Skip to content

DX-596-together-gpu-clusters: sync with mintlify-docs#887#18

Merged
zainhas merged 1 commit into
mainfrom
docs-sync/together-gpu-clusters/mintlify-docs-pr-887
May 27, 2026
Merged

DX-596-together-gpu-clusters: sync with mintlify-docs#887#18
zainhas merged 1 commit into
mainfrom
docs-sync/together-gpu-clusters/mintlify-docs-pr-887

Conversation

@zainhas

@zainhas zainhas commented May 27, 2026

Copy link
Copy Markdown
Collaborator

Syncs the together-gpu-clusters skill with the new Slurm startup scripts page merged in togethercomputer/mintlify-docs#887 ("[TCl-5106]docs: add Slurm startup scripts page", merge sha 2c0ecf2).

Changed docs files that triggered this sync

  • docs/slurm-startup-scripts.mdx (new) — matched via together-gpu-clusters glob docs/slurm*.mdx

Skill changes

  • references/cluster-management.md: added a new Startup Scripts (Slinky v1.0 only) subsection under Slurm Configuration, with a script-type table (worker init, login init, worker prolog, worker epilog, controller prolog, controller epilog, Extra slurm.conf), failure-mode bullets, the drained-node resume command, and operational rules.
  • references/cluster-management.md: added a TOC entry for the new subsection.
  • SKILL.md: added one High-Signal Rule covering Slinky-v1.0 scope, the DRAIN failure mode on worker prolog/epilog non-zero exit, and the prolog/epilog -> Slurm-command deadlock risk.
  • SKILL.md: added the new Slurm Startup Scripts page to Official Docs.

Validated locally with scripts/quick_validate.py and scripts/quality_check.py — both pass.

Generated by the Sync Skills Cursor Automation. Please review before merging.

Add Slurm Slinky v1.0 startup scripts coverage:

- references/cluster-management.md: new "Startup Scripts (Slinky v1.0 only)"
  subsection under Slurm Configuration covering worker/login init scripts,
  worker/controller prolog and epilog, Extra slurm.conf, failure modes
  (DRAIN, requeue, cancel), and rules (no Slurm calls inside prolog/epilog,
  PrologFlags=Alloc, configless caching, set -e / SLURM_JOB_ID).
- SKILL.md: new High-Signal Rule summarizing Slinky v1.0 scope and the
  DRAIN/deadlock failure modes; added Slurm Startup Scripts link to
  Official Docs.

Generated by the Sync Skills Cursor Automation. Please review before merging.
@zainhas zainhas merged commit 6cf512b into main May 27, 2026
2 checks passed
@zainhas zainhas deleted the docs-sync/together-gpu-clusters/mintlify-docs-pr-887 branch May 27, 2026 02:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants