I would like to seek some help regarding the issue I am facing while performing in series upgrade on the Antelope version.

We have currently deployed our cluster with 2023.1 Antelope version. The cluster was deployed on a clean environment by cleaning up any previous kolla openstack installation.

We have not observed any issues with this environment so far.

For a feature I am working on, I had to put a patch into GLANCE image and rebuild all kolla images with a new tag.

I have attempted to upgrade the cluster using both kolla-deploy and kolla-upgrade methods, but both methods have left the cluster in degraded condition.

The issue seems to be mainly with failure of communication between neutron-server and all/some neutron-agents, due to which the neutron-server sees all/some

neutron-agents dead. Rabbitmq itself seems to be doing fine. The issue is very consistent.

Is this a known behavior? And am I using the correct kolla tools to perform the in series upgrade.

Appreciate any help.

Thanks,

Shrishail