Hello,
I've just worked most of the way through a fresh install of Yoga on 5 Debian Bullseye systems. All systems are updated to include the Bullseye backports. The documentation doesn't mention backports, but I always install Debian this way - almost without thinking.
The specific config is based on openstack-user-config.yaml.prod.example with 3 infrastructure host, two compute hosts, and cinder/glance running on the infrastructure hosts as in the prod example file.
First, one surprise: For version of GlusterFS in Bullseye Backports, /usr/sbin/gluster as been moved to a separate package - glusterfs-cli. I installed this manually in the repo containers to get through setup-infrastructure.yml.
In setup-openstack.yml, I'm stopped at "TASK [os_nova : Run nova-status upgrade check to validate a healthy configuration]". "nova-status upgrade check" is failing. "nova-manage cell_v2 list_hosts" is not showing any.
Oh, and there are warnings about eventlet monkey patching and urllib3.
So I'm not quite sure how to dig into this. The nova-api container seems to be running on the infra hosts, and nova-compute.service is up on both compute hosts, although there are warnings about "Timed out waiting for nova-conductor" The nova-api containers are able to ing the compute hosts on br-mgmt.
I do have to wonder if this has anything to do with being upgraded to backpors.
Any hints on how to analyse this (or how to fix it)?
Thanks.
-Dave