Have replication clients remove _INT_STREAM_POS (#15309)

* Have replication clients remove _INT_STREAM_POS

Suppose worker A makes an internal http request from worker B. B may
make changes that A later learns about over replication. We want A's
request to block until it has seen those changes—mainly to ensure A's
caches are invalidated promptly. This helps provide read-after-write
consistency, eliminating entire categories of races and test flakes.

To implement this, B includes a top-level field `_INT_STREAM_POS` in its
response JSON. Roughly speaking, the field's value tells A what to wait
for. But we weren't removing that internal field before A's request
completed!

Introduced in https://github.com/matrix-org/synapse/pull/14820.
Fixes #15308.

* Changelog
This commit is contained in:
David Robertson 2023-03-22 12:53:55 +00:00 committed by GitHub
parent 72f3f23c4d
commit 1bc9985eb7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 2 additions and 1 deletions

1
changelog.d/15309.bugfix Normal file
View File

@ -0,0 +1 @@
Fix a bug introduced in Synapse 1.76.0 where responses from worker deployments could include an internal `_INT_STREAM_POS` key.

View File

@ -345,7 +345,7 @@ class ReplicationEndpoint(metaclass=abc.ABCMeta):
_outgoing_request_counter.labels(cls.NAME, 200).inc() _outgoing_request_counter.labels(cls.NAME, 200).inc()
# Wait on any streams that the remote may have written to. # Wait on any streams that the remote may have written to.
for stream_name, position in result.get( for stream_name, position in result.pop(
_STREAM_POSITION_KEY, {} _STREAM_POSITION_KEY, {}
).items(): ).items():
await replication.wait_for_stream_position( await replication.wait_for_stream_position(