we have the full auth chain for each event, but can happen for existing rooms.
c.f. #15245
We were seeing serialization errors when taking out multiple read locks.
The transactions were retried, so isn't causing any failures.
Introduced in #15782.
Adds three new configuration variables:
* destination_min_retry_interval is identical to before (10mn).
* destination_retry_multiplier is now 2 instead of 5, the maximum value will
be reached slower.
* destination_max_retry_interval is one day instead of (essentially) infinity.
Capping this will cause destinations to continue to be retried sometimes instead
of being lost forever. The previous value was 2 ^ 62 milliseconds.
Allow configuring the set of workers to proxy outbound federation traffic through (`outbound_federation_restricted_to`).
This is useful when you have a worker setup with `federation_sender` instances responsible for sending outbound federation requests and want to make sure *all* outbound federation traffic goes through those instances. Before this change, the generic workers would still contact federation themselves for things like profile lookups, backfill, etc. This PR allows you to set more strict access controls/firewall for all workers and only allow the `federation_sender`'s to contact the outside world.
Allow configuring the set of workers to proxy outbound federation traffic through (`outbound_federation_restricted_to`).
This is useful when you have a worker setup with `federation_sender` instances responsible for sending outbound federation requests and want to make sure *all* outbound federation traffic goes through those instances. Before this change, the generic workers would still contact federation themselves for things like profile lookups, backfill, etc. This PR allows you to set more strict access controls/firewall for all workers and only allow the `federation_sender`'s to contact the outside world.
The original code is from @erikjohnston's branches which I've gotten in-shape to merge.
* Fix#15669: always populate instance map even if it was empty
* Fix some tests
* Fix more tests
* Newsfile
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
* CI fix: don't forget to update apt repository sources before installing olddeps deps
* Add test testing the backwards compatibility
---------
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
The cached decorators always return a Deferred, which was not
properly propagated. It was close enough when wrapping coroutines,
but failed if a bare function was wrapped.
Before this change:
* `PerspectivesKeyFetcher` and `ServerKeyFetcher` write to `server_keys_json`.
* `PerspectivesKeyFetcher` also writes to `server_signature_keys`.
* `StoreKeyFetcher` reads from `server_signature_keys`.
After this change:
* `PerspectivesKeyFetcher` and `ServerKeyFetcher` write to `server_keys_json`.
* `PerspectivesKeyFetcher` also writes to `server_signature_keys`.
* `StoreKeyFetcher` reads from `server_keys_json`.
This results in `StoreKeyFetcher` now using the results from `ServerKeyFetcher`
in addition to those from `PerspectivesKeyFetcher`, i.e. keys which are directly
fetched from a server will now be pulled from the database instead of refetched.
An additional minor change is included to avoid creating a `PerspectivesKeyFetcher`
(and checking it) if no `trusted_key_servers` are configured.
The overall impact of this should be better usage of cached results:
* If a server has no trusted key servers configured then it should reduce how often keys
are fetched.
* if a server's trusted key server does not have a requested server's keys cached then it
should reduce how often keys are directly fetched.
* Change `store_server_verify_keys` to take a `Mapping[(str, str), FKR]`
This is because we already can't handle duplicate keys — leads to cardinality violation
* Newsfile
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
---------
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
* Revert "Fix registering a device on an account with lots of devices (#15348)"
This reverts commit f0d8f66eaa.
* Revert "Delete stale non-e2e devices for users, take 3 (#15183)"
This reverts commit 78cdb72cd6.
Previously, we would spin in a tight loop until
`update_state_for_partial_state_event` stopped raising
`FederationPullAttemptBackoffError`s. Replace the spinloop with a wait
until the backoff period has expired.
Signed-off-by: Sean Quah <seanq@matrix.org>
This should help reduce the number of devices e.g. simple bots the repeatedly login rack up.
We only delete non-e2e devices as they should be safe to delete, whereas if we delete e2e devices for a user we may accidentally break their ability to receive e2e keys for a message.
Additionally:
* Consistently use `freeze()` in test
---------
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
Co-authored-by: 6543 <6543@obermui.de>
* Fix a long-standing bug where non-ASCII characters in search terms,
including accented letters, would not match characters in a different
case.
* Fix a long-standing bug where search terms using combining accents
would not match display names using precomposed accents and vice
versa.
To fully take effect, the user directory must be rebuilt after this
change.
Fixes#14630.
Signed-off-by: Sean Quah <seanq@matrix.org>
* Change `create_room` return type
* Don't return room alias from /createRoom
* Update other callsites
* Fix up mypy complaints
It looks like new_room_user_id is None iff new_room_id is None. It's a
shame we haven't expressed this in a way that mypy can understand.
* Changelog
The previous assumption was that the stream_id column was unique
(for a room ID, receipt type, user ID tuple), but this turned out to be
incorrect.
Now find the max stream ID, then map this back to a database-specific
row identifier and delete other rows which match the (room ID, receipt type,
user ID) tuple, but *not* the row ID.
