ocp-nodes

name	ocp-nodes
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 node guidance from official OCP documentation: node and pod overview, containers, pod autoscaling, pod scheduling and placement, scheduler profiles, pod affinity and anti-affinity, node affinity, node selectors, taints and tolerations, topology spread constraints, descheduler, secondary scheduler, jobs, daemon sets, viewing nodes and pods on nodes, node labels, cordon/drain/unschedulable behavior, node deletion, node modification, control-plane schedulability, kernel arguments, swap, image pulls, maximum pods per node, Node Tuning Operator, node rebooting, garbage collection, node metrics, remote workers, single-node OpenShift worker behavior, and sigstore. Do NOT use for MachineSet or cloud instance provisioning; use future ocp-machine-management-aws. Do NOT use for RHOAI GPU hardware profiles or accelerator component behavior; use rhoai-nvidia-gpu-accelerators. Do NOT run live cluster operations without the env safety guard.

OCP Nodes

Use this skill to ground OpenShift node, pod placement, node operations, resource review, and node-health runbooks in the official OpenShift Container Platform Nodes guide 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 covers OCP node and scheduling behavior; it does not replace Machine API, MachineSet, Node Feature Discovery, GPU Operator, or RHOAI accelerator skills.

Demo Node Posture

For this AWS-hosted RHOAI demo:

• Discover node labels, taints, capacity, allocatable resources, and conditions from the target cluster before writing placement rules.
• Keep GPU node lifecycle and AWS instance provisioning in future ocp-machine-management-aws and existing rhoai-nvidia-gpu-accelerators handoffs.
• Use node selectors, node affinity, taints, tolerations, and topology spread intentionally; avoid over-constraining pods so they cannot schedule.
• Treat node drain, reboot, deletion, control-plane schedulability, kernel arguments, swap, maximum-pods-per-node, and Node Tuning Operator changes as live platform operations that require explicit approval.
• Keep RHOAI component resource requests and limits in rhoai-* skills; use this skill for the OpenShift scheduling and node mechanics around them.

OCP Node Model

Use the official docs to frame:

• Nodes: worker or control-plane machines that run pods and expose capacity, allocatable resources, labels, taints, conditions, and metrics.
• Pods: the unit scheduled onto nodes, with scheduling shaped by resource requests, affinity, selectors, tolerations, topology spread, and scheduler behavior.
• Containers: application runtime units inside pods, with operations such as logs, exec, file copy, and port forwarding.
• Scheduling controls: cluster scheduler profiles, pod affinity, node affinity, taints/tolerations, node selectors, topology spread constraints, descheduler, and secondary scheduler.
• Node operations: listing, inspecting, cordoning, draining, labeling, deleting, rebooting, garbage collection, and resource monitoring.

Workflow

1. Confirm the active OpenShift baseline in docs/PLATFORM_BASELINE.md.
2. Read references/official-doc-extraction.md.
3. Identify whether the task concerns:
   • node inventory and health
   • pod scheduling or placement
   • labels, taints, tolerations, or selectors
   • autoscaling, jobs, daemon sets, or pod lifecycle
   • node maintenance, drain, reboot, deletion, or tuning
   • node metrics and troubleshooting
4. For GitOps manifests, verify all labels, taints, scheduler names, and topology keys against the target cluster before committing placement rules.
5. For live operations, use the repo environment guard and pair this skill with env-troubleshoot, env-manage-resources, or env-deploy-and-evaluate.
6. Validate the output with references/validation-checklist.md.

Related Skills

• Use ocp-storage for node-related volume attachment and storage behavior.
• Use future ocp-machine-management-aws for AWS MachineSets, node creation, and instance lifecycle.
• Use ocp-node-feature-discovery for NFD-specific labels and specialized hardware detection.
• Use rhoai-nvidia-gpu-accelerators for NVIDIA GPU enablement and RHOAI hardware profile intent.
• Use rhoai-component-resource-customization for RHOAI component resource requests and limits.
• Use project-gitops-authoring and project-manifest-review for manifests.

References

• references/source-capture.md
• references/official-doc-extraction.md
• references/validation-checklist.md
• examples/node-placement-patterns.md