Troubleshooting And Diagnostics practice questions

oc adm upgrade resume

Why wrong: CVO upgrades are managed via clusterversion, not machineconfigpools directly with this command.

oc get mcp --show-paused

Why wrong: There is no --show-paused flag for oc get mcp.

oc resume machineconfigpool/worker

Why wrong: 'oc resume' is not a valid OpenShift CLI verb.

The ingress router certificate expiration.

Why wrong: Certificates affect TLS endpoints, not local pod volume mount permissions.

oc get clusteroperator etcd

Why wrong: Etcd operator status is unrelated to HTTP application route load balancing.

oc get events -n openshift-config-managed

Why wrong: Cluster events show state changes but lack the deep HTTP client stack trace for OAuth token validation.

oc get clusteroperator authentication -o yaml

Why wrong: This only displays the status conditions and does not provide the underlying error logs.

oc adm must-gather --image=quay.io/openshift-release-dev/ocp-v4.0-art-dev@sha256:...

Why wrong: Must-gather is for broad data collection, not the targeted real-time log inspection needed for specific routing validation failures.

oc top pod <pod-name> --containers

Why wrong: oc top shows current real-time usage, but a crashed pod will return an error since it is no longer running.

oc get pod <pod-name> -o yaml

Why wrong: YAML output shows static configuration but does not summarize exit reasons and last termination state cleanly.

oc adm diagnostics

Why wrong: adm diagnostics is a general cluster health check tool, not for querying specific pod OOM states.

oc describe machineconfigpool worker

Why wrong: Describing the pool shows high-level status conditions and node counts, but not the detailed MCD failure script output.

oc logs -n openshift-cluster-version-operator deployment/cluster-version-operator

Why wrong: The CVO coordinates cluster upgrades as a whole, but individual node file deployment is handled by the MCO.

journalctl -u kubelet -e

Why wrong: The kubelet manages container lifecycles, not the specific OS-level MachineConfig file applications handled by MCO.

oc get events --all-namespaces

Why wrong: API rate-limiting is handled at the transport layer and logged internally in the API server, rarely creating standard Kubernetes events.

oc logs -n openshift-ingress-operator deployment/ingress-operator

Why wrong: The ingress operator manages router deployments, not internal Kubernetes API request traffic.

oc logs -n openshift-kube-apiserver-operator deployment/cluster-kube-apiserver-operator

Why wrong: This operator manages the API server lifecycle, not the API server traffic requests themselves.

oc logs pod/my-pod backend

Why wrong: The syntax requires the -c flag before the container name.

oc rsh pod/my-pod backend

Why wrong: oc rsh opens a shell and does not output container logs.

oc describe pod/my-pod -c backend

Why wrong: oc describe does not take a container flag to filter logs.

oc get events -n openshift-storage

Why wrong: While events can show errors, the detailed CSI driver and attacher plugin logs are necessary for deep storage troubleshooting.

oc debug node/worker-0

Why wrong: Debugging the node is premature before checking the storage controller events and PVC description.

oc logs -n openshift-cluster-storage-operator deployment/csi-snapshot-controller-operator

Why wrong: This operator manages snapshots, not the volume attachment and mounting operations.

oc logs pod/my-app-6b79f-abcde --previous

Why wrong: oc logs --previous shows standard output from the previous container instance but does not allow file system access.

oc rsh deployment/my-app

Why wrong: oc rsh requires a running container to attach to; a crashing container cannot be accessed via rsh.

oc exec -it pod/my-app-6b79f-abcde -- /bin/sh

Why wrong: oc exec also requires the container to be running and healthy to execute commands.

oc adm diagnose

Why wrong: This command is deprecated and not used for comprehensive data collection.

oc cluster-info dump

Why wrong: This is an upstream Kubernetes command and does not collect all OpenShift-specific operator data.

oc get clusteroperators -o yaml

Why wrong: This only outputs cluster operator resources and does not gather pod logs or node diagnostics.

oc get routes

Why wrong: Routes deal with external ingress traffic, not internal container probes.

oc get pod <pod-name> -o json

Why wrong: JSON output contains raw specifications and status trees that are less readable than describe for probe errors.

oc adm diagnostics

Why wrong: Diagnostics checks cluster health, not individual application probe failures.

oc adm upgrade status

Why wrong: There is no 'adm upgrade status' subcommand in OpenShift; 'oc get clusterversion' is used.

oc get clusteroperators

Why wrong: This lists operators, but does not provide the specific CVO payload target version details.

oc version

Why wrong: oc version shows the client and server software versions but not the upgrade status or CVO progress.

/rootfs/var/log

Why wrong: /rootfs is not the standard mount point used by OpenShift's oc debug node implementation; /host is.

/mnt/host/log

Why wrong: This is an incorrect directory path for the host mount.

/var/log

Why wrong: /var/log inside the debug pod points to the debug container's own log directory, not the host.

oc adm inspect etcd

Why wrong: oc adm inspect is not a valid OpenShift command for etcd performance testing.

oc adm diagnostics etcd-latency

Why wrong: No such diagnostic subcommand exists for etcd latency.

oc get etcd -o yaml

Why wrong: The custom resource definition for etcd does not contain runtime peer latency numbers.

oc watch deployment/web-frontend

Why wrong: oc watch monitors resource metadata changes, not container output logs.

oc logs pods -l app=web-frontend

Why wrong: This command is syntactically invalid; label selectors require the --selector flag.

oc rsh deployment/web-frontend

Why wrong: oc rsh cannot attach to a deployment directly.

The liveness probe timed out before the application started.

Why wrong: Probe failures occur after the container has been created and started.

The container image pull policy is set to Always and the registry is down.

Why wrong: This causes an ErrImagePull or ImagePullBackOff error, not a config error.

The container exceeded its memory limit upon initialization.

Why wrong: This results in an OOMKilled state during runtime or startup, not a configuration error.

oc adm diagnostics ovs

Why wrong: No such diagnostic tool exists for OVS flows in OpenShift.

oc exec daemonset/ovn-controller -- ovs-ofctl dump-flows

Why wrong: OVN in OpenShift 4 uses OVN-Kubernetes, and the command syntax or pod names do not match this structure.

oc rsh daemonset/ovnkube-node

Why wrong: ovnkube-node runs in a container that may not have direct access to host OVS utilities without proper namespace sharing.

oc adm events --all

Why wrong: There is no 'oc adm events' subcommand.

oc show events --recent

Why wrong: 'show' is not a valid oc verb.

oc cluster-events

Why wrong: 'cluster-events' is not a native oc command.

Inspect the etcd cluster operator status.

Why wrong: Etcd has no bearing on pulling container images from container registries.

Run 'oc debug' to inspect the container file system.

Why wrong: The container has not been created yet because the image cannot be pulled, making debugging useless.

Restart the cluster version operator.

Why wrong: CVO manages cluster upgrades, not application container image pulling.