+++ This bug was initially created as a clone of Bug #1883461 +++ k8s.io/klog moved to v2 in client-go and frieds. The operator does not use v2 yet and hence, we lose one half of the logging output (v2). --- Additional comment from Stefan Schimanski on 2020-09-29 11:09:06 UTC --- I did some tests using openshift-apiserver that uses kube component-base to add klog command line flags: $ ./openshift-apiserver start --config asfd --v=4 I0929 13:04:18.354425 53628 cmd.go:57] v1 klog I0929 13:04:18.354506 53628 cmd.go:58] v2 klog I0929 13:04:18.354512 53628 cmd.go:61] v2 V(2) klog I0929 13:04:18.354515 53628 cmd.go:64] v2 V(4) klog So this means that we DO NOT set verbosity for v1 klog, and hence lose all the klogv1.V(2).Infof output.
We didn't have ability to change log levels before, so it's not a blocker for 4.6.
A PR already exists for this one, awaiting review. Other than that, we still need to wait for 4.7 cycle to start.
Thanks Ricardo and Oleg! Verified on 4.7.0-0.nightly-2020-10-24-155529: 1. With a GCP cluster, change spec.operatorlevel=TraceAll; 2. Set the image registry to Removed 3. Wait for some seconds, log like below will appear: $ oc logs pods/cluster-image-registry-operator-5b6f5ff8b-rwwgp | grep deleting I1026 07:32:48.770962 1 gcs.go:327] deleting all objects in bucket wzheng-47-2tfrc-image-registry-us-central1-fcuhhqklefmcoryijnt
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2020:5633