docs: drain the node before changing the operating mode

[moezdil]

What type of PR is this?

/kind documentation

What this PR does / why we need it:

The docs only said a mode change needs a plugin restart. Switching between cu and partition with GPU Pods running leaves them under the old slot count and CU layout (reproduced on an MI325X VF: new Pods failed admission or were mis-scheduled before #159). This adds the drain, change, restart, verify and uncordon steps, mentions the AMDGPUModeChangedInUse event from #160, and warns that upgrading from a partition-mode build older than #159 also needs a drained node.

Which issue(s) this PR fixes:

Follow-up to #17, #159 and #160.

Special notes for your reviewer:

Docs only. The steps match what was run on the MI325X VF in the end-to-end mode-switch tests.

Does this PR introduce a user-facing change?:

No.

AI Disclosure:

AI-assisted, reviewed by the author.

[hami-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: moezdil

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: dbd28edb-c699-4869-8558-11fb83b58636

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}