Skip to content

Conversation

@kjnilsson
Copy link
Contributor

@kjnilsson kjnilsson commented Jun 27, 2019

As this will make these more work reliably when leadership is in flux.

Also ensure that servers started and added after the initial declare
uses the same configuration.

requires rabbitmq/ra#100.

As this will make these more work reliably when leadership is in flux.

Also ensure that servers started and added after the initial declare
uses the same configuration.
@gerhard
Copy link
Contributor

gerhard commented Jun 27, 2019

Looks good, ship it. How to we stress-test this in LREs? Would this help? https://github.com/rabbitmq/rabbitmq-server-boshrelease/tree/master/jobs/periodic_shutdown

@kjnilsson
Copy link
Contributor Author

I think it's tricky to test this specific thing but lots of queue declare during leader elections would be the way I'd approach it.

@gerhard
Copy link
Contributor

gerhard commented Jul 1, 2019

If nodes get periodically shut-down and consumers get forcefully disconnect, when consumers reconnect they should first try to declare queues before consuming from them. Would that work?

@michaelklishin
Copy link
Collaborator

@kjnilsson @dcorbacho FYI, this has a legitimate conflict with #2053.

@michaelklishin michaelklishin merged commit 098ac57 into master Jul 4, 2019
@michaelklishin michaelklishin deleted the qq-sync-op-fixes branch July 4, 2019 17:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants