| name | ocp-etcd |
| metadata | {"author":"rhoai3-demo","version":"1.0.0","platform-family":"ocp","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"OpenShift Platform"} |
| description | Use when documenting, reviewing, or rebuilding OpenShift Container Platform etcd guidance from official OCP documentation: operator-managed etcd behavior, control plane quorum, storage and latency practices, hardware validation, disk latency and consensus monitoring, defragmentation, moving etcd to a different disk, timer tuning, database size, network jitter and peer RTT, Kubernetes API transaction rate, backup and restore, unhealthy member replacement, disaster recovery, etcd encryption, and stretched-cluster caveats. Do NOT use for RHOAI component CRs; use rhoai-* skills. Do NOT use for ODF storage service behavior; use odf-* skills once created. Do NOT run live cluster operations without the env safety guard and explicit approval.
|
OCP etcd
Use this skill to ground OpenShift etcd documentation, runbooks, GitOps review
notes, and live-environment planning in the official OpenShift Container
Platform etcd documentation for the active baseline in
docs/PLATFORM_BASELINE.md.
Source Grounding
Read references/source-capture.md before using product behavior. Official Red
Hat documentation is product authority. This skill extracts supported etcd
operations and validation patterns; it does not replace the full official
procedure for destructive recovery work.
Safety Posture
Treat etcd as cluster-critical state:
- Prefer read-only health checks and documentation review before mutation.
- Take an official etcd backup before any maintenance, replacement, restore,
encryption, disk-move, or tuning change that can affect the control plane.
- Do not store etcd snapshots, static pod resources, encryption keys, or
kubeconfigs in this repository.
- Do not patch
unsupportedConfigOverrides unless following an official Red
Hat procedure and the user has explicitly approved the recovery action.
- Do not treat a cluster that spans sites as a disaster-recovery design.
- For live
oc or kubectl commands, use the repo environment guard and pair
this skill with the appropriate env-* skill.
OCP etcd Model
Use the official docs to frame etcd as the OpenShift control plane data store
managed by the OpenShift etcd Operator. The main responsibilities for this demo
are:
- keep etcd quorum and control plane health visible in operations docs
- avoid unsupported low-level etcd changes in GitOps manifests
- document backup, restore, and disaster-recovery boundaries clearly
- review latency, disk, storage, and stretched-cluster assumptions for AWS demo
environments
- ensure RHOAI and ODF planning respects etcd and API server performance limits
Workflow
- Confirm the active OpenShift baseline in
docs/PLATFORM_BASELINE.md.
- Read
references/official-doc-extraction.md.
- Decide whether the task is read-only guidance, planned maintenance, or
disaster recovery.
- For read-only guidance, use the validation commands in
references/validation-checklist.md.
- For any mutation or recovery action, capture:
- target cluster and API server guard status
- current
ClusterVersion and control plane operator status
- backup location and age
- official procedure URL and section name
- rollback or support-escalation plan
- For README or architecture content, describe the operational value without
teaching unsupported etcd internals.
- Validate the output with
references/validation-checklist.md.
Related Skills
- Use
project-red-hat-doc-skill-authoring to create additional ocp-*
skills from official OCP chapters.
- Use
project-gitops-authoring for demo GitOps manifests.
- Use
project-manifest-review for Kubernetes/OpenShift manifest review.
- Use
env-troubleshoot for live environment diagnostics.
- Use
env-deploy-and-evaluate or env-manage-resources for guarded live
deployment and resource-management workflows once active scripts exist.
References
references/source-capture.md
references/official-doc-extraction.md
references/validation-checklist.md
examples/etcd-operations-patterns.md