The extended control node is rabbitmq error reporting

[LiKui]

Rabbitmq exception when extending a single control node to a 3 control node
On the extended two control nodes, the following is shown
#docker logs -f rabbitmq
.......
=INFO REPORT==== 8-Nov-2017::21:43:22 ===
Multiple different configurations with equal version numbers detected. Shutting down.

[Huikai Ran (huikai.ran)]

I have the same issue. with one rabbitmq, it works fine, after I deployed other two rabbitmq, the other two cannot startup, keeping restarting.

[doubor (doubor)]

I have the same issue.

[Florian Faltermeier (florianfa)]

Same issue on rocky:
=INFO REPORT==== 15-Feb-2019::14:21:27 ===
Multiple different configurations with equal version numbers detected. Shutting down.

[Michal Nasiadka (mnasiadka)]

Currently a workaround is to manually bump up the config version.
TODO: test on Stein (if new clusterer setup is fixing that)

[Grzegorz Sikorski (gregors)]

Please see attached patch file. This solves the issue for us.

[Radosław Piliszek (yoctozepto)]

Well, we can't really propose to raise the cluster version with each run.

I haven't seen this issue since Stein - could it be gone?

[Grzegorz Sikorski (gregors)]

Possibly, we are using Rocky and Ocata, haven't yet tested with further version. Let me try to refactor this piece to be idempotent.

[Grzegorz Sikorski (gregors)]

Now bumping of the version only occurs if anything in the configs is modified.

[Radosław Piliszek (yoctozepto)]

Let's say I'm not entirely convinced you got it to work considering it has a typo: "cluser_version" in template. Otherwise looks pretty legit. Please propose to gerrit.

[Grzegorz Sikorski (gregors)]

Sorry, the typo was corrected and tested afterwards. I must have done something stupid with git.

[Radosław Piliszek (yoctozepto)]

All right, see https://docs.openstack.org/kolla/latest/contributor/CONTRIBUTING.html

[Grzegorz Sikorski (gregors)]

OK, I'll push this to Gerrit as soon as some legal stuff pass here :D

[Grzegorz Sikorski (gregors)]

https://review.opendev.org/713628

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/713628
Committed: https://git.openstack.org/cgit/openstack/kolla-ansible/commit/?id=7408d499799e4d3a043f5e5ebf5f61c206bc0d2e
Submitter: Zuul
Branch: stable/rocky

commit 7408d499799e4d3a043f5e5ebf5f61c206bc0d2e
Author: Grzegorz Sikorski <email address hidden>
Date: Tue Mar 17 09:07:03 2020 +0100

    rabbitmq: update cluster version on reconfigure

    This fixes https://bugs.launchpad.net/kolla-ansible/+bug/1731142. The
    solution is to bump a version number of the cluster configuration
    file (/etc/kolla/rabbitmq/rabbitmq-clusterer.config) every time
    RabbitMQ configuration files change.

    Closes-bug: #1731142
    Change-Id: I26cce92e52b2ba6b905658e11d794c65e361f545

[OpenStack Infra (hudson-openstack)]

Change abandoned by Grzegorz Sikorski (<email address hidden>) on branch: stable/ocata
Review: https://review.opendev.org/713878

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/kolla-ansible rocky-eol release.