+++ This bug was initially created as a clone of Bug #1804693 +++ we skipped a file check on etcd-member, but we can remove that check now that we have been able to remove the file overall
We are not going to port these changes into 4.4 to support upgrade from 4.3 to 4.4 (in which case the etcd-member may still be present). We need to close this bug as NOTABUG.
Upgrade from 4.3 was failing if we do not skip file checks: ``` E0311 22:17:22.214558 155963 writer.go:135] Marking Degraded due to: unexpected on-disk state validating against rendered-master-b4ac8feed03200284df2a016c62c65a1 I0311 22:18:22.226749 155963 daemon.go:724] Current+desired config: rendered-master-b4ac8feed03200284df2a016c62c65a1 I0311 22:18:22.235481 155963 daemon.go:958] Validating against current config rendered-master-b4ac8feed03200284df2a016c62c65a1 E0311 22:18:22.243624 155963 daemon.go:1337] could not stat file: "/etc/kubernetes/manifests/etcd-member.yaml", error: lstat /etc/kubernetes/manifests/etcd-member.yaml: no such file or directory ```
Reopening this bug based on https://github.com/openshift/machine-config-operator/pull/1562 I just hit it on an upgrade in 4.3 and that PR would improve the upgrade behaviour.
> Why this Bugzilla has been closed as I see the PR hasnt merged yet : Excuse the confusion the PR was not what we wanted as per[1] I just closed it. > I just hit it on an upgrade in 4.3 and that PR would improve the upgrade behaviour. Can you clarify what exactly you hit as per the details in [1] this PR [2] would actually increase failures. I am closing this bug as it was directly related to the backport which is not what we want.If you have a 4.3 -> 4.4 upgrade issue please file a new BZ with details of the failure. [1] https://github.com/openshift/machine-config-operator/pull/1562#issuecomment-694185100 [2] https://github.com/openshift/machine-config-operator/pull/1562