TECHNICAL GUIDE
Troubleshooting · TRB-03

Kubernetes/OpenShift Troubleshooting

Keep failed Kubernetes/OpenShift objects in place long enough to inspect them. Pod state, events, PVCs, and container logs usually identify the first rejected dependency.

DIAGNOSE

Preserve the rejected object

Do not immediately wipe the release. Inspect pod status, events, PVC state, previous logs, and the rendered chart. These usually separate image, SCC, storage, device, host-network, and STRATUM-startup failures.

CHECK

Escalate with context

When collecting evidence, include the failing pod description, relevant logs, PVC state, selected node, and values layers. Avoid sending only a final generic Helm error after the underlying pod has been deleted.

COMMANDS

Commands

Replace example node, registry, namespace, and endpoint values with the values for the target environment.

shell
oc get pods -n stratum -o wide
oc get pvc -n stratum
oc describe pod -n stratum stratum-stratum-controller-0
oc logs -n stratum stratum-stratum-controller-0 -c stratum --previous