Description of problem: 1. The cluster-etcd-operator introduced an init container that waits service account(namespace, token, ca.crt, service-ca.crt) related files to be synced to the static pod. The CI runs have indicated takes a long time for service-ca.crt to be synced, even after kube is up. 2. The cluster-etcd-operator also introduced a new wait-for-ceo[1] command in the installer, that blocks the installer from destroying the bootstrap node before etcd has scaled to all 4 masters. The wait-for-ceo command also needs to wait for kube apiserver to roll out valid config. The loop checking this config errors out without checking all the config. It should collect the errors and attempt to check the config for all kube-apiserver pods. [1] https://github.com/openshift/installer/blob/993c8cec1e8b1dbf58383fa52b264695187656be/data/data/bootstrap/files/usr/local/bin/bootkube.sh.template#L403 Version-Release number of selected component (if applicable):
This issue seems still be hit in installation randomly, https://bugzilla.redhat.com/show_bug.cgi?id=1798945, we may trace this issue with one bug, so close this one, pls correct me if there is misunderstanding.
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, 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/RHBA-2020:0581