quincy: cephadm: handle exceptions applying extra services during bootstrap #50904
Conversation
Otherwise we risk hitting a mismatch between the cephadm binary version and the container image version we're bootstrapping on, resulting in bootstrap failing. Example in the tracker. Fixes: https://tracker.ceph.com/issues/59082 Signed-off-by: Adam King <adking@redhat.com> (cherry picked from commit a57fc00)
|
4 dead jobs are a general issue I'm seeing with the upgrade-with-workload task. It seems the upgrade completes but the workload test at the end never does for some reason. This started happening on multiple branches simultaneously. For example, it came up in the reef baseline run (https://pulpito.ceph.com/yuriw-2023-04-08_15:50:42-orch-reef-distro-default-smithi/7235797). Due to it showing up in multiple stable branches at once with different codebases, thinking it's likely not an actual bug within ceph and shouldn't block PR merging, although it will require investigation. Rerun of failed jobs: https://pulpito.ceph.com/adking-2023-04-08_15:01:19-orch:cephadm-wip-adk3-testing-2023-04-05-1604-quincy-distro-default-smithi/ After reruns, 2 failures. Both are same test that is failing because it expects some command to give a zero error code but give an error message and instead it's giving a nonzero return code. To clarify, the command is meant to fail (it's checking on a nonexistent cluster) it's just the test expects it to fail a different way. I think it's just an issue of the test being backported without the change to the actual function's returncode, and shouldn't block merging. |
backport tracker: https://tracker.ceph.com/issues/59302
backport of #50548
parent tracker: https://tracker.ceph.com/issues/59082
this backport was staged using ceph-backport.sh version 16.0.0.6848
find the latest version at https://github.com/ceph/ceph/blob/main/src/script/ceph-backport.sh