prevent event loop polling from stopping on active redis connections #1734
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes: celery/celery#7276
Caused by: #1476
Currently, the Redis transport will always remove the last function added to the event loop (
on_tick
) regardless of which connection is disconnected. If you restart Redis enough (and repeatedly cause connection errors), eventually_on_connection_disconnect
will remove a function from the event loop for the only active connection and the worker will get stuck.I've been seeing these issues with workers getting stuck after I migrated from the RabbitMQ broker to Redis on Celery 5.2.6 and Kombu 5.2.3.
This PR changes it to track which event loop function belongs to which connection and remove the correct function from
on_tick
.