This is a clone of issue OCPBUGS-17808. The following is the description of the original issue:
—
This is a clone of issue OCPBUGS-16804. The following is the description of the original issue:
—
This is a clone of issue OCPBUGS-15327. The following is the description of the original issue:
—
Description of problem:

On OpenShift Container Platform, the etcd Pod is showing messages like the following:

2023-06-19T09:10:30.817918145Z {"level":"warn","ts":"2023-06-19T09:10:30.817Z","caller":"fileutil/purge.go:72","msg":"failed to lock file","path":"/var/lib/etcd/member/wal/000000000000bc4b-00000000183620a4.wal","error":"fileutil: file already locked"}


This is described in KCS https://access.redhat.com/solutions/7000327

Version-Release number of selected component (if applicable):

any currently supported version (> 4.10) running with 3.5.x

How reproducible:

always

Steps to Reproduce:

happens after running etcd for a while

This has been discussed in https://github.com/etcd-io/etcd/issues/15360

It's not a harmful error message, it merely indicates that some WALs have not been included in snapshots yet.

This was caused by changing default numbers: https://github.com/etcd-io/etcd/issues/13889

This was fixed in https://github.com/etcd-io/etcd/pull/15408/files but never backported to 3.5.

To mitigate that error and stop confusing people, we should also supply that argument when starting etcd in: https://github.com/openshift/cluster-etcd-operator/blob/master/bindata/etcd/pod.yaml#L170-L187

That way we're not surprised by changes of the default values upstream.