-
Bug
-
Resolution: Done
-
Undefined
-
None
Clowder in stage appears is misrepresenting the state of a deployment in cloudigrade-stage. Here it says it's not ready:
❯ oc get ClowdApp/postigrade
NAME READY MANAGED ENVNAME AGE
postigrade 0 1 stage 442d
But here it says it's ready:
❯ oc get ClowdApp/postigrade -o json | jq .status.ready
true
This is blocking cloudigrade's release process. The qontract reconcile tekton pipeline appears to be stuck in a loop waiting forever for the ready status:
[2023-01-11 20:17:21] [INFO] [openshift_base.py:validate_realized_data:868] - ['validating', 'crcs02ue1', 'cloudigrade-stage', 'ClowdApp', 'postigrade']
[2023-01-11 20:17:21] [INFO] [openshift_base.py:validate_realized_data:928] - ClowdApp has deployments that are not ready (0 ready / 1 total)
psavage@redhat.com confirmed in Slack that this is a bug in Clowder:
So yeh - we made a fix a few weeks back which only wrote the status of a resource when there was an actual change. The problem with this was the routine that detects if there has been a change didn't know about the deployment stats changes we made.
Slack discussion thread: https://redhat-internal.slack.com/archives/C022YV4E0NA/p1673468424641509
Pete's initial proposed fix: https://github.com/RedHatInsights/clowder/pull/752