Liveness check should probe the storage backend for OpenShift
Issue
- registry is stuck in a non-usable state
- There's a liveness check in the registry, but it only checks the API, not the underlying storage, so it doesn't restart automatically
Environment
- OpenShift Container Platform 3.6
- integrated OpenShift registry is using glusterfs (CNS).
- after a loss of connection from the registry to the glusterfs storage with the connection to glusterfs is restored.
- Connecting to the pod reveals a broken mount point from the registry pod to the glusterfs.
- In DR scenarios when glusterfs nodes are moved to a different host, or moved back.
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
