Upgrade to OCP 4.6 is failing on vSphere IPI installation if master node has changed IP
Issue
- The upgrade gets stuck when updating the machine-config operator (
MCO
). - Master
machine config pool
(mcp
) is degraded. - Master node in
SchedulingDisabled
status after being restarted. - Master node has changed IP after being restarted.
-
Errors and warnings in etcd logs:
E | etcdserver: publish error: etcdserver: request timed out W | rafthttp: health check for peer xxxxxxxxxxxxxxxx could not connect: EOF
-
Errors in etcd-quorum-guard logs:
embed: rejected connection from "10.0.0.101:55555" (error "tls: \"10.0.0.101\" does not match any of DNSNames [\"localhost\" \"master-0.example.com\" \"10.0.0.110\"] xxxxx-master-0 was 10.0.0.110 is now 10.0.0.101
Environment
- Red Hat OpenShift Container Platform
- 4.5
- 4.6
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.