Debug for #8631.
I'm having a hard time tracking down what's going wrong in that issue.
In the reported example, I could see server A sending federation traffic
to server B and all was well. Yet B reports out-of-sync device updates
from A.
I couldn't see what was _in_ the events being sent from A to B. So I
have added some crude logging to track
- when we have updates to send to a remote HS
- the edus we actually accumulate to send
- when a federation transaction includes a device list update edu
- when such an EDU is received
This is a bit of a sledgehammer.
I've never found this terribly useful. I think it was added in the early days
of Synapse, without much thought as to what would actually be useful to log,
and has just been cargo-culted ever since.
Rather, it tends to clutter up debug logs with useless information.
Always add state.room_id after the configurable ORDER BY. Otherwise,
for any sort, certain pages can contain results from
other pages. (Especially when sorting by creator, since there may
be many rooms by the same creator)
* Document different order direction of numerical fields
"joined_members", "joined_local_members", "version" and "state_events"
are ordered in descending direction by default (dir=f). Added a note
in tests to explain the differences in ordering.
Signed-off-by: Daniël Sonck <daniel@sonck.nl>
This makes the serialization of events synchronous (and it no
longer access the database), but we must manually calculate and
provide the bundled aggregations.
Overall this should cause no change in behavior, but is prep work
for other improvements.
* Push `get_room_{min,max_stream_ordering}` into StreamStore
Both implementations of this are identical, so we may as well push it down and
get rid of the abstract base class nonsense.
* Remove redundant `StreamStore` class
This is empty now
* Remove redundant `get_current_events_token`
This was an exact duplicate of `get_room_max_stream_ordering`, so let's get rid
of it.
* newsfile
A couple of safety-checks to hopefully stop people doing what I just did, and create a storage
function which only works the first time it is called (and not when it is re-run due to a database
concurrency error or similar).
Create a new dict helper method `simple_insert_many_values_txn`, which takes
raw row values, rather than {key=>value} dicts. This saves us a bunch of dict
munging, and makes it easier to use generators rather than creating
intermediate lists and dicts.
This mainly consists of docstrings and inline comments. There are one or two type annotations and variable renames thrown in while I was here.
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
MSC3030: https://github.com/matrix-org/matrix-doc/pull/3030
Client API endpoint. This will also go and fetch from the federation API endpoint if unable to find an event locally or we found an extremity with possibly a closer event we don't know about.
```
GET /_matrix/client/unstable/org.matrix.msc3030/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>
{
"event_id": ...
"origin_server_ts": ...
}
```
Federation API endpoint:
```
GET /_matrix/federation/unstable/org.matrix.msc3030/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>
{
"event_id": ...
"origin_server_ts": ...
}
```
Co-authored-by: Erik Johnston <erik@matrix.org>
Part of https://github.com/matrix-org/synapse/issues/11300
Call stack:
- `_persist_events_and_state_updates` (added `use_negative_stream_ordering`)
- `_persist_events_txn`
- `_update_room_depths_txn` (added `update_room_forward_stream_ordering`)
- `_update_metadata_tables_txn`
- `_store_room_members_txn` (added `inhibit_local_membership_updates`)
Using keyword-only arguments (`*`) to reduce the mistakes from `backfilled` being left as a positional argument somewhere and being interpreted wrong by our new arguments.
The previous fix for the ongoing event fetches counter
(8eec25a1d9) was both insufficient and
incorrect.
When the database is unreachable, `_do_fetch` never gets run and so
`_event_fetch_ongoing` is never decremented.
The previous fix also moved the `_event_fetch_ongoing` decrement outside
of the `_event_fetch_lock` which allowed race conditions to corrupt the
counter.
* remove background update code related to deprecated config flag
* changelog entry
* update changelog
* Delete 11394.removal
Duplicate, wrong number
* add no-op background update and change newfragment so it will be consolidated with associated work
* remove unused code
* Remove code associated with deprecated flag from legacy docker dynamic config file
Co-authored-by: reivilibre <oliverw@matrix.org>
Instead of only known relation types. This also reworks the background
update for thread relations to crawl events and search for any relation
type, not just threaded relations.
Adds validation to the Client-Server API to ensure that
the potential thread head does not relate to another event
already. This results in not allowing a thread to "fork" into
other threads.
If the target event is unknown for some reason (maybe it isn't
visible to your homeserver), but is the target of other events
it is assumed that the thread can be created from it. Otherwise,
it is rejected as an unknown event.
==============================
Bugfixes
--------
- Fix a bug introduced in 1.47.0rc1 which caused worker processes to not halt startup in the presence of outstanding database migrations. ([\#11346](https://github.com/matrix-org/synapse/issues/11346))
- Fix a bug introduced in 1.47.0rc1 which prevented the 'remove deleted devices from `device_inbox` column' background process from running when updating from a recent Synapse version. ([\#11303](https://github.com/matrix-org/synapse/issues/11303), [\#11353](https://github.com/matrix-org/synapse/issues/11353))
-----BEGIN PGP SIGNATURE-----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=iNww
-----END PGP SIGNATURE-----
gpgsig -----BEGIN PGP SIGNATURE-----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=c7gi
-----END PGP SIGNATURE-----
Merge tag 'v1.47.0rc3' into develop
Synapse 1.47.0rc3 (2021-11-16)
==============================
Bugfixes
--------
- Fix a bug introduced in 1.47.0rc1 which caused worker processes to not halt startup in the presence of outstanding database migrations. ([\#11346](https://github.com/matrix-org/synapse/issues/11346))
- Fix a bug introduced in 1.47.0rc1 which prevented the 'remove deleted devices from `device_inbox` column' background process from running when updating from a recent Synapse version. ([\#11303](https://github.com/matrix-org/synapse/issues/11303), [\#11353](https://github.com/matrix-org/synapse/issues/11353))
It already seems to pass mypy. I wonder what changed, given that it was
on the exclusion list. So this commit consists of me ensuring
`--disallow-untyped-defs` passes and a minor fixup to a function that
returned either `True` or `None`.
* remove unused tables room_stats_historical and user_stats_historical
* update changelog number
* Bump schema compat version comment
* make linter happy
* Update comment to give more info
Co-authored-by: reivilibre <oliverw@matrix.org>
Co-authored-by: reivilibre <oliverw@matrix.org>
* Prefer `HTTPStatus` over plain `int`
This is an Opinion that no-one has seemed to object to yet.
* `--disallow-untyped-defs` for `tests.rest.client.test_directory`
* Improve synapse's annotations for deleting aliases
* Test case for deleting a room alias
* Changelog
* change display names/avatar URLS to None if they contain null bytes
* add changelog
* add POC test, requested changes
* add a saner test and remove old one
* update test to verify that display name has been changed to None
* make test less fragile
* Make DataStore inherit from EventForwardExtremitiesStore before CacheInvalidationWorkerStore
the former implicitly inherits from the latter, so they should be
ordered like this when used.
Co-authored-by: Dirk Klimpel <5740567+dklimpel@users.noreply.github.com>
Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com>
* Make lock better handle process being killed
If the process gets killed and restarted (so that it didn't have a
chance to drop its locks gracefully) then there may still be locks in
the DB that are for the same instance that haven't yet timed out but are
safe to delete.
We handle this case by a) checking if the current instance already has
taken out the lock, and b) if not then ignoring locks that are for the
same instance.
* Periodically check for old staged events
This is to protect against other instances dying and their locks timing
out.
When an event fetcher aborts due to an exception, `_event_fetch_ongoing`
must be decremented, otherwise the event fetcher would never be
replaced. If enough event fetchers were to fail, no more events would be
fetched and requests would get stuck waiting for events.
