Created attachment 1729751 [details]
List of deployments during failure

Description of problem:

Tests below do not log enough information to make it clear why the ephemeral image registry deployment is failing:

[sig-imageregistry][Feature:ImageStreamImport][Serial][Slow] ImageStream API TestImportRepositoryFromBlockedRegistry expand_more	5m20s
[sig-imageregistry][Feature:ImageStreamImport][Serial][Slow] ImageStream API TestImportImageFromBlockedRegistry expand_more	5m20s
[sig-imageregistry][Feature:ImageStreamImport][Serial][Slow] ImageStream API TestImportImageFromInsecureRegistry expand_more	5m20s
[sig-imageregistry][Feature:ImageStreamImport][Serial][Slow] ImageStream API TestImportRepositoryFromInsecureRegistry	5m20s

By looking at the gather-extra information I could see that the deployments did not contain anything on their statuses. By increasing the log verbosity we can at least see what was on the deployment status and how long it is taking to succeed.


How reproducible:

Not able to tell how often it does happen, but you might find useful info through https://testgrid.k8s.io/redhat-openshift-ocp-release-4.7-blocking#release-openshift-ocp-installer-e2e-aws-4.7


Expected results:

On failure we can see where exactly it failed and what was registered in the deployment statuses (conditions, available replicas, etc).


Additional info:

This might be an outcome of recent docker.io limiting. Follow attached here the list of deployments during the failure, we can see that there is no "status" on ephemeral registry deployments (we could see the deployments on gather-extra just by luck as the namespaces were already flagged to be deleted).