Skip to content

Refine rollout worker health check and recovery lifecycle#1877

Open
YanhuiDua wants to merge 2 commits into
InternLM:mainfrom
YanhuiDua:fix-health-check-part3
Open

Refine rollout worker health check and recovery lifecycle#1877
YanhuiDua wants to merge 2 commits into
InternLM:mainfrom
YanhuiDua:fix-health-check-part3

Conversation

@YanhuiDua

@YanhuiDua YanhuiDua commented Jun 5, 2026

Copy link
Copy Markdown
Collaborator

Summary

Centralize rollout worker lifecycle and recovery in RolloutHealthChecker.

Changes

  • Move failed worker recovery out of RolloutController.
  • Track worker lifecycle with ACTIVE / INACTIVE / RECOVERING.
  • Add all-down recovery in health check to avoid generation hang.

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.

1 participant