Skip to content

feat(supervisor): expose concurrent UDS auto-start - #905

Closed
kvinwang wants to merge 1 commit into
masterfrom
codex/feat-supervisor-trusted-uds-autostart
Closed

feat(supervisor): expose concurrent UDS auto-start#905
kvinwang wants to merge 1 commit into
masterfrom
codex/feat-supervisor-trusted-uds-autostart

Conversation

@kvinwang

@kvinwang kvinwang commented Jul 31, 2026

Copy link
Copy Markdown
Collaborator

Summary

Expose the Supervisor client's existing Unix-socket auto-start flow through the CLI and make concurrent starters converge on one daemon.

Changes

  • Add CLI options for auto-starting a missing Supervisor over a Unix socket.
  • Serialize concurrent starts with an owner-only lock and re-probe after acquiring it.
  • Start the Supervisor with umask 077, so its runtime files and socket are not group/world writable.
  • Keep structured command output on stdout by sending tracing output to stderr.
  • Return the shutdown response before notifying Rocket to stop.

This branch is rebased directly onto master. It does not include the client-side socket path validation from the closed PR #893; normal Unix socket and parent-directory permissions remain the access-control boundary.

Verification

  • cargo check --manifest-path dstack/Cargo.toml -p supervisor-client -p supervisor
  • cargo test --manifest-path dstack/Cargo.toml -p supervisor-client -p supervisor
  • cargo fmt --manifest-path dstack/Cargo.toml --all
  • git diff --check

Copilot AI review requested due to automatic review settings July 31, 2026 03:02

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

@kvinwang
kvinwang force-pushed the codex/feat-supervisor-trusted-uds-autostart branch from 09084ca to 4e746f6 Compare August 4, 2026 09:10
@kvinwang kvinwang changed the title [STACKED on #893] feat(supervisor): expose trusted UDS auto-start feat(supervisor): expose concurrent UDS auto-start Aug 4, 2026
@kvinwang
kvinwang changed the base branch from codex/fix-supervisor-trusted-client-sockets to master August 4, 2026 09:10
@kvinwang

kvinwang commented Aug 4, 2026

Copy link
Copy Markdown
Collaborator Author

Closing the concurrent UDS auto-start feature because the supported deployment has one VMM owner per runtime directory. The two independent fixes were split into #994 (machine-readable client output) and #995 (graceful shutdown response). PR #841 was updated to test the client lifecycle directly without requiring CLI auto-start or concurrent-starter behavior.

@kvinwang kvinwang closed this Aug 4, 2026
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.

2 participants