InstallPlan cannot resolve a change to immutable field in operator Deployment spec

[aharbis]

Bug Report

What did you do?

Between two releases of our operator, we modified the labels on our operator's Deployment resource.

What did you expect to see?

OLM successfully resolve the new InstallPlan and Deployment spec, and perform the update.

What did you see instead? Under which circumstances?

OLM failed to perform the update, and the InstallPlan failed with an error such as:

install strategy failed: Deployment.apps "datapower-operator" is invalid: spec.selector: Invalid value: v1.LabelSelector{MatchLabels:map[string]string{"app.kubernetes.io/instance":"datapower-operator", "app.kubernetes.io/managed-by":"datapower-operator", "app.kubernetes.io/name":"datapower-operator", "name":"datapower-operator"}, MatchExpressions:[]v1.LabelSelectorRequirement(nil)}: field is immutable

Environment

• operator-lifecycle-manager version:

OCP 4.4.3, not sure on specific OLM version.

• Kubernetes version information:

$ oc version
Client Version: openshift-clients-4.3.0-201910250623-88-g6a937dfe
Server Version: 4.4.3
Kubernetes Version: v1.17.1

• Kubernetes cluster kind: OCP

Possible Solution

If some change is found during the InstallPlan such as this, which causes a diff on an immutable field, the Deployment could be deleted by OLM and recreated to resolve the issue (this is what we had to do manually as a workaround).

Additional context
Add any other context about the problem here.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[aharbis]

bump

[prafull01]

@exdx If no one is working on this, I can pick it up..

[cdjohnson]

@aharbis There was a workaround documented here:
#952 (comment)

Basically, if you are adding a label or otherwise immutable change, you can just rename the deployment, which will cause a new one to be created instead of the existing updated. Is this an acceptable workaround?

[aharbis]

@cdjohnson Yes, that's the workaround that we documented in our upgrade docs. I still believe this is something that should be handled by OLM gracefully.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had any recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contribution.
For more help on your issue, check out the olm-dev channel on the kubernetes slack [1] and the OLM Dev Working Group [2] [1] https://kubernetes.slack.com/archives/C0181L6JYQ2 [2] https://github.com/operator-framework/community#operator-lifecycle-manager-wg

[stale]

This issue has been automatically closed because it has not had any recent activity. Thank you for your contribution.