Fix race conditions in the async cache invalidation logic, by separating
the async & local invalidation calls and ensuring any async call i
executed first.
Signed off by Nick @ Beeper (@Fizzadar).
More prep work for asyncronous caching, also makes all process_replication_rows methods consistent (presence handler already is so).
Signed off by Nick @ Beeper (@Fizzadar)
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.
There are two fixes here:
1. A long-standing bug where we incorrectly calculated `delta_ids`; and
2. A bug introduced in #13267 where we got current state incorrect.
Some experimental prep work to enable external event caching based on #9379 & #12955. Doesn't actually move the cache at all, just lays the groundwork for async implemented caches.
Signed off by Nick @ Beeper (@Fizzadar)
* Replace `get_new_events_for_appservice` with `get_all_new_events_stream`
The functions were near identical and this brings the AS worker closer
to the way federation senders work which can allow for multiple workers
to handle AS traffic.
* Pull received TS alongside events when processing the stream
This avoids an extra query -per event- when both federation sender
and appservice pusher process events.
Bounce recalculation of current state to the correct event persister and
move recalculation of current state into the event persistence queue, to
avoid concurrent updates to a room's current state.
Also give recalculation of a room's current state a real stream
ordering.
Signed-off-by: Sean Quah <seanq@matrix.org>
This happened if we encountered a stream ordering in `event_push_actions` that had more rows than the batch size of the delete, as If we don't delete any rows in an iteration then the next time round we get the exact same stream ordering and get stuck.
Whenever we want to persist an event, we first compute an event context,
which includes the state at the event and a flag indicating whether the
state is partial. After a lot of processing, we finally try to store the
event in the database, which can fail for partial state events when the
containing room has been un-partial stated in the meantime.
We detect the race as a foreign key constraint failure in the data store
layer and turn it into a special `PartialStateConflictError` exception,
which makes its way up to the method in which we computed the event
context.
To make things difficult, the exception needs to cross a replication
request: `/fed_send_events` for events coming over federation and
`/send_event` for events from clients. We transport the
`PartialStateConflictError` as a `409 Conflict` over replication and
turn `409`s back into `PartialStateConflictError`s on the worker making
the request.
All client events go through
`EventCreationHandler.handle_new_client_event`, which is called in
*a lot* of places. Instead of trying to update all the code which
creates client events, we turn the `PartialStateConflictError` into a
`429 Too Many Requests` in
`EventCreationHandler.handle_new_client_event` and hope that clients
take it as a hint to retry their request.
On the federation event side, there are 7 places which compute event
contexts. 4 of them use outlier event contexts:
`FederationEventHandler._auth_and_persist_outliers_inner`,
`FederationHandler.do_knock`, `FederationHandler.on_invite_request` and
`FederationHandler.do_remotely_reject_invite`. These events won't have
the partial state flag, so we do not need to do anything for then.
The remaining 3 paths which create events are
`FederationEventHandler.process_remote_join`,
`FederationEventHandler.on_send_membership_event` and
`FederationEventHandler._process_received_pdu`.
We can't experience the race in `process_remote_join`, unless we're
handling an additional join into a partial state room, which currently
blocks, so we make no attempt to handle it correctly.
`on_send_membership_event` is only called by
`FederationServer._on_send_membership_event`, so we catch the
`PartialStateConflictError` there and retry just once.
`_process_received_pdu` is called by `on_receive_pdu` for incoming
events and `_process_pulled_event` for backfill. The latter should never
try to persist partial state events, so we ignore it. We catch the
`PartialStateConflictError` in `on_receive_pdu` and retry just once.
Refering to the graph of code paths in
https://github.com/matrix-org/synapse/issues/12988#issuecomment-1156857648
may make the above make more sense.
Signed-off-by: Sean Quah <seanq@matrix.org>
==============================
Bugfixes
--------
- Update the version of the [ldap3 plugin](https://github.com/matrix-org/matrix-synapse-ldap3/) included in the `matrixdotorg/synapse` DockerHub images and the Debian packages hosted on `packages.matrix.org` to 0.2.1. This fixes [a bug](https://github.com/matrix-org/matrix-synapse-ldap3/pull/163) with usernames containing uppercase characters. ([\#13156](https://github.com/matrix-org/synapse/issues/13156))
- Fix a bug introduced in Synapse 1.62.0rc1 affecting unread counts for users on small servers. ([\#13168](https://github.com/matrix-org/synapse/issues/13168))
-----BEGIN PGP SIGNATURE-----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=SpI6
-----END PGP SIGNATURE-----
Merge tag 'v1.62.0rc3' into develop
Synapse 1.62.0rc3 (2022-07-04)
==============================
Bugfixes
--------
- Update the version of the [ldap3 plugin](https://github.com/matrix-org/matrix-synapse-ldap3/) included in the `matrixdotorg/synapse` DockerHub images and the Debian packages hosted on `packages.matrix.org` to 0.2.1. This fixes [a bug](https://github.com/matrix-org/matrix-synapse-ldap3/pull/163) with usernames containing uppercase characters. ([\#13156](https://github.com/matrix-org/synapse/issues/13156))
- Fix a bug introduced in Synapse 1.62.0rc1 affecting unread counts for users on small servers. ([\#13168](https://github.com/matrix-org/synapse/issues/13168))
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.
* 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.
By always using delete_devices and sometimes passing a list
with a single device ID.
Previously these methods had gotten out of sync with each
other and it seems there's little benefit to the single-device
variant.
* Update worker docs to remove group endpoints.
* Removes an unused parameter to `ApplicationService`.
* Break dependency between media repo and groups.
* Avoid copying `m.room.related_groups` state events during room upgrades.
Currently, we try to pull the event corresponding to a sync token from the database. However, when
we fetch redaction events, we check the target of that redaction (because we aren't allowed to send
redactions to clients without validating them). So, if the sync token points to a redaction of an event
that we don't have, we have a problem.
It turns out we don't really need that event, and can just work with its ID and metadata, which
sidesteps the whole problem.
* Fix room deletion
ae7858f broke room deletion by attempting to delete the entry from `rooms`
before the tables that reference it.
* faster_joins: remove database rows on purge
By always returning all requested values from the function
wrapped by cachedList. Otherwise implicit None values get
added into the cache, which are unexpected.
Implements the following behind an experimental configuration flag:
* A new push rule kind for mutually related events.
* A new default push rule (`.m.rule.thread_reply`) under an unstable prefix.
This is missing part of MSC3772:
* The `.m.rule.thread_reply_to_me` push rule, this depends on MSC3664 / #11804.
==============================
Synapse 1.59 makes several changes that server administrators should be aware of:
- Device name lookup over federation is now disabled by default. ([\#12616](https://github.com/matrix-org/synapse/issues/12616))
- The `synapse.app.appservice` and `synapse.app.user_dir` worker application types are now deprecated. ([\#12452](https://github.com/matrix-org/synapse/issues/12452), [\#12654](https://github.com/matrix-org/synapse/issues/12654))
See [the upgrade notes](https://github.com/matrix-org/synapse/blob/develop/docs/upgrade.md#upgrading-to-v1590) for more details.
Additionally, this release removes the non-standard `m.login.jwt` login type from Synapse. It can be replaced with `org.matrix.login.jwt` for identical behaviour. This is only used if `jwt_config.enabled` is set to `true` in the configuration. ([\#12597](https://github.com/matrix-org/synapse/issues/12597))
Bugfixes
--------
- Fix a bug introduced in Synapse 1.58.0 where `/sync` would fail if the most recent event in a room was rejected. ([\#12729](https://github.com/matrix-org/synapse/issues/12729))
-----BEGIN PGP SIGNATURE-----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=iA2S
-----END PGP SIGNATURE-----
gpgsig -----BEGIN PGP SIGNATURE-----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=JmDV
-----END PGP SIGNATURE-----
Merge tag 'v1.59.0rc2' into develop
Synapse 1.59.0rc2 (2022-05-16)
==============================
Synapse 1.59 makes several changes that server administrators should be aware of:
- Device name lookup over federation is now disabled by default. ([\#12616](https://github.com/matrix-org/synapse/issues/12616))
- The `synapse.app.appservice` and `synapse.app.user_dir` worker application types are now deprecated. ([\#12452](https://github.com/matrix-org/synapse/issues/12452), [\#12654](https://github.com/matrix-org/synapse/issues/12654))
See [the upgrade notes](https://github.com/matrix-org/synapse/blob/develop/docs/upgrade.md#upgrading-to-v1590) for more details.
Additionally, this release removes the non-standard `m.login.jwt` login type from Synapse. It can be replaced with `org.matrix.login.jwt` for identical behaviour. This is only used if `jwt_config.enabled` is set to `true` in the configuration. ([\#12597](https://github.com/matrix-org/synapse/issues/12597))
Bugfixes
--------
- Fix a bug introduced in Synapse 1.58.0 where `/sync` would fail if the most recent event in a room was rejected. ([\#12729](https://github.com/matrix-org/synapse/issues/12729))
Parse the `m.relates_to` event content field (which describes relations)
in a single place, this is used during:
* Event persistence.
* Validation of the Client-Server API.
* Fetching bundled aggregations.
* Processing of push rules.
Each of these separately implement the logic and each made slightly
different assumptions about what was valid. Some had minor / potential
bugs.
Refactor how the `EventContext` class works, with the intention of reducing the amount of state we fetch from the DB during event processing.
The idea here is to get rid of the cached `current_state_ids` and `prev_state_ids` that live in the `EventContext`, and instead defer straight to the database (and its caching).
One change that may have a noticeable effect is that we now no longer prefill the `get_current_state_ids` cache on a state change. However, that query is relatively light, since its just a case of reading a table from the DB (unlike fetching state at an event which is more heavyweight). For deployments with workers this cache isn't even used.
Part of #12684
* Add mau_appservice_trial_days
* Add a test
* Tweaks
* changelog
* Ensure we sync after the delay
* Fix types
* Add config statement
* Fix test
* Reinstate logging that got removed
* Fix feature name
* Changes hidden read receipts to be a separate receipt type
(instead of a field on `m.read`).
* Updates the `/receipts` endpoint to accept `m.fully_read`.
The `latest_event` field of the bundled aggregations for `m.thread` relations
did not include bundled aggregations itself. This resulted in clients needing to
immediately request the event from the server (and thus making it useless that
the latest event itself was serialized instead of just including an event ID).
I've seen a few errors which can only plausibly be explained by the calculated
event id for an event being different from the ID of the event in the
database. It should be cheap to check this, so let's do so and raise an
exception.
This works by taking a row level lock on the `rooms` table at the start of both transactions, ensuring that they don't run at the same time. In the event persistence transaction we also check that there is an entry still in the `rooms` table.
I can't figure out how to do this in SQLite. I was just going to lock the table, but it seems that we don't support that in SQLite either, so I'm *really* confused as to how we maintain integrity in SQLite when using `lock_table`....
Over time we've begun to use newer versions of mypy, typeshed, stub
packages---and of course we've improved our own annotations. This makes
some type ignore comments no longer necessary. I have removed them.
There was one exception: a module that imports `select.epoll`. The
ignore is redundant on Linux, but I've kept it ignored for those of us
who work on the source tree using not-Linux. (#11771)
I'm more interested in the config line which enforces this. I want
unused ignores to be reported, because I think it's useful feedback when
annotating to know when you've fixed a problem you had to previously
ignore.
* Installing extras before typechecking
Lacking an easy way to install all extras generically, let's bite the bullet and
make install the hand-maintained `all` extra before typechecking.
Now that https://github.com/matrix-org/backend-meta/pull/6 is merged to
the release/v1 branch.
Try to avoid an OOM by checking fewer extremities.
Generally this is a big rewrite of _maybe_backfill, to try and fix some of the TODOs and other problems in it. It's best reviewed commit-by-commit.
Multiple calls to `EventsWorkerStore._get_events_from_cache_or_db` can
reuse the same database fetch, which is initiated by the first call.
Ensure that cancelling the first call doesn't cancel the other calls
sharing the same database fetch.
Signed-off-by: Sean Quah <seanq@element.io>
This will mainly be useful when dealing with module callbacks, which are
all typed as returning `Awaitable`s instead of coroutines or
`Deferred`s.
Signed-off-by: Sean Quah <seanq@element.io>
When we join a room via the faster-joins mechanism, we end up with "partial
state" at some points on the event DAG. Many parts of the codebase need to
wait for the full state to load. So, we implement a mechanism to keep track of
which events have partial state, and wait for them to be fully-populated.
We work through all the events with partial state, updating the state at each
of them. Once it's done, we recalculate the state for the whole room, and then
mark the room as having complete state.
* Add some type hints to datastore
* newsfile
* change `Collection` to `List`
* refactor return type of `select_users_txn`
* correct type hint in `stream.py`
* Remove `Optional` in `select_users_txn`
* remove not needed return type in `__init__`
* Revert change in `get_stream_id_for_event_txn`
* Remove import from `Literal`
Consider the requester's ignored users when calculating the
bundled aggregations.
See #12285 / 4df10d3214
for corresponding changes for the `/relations` endpoint.
Of note:
* No untyped defs in `register_new_matrix_user`
This one might be contraversial. `request_registration` has three
dependency-injection arguments used for testing. I'm removing the
injection of the `requests` module and using `unitest.mock.patch` in the
test cases instead.
Doing `reveal_type(requests)` and `reveal_type(requests.get)` before the
change:
```
synapse/_scripts/register_new_matrix_user.py:45: note: Revealed type is "Any"
synapse/_scripts/register_new_matrix_user.py:46: note: Revealed type is "Any"
```
And after:
```
synapse/_scripts/register_new_matrix_user.py:44: note: Revealed type is "types.ModuleType"
synapse/_scripts/register_new_matrix_user.py:45: note: Revealed type is "def (url: Union[builtins.str, builtins.bytes], params: Union[Union[_typeshed.SupportsItems[Union[builtins.str, builtins.bytes, builtins.int, builtins.float], Union[builtins.str, builtins.bytes, builtins.int, builtins.float, typing.Iterable[Union[builtins.str, builtins.bytes, builtins.int, builtins.float]], None]], Tuple[Union[builtins.str, builtins.bytes, builtins.int, builtins.float], Union[builtins.str, builtins.bytes, builtins.int, builtins.float, typing.Iterable[Union[builtins.str, builtins.bytes, builtins.int, builtins.float]], None]], typing.Iterable[Tuple[Union[builtins.str, builtins.bytes, builtins.int, builtins.float], Union[builtins.str, builtins.bytes, builtins.int, builtins.float, typing.Iterable[Union[builtins.str, builtins.bytes, builtins.int, builtins.float]], None]]], builtins.str, builtins.bytes], None] =, data: Union[Any, None] =, headers: Union[Any, None] =, cookies: Union[Any, None] =, files: Union[Any, None] =, auth: Union[Any, None] =, timeout: Union[Any, None] =, allow_redirects: builtins.bool =, proxies: Union[Any, None] =, hooks: Union[Any, None] =, stream: Union[Any, None] =, verify: Union[Any, None] =, cert: Union[Any, None] =, json: Union[Any, None] =) -> requests.models.Response"
```
* Drive-by comment in `synapse.storage.types`
* No untyped defs in `synapse_port_db`
This was by far the most painful. I'm happy to break this up into
smaller pieces for review if it's not managable as-is.
Principally, `prometheus_client.REGISTRY.register` now requires its argument to
extend `prometheus_client.Collector`.
Additionally, `Gauge.set` is now annotated so that passing `Optional[int]`
causes an error.
Refactor and convert `Linearizer` to async. This makes a `Linearizer`
cancellation bug easier to fix.
Also refactor to use an async context manager, which eliminates an
unlikely footgun where code that doesn't immediately use the context
manager could forget to release the lock.
Signed-off-by: Sean Quah <seanq@element.io>
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.
Patrick Cloke