-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
rhos-18.0.15
-
None
-
0
-
False
-
-
False
-
?
-
rhos-ops-day1day2-upgrades
-
None
-
-
-
-
Important
To Reproduce Steps to reproduce the behavior:
Customer tried to update containers from 18.0.7-20250408.2 bundle to 18.0.15-20251126.192455 bundle in RHOSO environment by running data plane deployment with servicesOverride set to update-services.
Results were mixed:
- all RHOSO containers except logrotate_crond were re-created (customer is also running Trillio containers, let's put them aside);
- in deployment's metadata desired version (18.0.15-20251126.192455 is listed);
- some containers are still using images from original bundle (18.0.7-20250408.2): ceilometer_agent_compute, logrotate_crond, podman_exporter, openstack_network_exporter, node_exporter, neutron-haproxy-ovnmeta-UUID
- remaining RHOSO containers were moved to new images
It looks like for some reason appropriate container images were not translated from RHOSO operators to Ansible variables.
Expected behavior
Containers are using consistent container images
Bug impact
Update is blocked
Known workaround
None
Additional context
- sosreport from one of affected compute nodes is attached to support case
- must-gather is attached to support case