Skip to content

microvm: find a restored actor's CH api-socket without ateom's record - #1136

Open
Troy Chiu (troychiu) wants to merge 1 commit into
agent-substrate:mainfrom
troychiu:split/2-microvm-ch-socket
Open

microvm: find a restored actor's CH api-socket without ateom's record#1136
Troy Chiu (troychiu) wants to merge 1 commit into
agent-substrate:mainfrom
troychiu:split/2-microvm-ch-socket

Conversation

@troychiu

@troychiu Troy Chiu (troychiu) commented Aug 21, 2026

Copy link
Copy Markdown
Contributor

Towards #372

When ateom restarts, it loses in-memory actor records and falls back to socket lookup on disk. Restored actors listen on clh-api-restore.sock rather than clh-api.sock, causing terminateWorkload to miss the VMM (leaving orphan guests running) and CheckpointWorkload to time out.

This change:

  • Uses chSocketFor to probe disk for whichever API socket exists when no in-memory record is present.
  • Adds kata.RestoredCLHSocketPath alongside kata.CLHSocketPath for consistent socket naming across launch, restore, and teardown.

AI has assisted with this PR and I have verified all the changes.

  • Tests pass
  • Appropriate changes to documentation are included in the PR

@troychiu
Troy Chiu (troychiu) marked this pull request as ready for review August 24, 2026 06:58
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