Upgrade to OCP 4.14 using OVN-Kubernetes leaves the cluster severely degraded when networks overlap the 100.88.0.0/16 subnet
Issue
-
Planning to upgrade OpenShift cluster using
OVN-Kubernetes CNIplugin to 4.14, but upon checking the official documentation 1 and 2 it raises concerns since cluster is configured with network ranges that conflict with the newOVNtransit switchCIDR. -
Cluster upgraded and now noticing severe network performance issues and multiple issues in the cluster general functionality, with cluster operators flipping or stuck to degraded state.
-
Errors in OVNKube: illegal network configuration: transit switch subnet "100.88.0.0/16" overlaps service subnet "100.88.0.0/16" or similar messaging.
Environment
- Red Hat OpenShift Container Platform (RHOCP)
- 4.14
- OVN-Kubernetes CNI (OVN-K8S)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.