Question 340A Pod fails to mount an otherwise Bound PVC because its volumeMode is Block while the application expects a filesystem path. Which Kubernetes administration action most directly addresses the observed failure? Inspect VolumeAttachment objects, CSI controller and node logs, then clear a genuinely stale attachment using the documented driver process. Match the claim's accessModes, storage class and capacity to the intended static PV. Use a filesystem-mode volume or configure the consumer for a raw block device consistently. Perform the approved reclaim and rebind procedure, including handling stale claimRef where safe. ExplanationCorrect Answer C:Use a filesystem-mode volume or configure the consumer for a raw block device consistently. This corrects the specific failure because volumeMode distinguishes raw Block devices from mounted Filesystem volumes. Verify it by checking: The Pod sees the expected filesystem path or explicitly mapped blockDevice as designed.Incorrect Answers A:A Bound PVC proves allocation but not that attach, stage and publish operations have succeeded. That is relevant to a different intervention. For the stated situation, the required action is: Use a filesystem-mode volume or configure the consumer for a raw block device consistently. The governing reason is: volumeMode distinguishes raw Block devices from mounted Filesystem volumes. B:A PVC must satisfy PV binding predicates, not merely request a compatible size. That is relevant to a different intervention. For the stated situation, the required action is: Use a filesystem-mode volume or configure the consumer for a raw block device consistently. The governing reason is: volumeMode distinguishes raw Block devices from mounted Filesystem volumes. D:A Released PV may preserve its previous claim reference until an administrator makes it available. That is relevant to a different intervention. For the stated situation, the required action is: Use a filesystem-mode volume or configure the consumer for a raw block device consistently. The governing reason is: volumeMode distinguishes raw Block devices from mounted Filesystem volumes.
Question 341A team deletes a PVC expecting to preserve its disk, but the dynamically created PV uses a Delete reclaim policy. An engineer recommends: Set the appropriate Retain policy before deleting the claim and plan explicit data recovery or disposal. Which Kubernetes behavior best explains why this correction fits? Mount-level read-only settings can prohibit writes regardless of a backend's broader capabilities. emptyDir follows the Pod lifetime and is removed when that Pod ceases to exist. Reclaim policy determines what happens to backend storage after a bound claim is released. ReadWriteOnce restricts volume mounting to one node, not globally one Pod; multi-node sharing needs backend support. ExplanationCorrect Answer C:Reclaim policy determines what happens to backend storage after a bound claim is released. This explains why the proposed action—set the appropriate Retain policy before deleting the claim and plan explicit data recovery or disposal.—is appropriate. Direct confirmation is: After PVC deletion, the retained PV and backend volume still exist for approved recovery.Incorrect Answers A:Mount-level read-only settings can prohibit writes regardless of a backend's broader capabilities. This does not establish the proposed intervention. The governing principle here is: Reclaim policy determines what happens to backend storage after a bound claim is released. Apply it by: Set the appropriate Retain policy before deleting the claim and plan explicit data recovery or disposal. B:emptyDir follows the Pod lifetime and is removed when that Pod ceases to exist. This does not establish the proposed intervention. The governing principle here is: Reclaim policy determines what happens to backend storage after a bound claim is released. Apply it by: Set the appropriate Retain policy before deleting the claim and plan explicit data recovery or disposal. D:ReadWriteOnce restricts volume mounting to one node, not globally one Pod; multi-node sharing needs backend support. This does not establish the proposed intervention. The governing principle here is: Reclaim policy determines what happens to backend storage after a bound claim is released. Apply it by: Set the appropriate Retain policy before deleting the claim and plan explicit data recovery or disposal.
Question 342An application serves traffic before its cache has warmed and early requests fail during every restart. An engineer recommends: Add a readiness probe that becomes successful only when the application can serve requests. Which Kubernetes behavior best explains why this correction fits? Liveness failures restart a container; readiness failures alone only affect traffic eligibility. Startup probes suppress other probe evaluation until the startup check succeeds. A failed readiness probe removes Pod endpoints from Services without necessarily restarting the container. Requests guide scheduling while limits constrain runtime resource consumption. ExplanationCorrect Answer C:A failed readiness probe removes Pod endpoints from Services without necessarily restarting the container. This explains why the proposed action—add a readiness probe that becomes successful only when the application can serve requests.—is appropriate. Direct confirmation is: The new Pod is excluded from ready endpoints until its cache is usable.Incorrect Answers A:Liveness failures restart a container; readiness failures alone only affect traffic eligibility. This does not establish the proposed intervention. The governing principle here is: A failed readiness probe removes Pod endpoints from Services without necessarily restarting the container. Apply it by: Add a readiness probe that becomes successful only when the application can serve requests. B:Startup probes suppress other probe evaluation until the startup check succeeds. This does not establish the proposed intervention. The governing principle here is: A failed readiness probe removes Pod endpoints from Services without necessarily restarting the container. Apply it by: Add a readiness probe that becomes successful only when the application can serve requests. D:Requests guide scheduling while limits constrain runtime resource consumption. This does not establish the proposed intervention. The governing principle here is: A failed readiness probe removes Pod endpoints from Services without necessarily restarting the container. Apply it by: Add a readiness probe that becomes successful only when the application can serve requests.
Question 343An API endpoint appears slow but an administrator looks only at node CPU and never measures its container resource pressure. Following the proposed change—Inspect Pod-level CPU and memory use and component metrics with kubectl top and relevant monitoring—which observation best demonstrates that the particular fault is resolved? Throttle counters decline and application latency improves after a justified limit adjustment. The specific bottlenecking Pod is identified and its latency improves after targeted resource changes. Per-Pod and node metrics identify the actual source of growth and targeted action resolves pressure. Adjusted requests or added capacity make the Pods schedulable without overcommit beyond policy. ExplanationCorrect Answer B:The specific bottlenecking Pod is identified and its latency improves after targeted resource changes. This validates the stated intervention—inspect Pod-level CPU and memory use and component metrics with kubectl top and relevant monitoring.—rather than an unrelated cluster change. The relevant mechanism is: Node-wide utilization can hide a throttled or memory-constrained application container.Incorrect Answers A:CPU limits may cause throttling independent of overall node idle time. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The specific bottlenecking Pod is identified and its latency improves after targeted resource changes. C:Node memory usage includes multiple consumers and cannot be attributed to one Pod without evidence. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The specific bottlenecking Pod is identified and its latency improves after targeted resource changes. D:Scheduling uses resource requests and constraints, not recent utilization percentages alone. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The specific bottlenecking Pod is identified and its latency improves after targeted resource changes.
Question 344A cluster operator assumes a CSI plugin handles Pod networking because it is deployed in the kube-system namespace. Following the proposed change—Verify the installed CNI and CSI components separately and test each extension boundary—which observation best demonstrates that the particular fault is resolved? CSIDriver and plugin Pods are healthy and a test PVC mounts successfully. Network connectivity and volume mount fixtures independently pass for their respective plugins. Representative policy and cross-node traffic tests pass under the new plugin before retirement of the old one. crictl info succeeds against the configured socket and a test workload starts. ExplanationCorrect Answer B:Network connectivity and volume mount fixtures independently pass for their respective plugins. This validates the stated intervention—verify the installed CNI and CSI components separately and test each extension boundary.—rather than an unrelated cluster change. The relevant mechanism is: CNI handles networking while CSI handles storage; shared namespaces do not imply interchangeable capabilities.Incorrect Answers A:CSI driver registration and controller/node capabilities are required for external volume provisioning and mounting. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Network connectivity and volume mount fixtures independently pass for their respective plugins. C:Network policy implementations and CNI-specific resources can differ despite common Pod-network requirements. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Network connectivity and volume mount fixtures independently pass for their respective plugins. D:The container runtime interface is the kubelet's integration boundary for Pod sandboxes and containers. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Network connectivity and volume mount fixtures independently pass for their respective plugins.
Question 345A security review finds a wildcard RBAC rule allowing all verbs on all API groups for a log collector. What is the most direct consequence if the underlying problem is left unresolved? Legitimate maintenance remains blocked or privileges are broadened unnecessarily. Broad fallback permissions can expose credentials beyond the job's purpose. Emergency diagnostics are blocked or an overly broad role exposes workloads. A compromised collector may manipulate arbitrary cluster resources. ExplanationCorrect Answer D:A compromised collector may manipulate arbitrary cluster resources. This directly follows from the failed condition. The preventive correction is: Replace wildcard rules with explicit get list watch permissions on the exact resources the collector reads. The governing Kubernetes behavior is: Least-privilege RBAC narrows blast radius by limiting the allowed verbs, resources, and API groups.Incorrect Answers A:Authentication establishes identity while authorization separately decides permitted API verbs and resources. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: A compromised collector may manipulate arbitrary cluster resources. B:Kubernetes service accounts do not automatically inherit administrator permissions from the Pod owner. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: A compromised collector may manipulate arbitrary cluster resources. C:RBAC treats interactive Pod subresources separately from operations on the parent Pod resource. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: A compromised collector may manipulate arbitrary cluster resources.
Question 346An operator installation is removed but its CRDs and custom instances remain with active finalizers. Which Kubernetes administration action most directly addresses the observed failure? Review finalizer ownership and perform a supported cleanup sequence before removing the responsible controller. Restore the controller/operator and check reconciliation logs and permissions before changing the custom object. Grant the controller the minimum required update or patch permission for the status subresource. Plan served and storage-version conversion carefully and test existing object reads before changing the CRD. ExplanationCorrect Answer A:Review finalizer ownership and perform a supported cleanup sequence before removing the responsible controller. This corrects the specific failure because finalizers block object deletion until the associated controller completes the required cleanup. Verify it by checking: Custom instances delete cleanly after their finalizers are processed and dependent resources are reviewed.Incorrect Answers B:A CRD defines the API schema while a running controller implements the desired behavior. That is relevant to a different intervention. For the stated situation, the required action is: Review finalizer ownership and perform a supported cleanup sequence before removing the responsible controller. The governing reason is: Finalizers block object deletion until the associated controller completes the required cleanup. C:Custom resource status updates are separately authorized from spec updates when status is enabled. That is relevant to a different intervention. For the stated situation, the required action is: Review finalizer ownership and perform a supported cleanup sequence before removing the responsible controller. The governing reason is: Finalizers block object deletion until the associated controller completes the required cleanup. D:CRD versioning separates served API versions from the storage version and may require conversion webhooks. That is relevant to a different intervention. For the stated situation, the required action is: Review finalizer ownership and perform a supported cleanup sequence before removing the responsible controller. The governing reason is: Finalizers block object deletion until the associated controller completes the required cleanup.
Question 347A planned kubeadm control-plane upgrade is attempted without first taking a recoverable etcd snapshot. Following the proposed change—Take, securely store, and verify an etcd backup as part of the documented upgrade rollback plan—which observation best demonstrates that the particular fault is resolved? The backup has the expected revision and an isolated restore rehearsal can list known objects. The upgrade plan lists the expected target and preflight checks pass before apply. kubectl version and node reports confirm supported version relationships after the planned sequence. kubectl get nodes shows Ready without SchedulingDisabled and a test Pod can land there. ExplanationCorrect Answer A:The backup has the expected revision and an isolated restore rehearsal can list known objects. This validates the stated intervention—take, securely store, and verify an etcd backup as part of the documented upgrade rollback plan.—rather than an unrelated cluster change. The relevant mechanism is: Cluster state durability requires more than rerunning kubeadm; an etcd snapshot protects API objects.Incorrect Answers B:kubeadm upgrade plan surfaces supported upgrade paths and potential blockers, not just an install command. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The backup has the expected revision and an isolated restore rehearsal can list known objects. C:Component version skew is constrained; a newer kubelet must not be assumed compatible with an older API server. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The backup has the expected revision and an isolated restore rehearsal can list known objects. D:Cordon status prevents new scheduling even when the node is otherwise Ready. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The backup has the expected revision and an isolated restore rehearsal can list known objects.
Question 348A PVC remains Pending although a statically created PV exists with the same capacity because access modes differ. Following the proposed change—Match the claim's accessModes, storage class and capacity to the intended static PV—which observation best demonstrates that the particular fault is resolved? The Pod's namespace contains the referenced claim and volume mounts successfully. The PV is Available or intentionally bound to the new approved PVC without losing data. The PVC binds the expected PV and a consumer Pod mounts it successfully. The storage driver reports successful attachment to the replacement node and the Pod mounts correctly. ExplanationCorrect Answer C:The PVC binds the expected PV and a consumer Pod mounts it successfully. This validates the stated intervention—match the claim's accessModes, storage class and capacity to the intended static PV.—rather than an unrelated cluster change. The relevant mechanism is: A PVC must satisfy PV binding predicates, not merely request a compatible size.Incorrect Answers A:PVCs are namespace-scoped and Pods cannot mount another namespace's claim by name. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The PVC binds the expected PV and a consumer Pod mounts it successfully. B:A Released PV may preserve its previous claim reference until an administrator makes it available. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The PVC binds the expected PV and a consumer Pod mounts it successfully. D:A Bound PVC proves allocation but not that attach, stage and publish operations have succeeded. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The PVC binds the expected PV and a consumer Pod mounts it successfully.
Question 349An HPA oscillates between minimum and maximum replicas, and the team only watches its current desired count. An engineer recommends: Inspect HPA events, metric availability and scaling behavior over time before tuning thresholds. Which Kubernetes behavior best explains why this correction fits? OOMKilled is evidence of memory-limit or node memory pressure, not an ordinary successful exit. Image storage pressure and image garbage collection are distinct from application CPU demand. Scheduling uses resource requests and constraints, not recent utilization percentages alone. Autoscaling dynamics depend on metric sampling, requests and stabilization parameters. ExplanationCorrect Answer D:Autoscaling dynamics depend on metric sampling, requests and stabilization parameters. This explains why the proposed action—inspect HPA events, metric availability and scaling behavior over time before tuning thresholds.—is appropriate. Direct confirmation is: A controlled load pattern produces bounded replica changes without repeated rapid oscillation.Incorrect Answers A:OOMKilled is evidence of memory-limit or node memory pressure, not an ordinary successful exit. This does not establish the proposed intervention. The governing principle here is: Autoscaling dynamics depend on metric sampling, requests and stabilization parameters. Apply it by: Inspect HPA events, metric availability and scaling behavior over time before tuning thresholds. B:Image storage pressure and image garbage collection are distinct from application CPU demand. This does not establish the proposed intervention. The governing principle here is: Autoscaling dynamics depend on metric sampling, requests and stabilization parameters. Apply it by: Inspect HPA events, metric availability and scaling behavior over time before tuning thresholds. C:Scheduling uses resource requests and constraints, not recent utilization percentages alone. This does not establish the proposed intervention. The governing principle here is: Autoscaling dynamics depend on metric sampling, requests and stabilization parameters. Apply it by: Inspect HPA events, metric availability and scaling behavior over time before tuning thresholds.
Question 350A tenant can edit operator configuration custom resources outside its namespace due to a broad custom API RoleBinding. Which Kubernetes administration action most directly addresses the observed failure? Install the CRD and wait for its Established condition before applying dependent custom resources. Restore the controller/operator and check reconciliation logs and permissions before changing the custom object. Restrict custom resource group, kind and namespace verbs using scoped RBAC policies. Grant the controller the minimum required update or patch permission for the status subresource. ExplanationCorrect Answer C:Restrict custom resource group, kind and namespace verbs using scoped RBAC policies. This corrects the specific failure because new custom resource types participate in authorization using their API group and scope. Verify it by checking: Impersonated edits fail outside the permitted namespace while intended local access remains.Incorrect Answers A:The API server must register an extension type before it can accept objects of that kind. That is relevant to a different intervention. For the stated situation, the required action is: Restrict custom resource group, kind and namespace verbs using scoped RBAC policies. The governing reason is: New custom resource types participate in authorization using their API group and scope. B:A CRD defines the API schema while a running controller implements the desired behavior. That is relevant to a different intervention. For the stated situation, the required action is: Restrict custom resource group, kind and namespace verbs using scoped RBAC policies. The governing reason is: New custom resource types participate in authorization using their API group and scope. D:Custom resource status updates are separately authorized from spec updates when status is enabled. That is relevant to a different intervention. For the stated situation, the required action is: Restrict custom resource group, kind and namespace verbs using scoped RBAC policies. The governing reason is: New custom resource types participate in authorization using their API group and scope.