Cluster upgrade causes Ingress VIP getting attached to two nodes at once
Issue
- After cluster upgrade from OpenShift v4.10 to v4.11 traffic is routed to
workernodes where defaultrouterpods is not present causingIngress,AuthenticationandConsolecluster operator going into degraded state withconnection refusederror. - After upgrading to OpenShift v4.11
Ingress VIPgets attached to multiple nodes at same time. - After an upgrade to v4.11, more than one node owns the API VIP or the kube-apiserver VIP
Environment
- Red Hat OpenShift Container Platform (RHOCP)
- 4.10
- 4.11
- 4.12
- 4.13
- IPI installation method
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.