For the record, the reason we need this is as follows:
each RDATA command comes down the redis pipe as a subscription message. txredisapi as written needs at least three reactor ticks to read each subscription message from the tcp buffer. Hence, once the process gets loaded, it starts getting behind, and eventually redis knifes the connection. it then takes ages for the master to work its way through the backlog, before it reconnects again, during which any commands from any workers are dropped.
An update of check-manifest shone some light on some issues with MANIFEST.in, specifically that we didn't ignore/prune the contrib directory, and that we were using prune instead of exclude for files. This fixes both issues.
Fixes#7403
We forgot to set the password on the subscriber connection, as well as
not calling super methods for overridden connectionMade/connectionLost
functions.
For in memory streams when fetching updates on workers we need to query the source of the stream, which currently is hard coded to be master. This PR threads through the source instance we received via `POSITION` through to the update function in each stream, which can then be passed to the replication client for in memory streams.
We move the processing of typing and federation replication traffic into their handlers so that `Stream.current_token()` points to a valid token. This allows us to remove `get_streams_to_replicate()` and `stream_positions()`.
By persisting the user interactive authentication sessions to the database, this fixes
situations where a user hits different works throughout their auth session and also
allows sessions to persist through restarts of Synapse.
This is primarily for allowing us to send those commands from workers, but for now simply allows us to ignore echoed RDATA/POSITION commands that we sent (we get echoes of sent commands when using redis). Currently we log a WARNING on the master process every time we receive an echoed RDATA.
For direct TCP connections we need the master to relay REMOTE_SERVER_UP
commands to the other connections so that all instances get notified
about it. The old implementation just relayed to all connections,
assuming that sending back to the original sender of the command was
safe. This is not true for redis, where commands sent get echoed back to
the sender, which was causing master to effectively infinite loop
sending and then re-receiving REMOTE_SERVER_UP commands that it sent.
The fix is to ensure that we only relay to *other* connections and not
to the connection we received the notification from.
Fixes#7334.
* Factor out functions for injecting events into database
I want to add some more flexibility to the tools for injecting events into the
database, and I don't want to clutter up HomeserverTestCase with them, so let's
factor them out to a new file.
* Rework TestReplicationDataHandler
This wasn't very easy to work with: the mock wrapping was largely superfluous,
and it's useful to be able to inspect the received rows, and clear out the
received list.
* Fix AssertionErrors being thrown by EventsStream
Part of the problem was that there was an off-by-one error in the assertion,
but also the limit logic was too simple. Fix it all up and add some tests.
Specifically some tests for the typing stream, which means we test streams that fetch missing updates via HTTP (rather than via the DB).
We also shuffle things around a bit so that we create two separate `HomeServer` objects, rather than trying to insert a slaved store into places.
Note: `test_typing.py` is heavily inspired by `test_receipts.py`
It doesn't seem to be documented anywhere and means that you suddenly start losing metrics without any obvious reason when you go from monolith to workers (e.g. #7312).
