Openshift forced upgrade in disconnected environment can bring cluster to non-operating state
Issue
During a forced upgrade of OpenShift Container Platform in disconnected environment, a race condition can happen which can lead cluster to non-operating state. In the worst case a cluster can be bricked, so one would be having to turn the cluster version operator off entirely, and push manifests one at a time on its behalf to slowly unwind any components that had been too tangled up.
Environment
- Red Hat OpenShift Container Platform 4.x
- Minor versions prior 4.3.12 and 4.4.0
- Restricted (disconnected) network environment
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.