Lab runs connect a restarted CMOnet again #149

Manually merged

vinny merged 0 commits from vinny/restart-reconnect into master 2026-10-17 11:16:32 -04:00 AGit

Conversation 0 Commits - Files changed -

vinny commented 2026-10-17 11:15:40 -04:00

Owner

Copy link

Summary

• The lab runner's restart_cmonet fault now connects the restarted CMOnet to CMO once the new process answers (connection.state not_initialized, generation 0), as the operator does after a restart, and then waits for the process rejoin.
• wait reports an error only while it is the latest answer.
• Why: lab run 10, trial 7 (tpl-resilience-reload-alternate) restarted CMOnet at game 10:30. The new process started 88 s later and waited, unconnected, for a Connect, as it is meant to. The runner waited 600 s for a rejoin that nothing had asked for, then failed the trial with "last error: timed out", a timeout from early in the wait. This was the first live run of the fault, since smoke runs skip it.
• CMOnet itself is unchanged: it still does not reconnect on its own after its own restart (decision 9 covers CMO restarts only).

Test plan

• tests/test_laboratory_runs.py::test_a_restarted_controller_is_connected_again_before_its_rejoin_is_awaited
• tests/test_laboratory_runs.py::test_a_wait_reports_an_error_only_while_it_is_the_latest_answer
• tests/test_laboratory_runs.py
• Full tools/check.py --rev in the background

🤖 Generated with Claude Code

## Summary - The lab runner's `restart_cmonet` fault now connects the restarted CMOnet to CMO once the new process answers (`connection.state` `not_initialized`, generation 0), as the operator does after a restart, and then waits for the process rejoin. - `wait` reports an error only while it is the latest answer. - Why: lab run 10, trial 7 (tpl-resilience-reload-alternate) restarted CMOnet at game 10:30. The new process started 88 s later and waited, unconnected, for a Connect, as it is meant to. The runner waited 600 s for a rejoin that nothing had asked for, then failed the trial with "last error: timed out", a timeout from early in the wait. This was the first live run of the fault, since smoke runs skip it. - CMOnet itself is unchanged: it still does not reconnect on its own after its own restart (decision 9 covers CMO restarts only). ## Test plan - [x] `tests/test_laboratory_runs.py::test_a_restarted_controller_is_connected_again_before_its_rejoin_is_awaited` - [x] `tests/test_laboratory_runs.py::test_a_wait_reports_an_error_only_while_it_is_the_latest_answer` - [x] `tests/test_laboratory_runs.py` - [ ] Full `tools/check.py --rev` in the background 🤖 Generated with [Claude Code](https://claude.com/claude-code)

vinny added 2 commits 2026-10-17 11:15:40 -04:00

Lab runs connect a restarted CMOnet again ... 32ef686a34

A controller that starts does not connect itself; the operator connects it
after a restart. Lab run 10's restart_cmonet fault (trial 7) waited 600 s for
a process rejoin nothing had asked for, and the trial failed. The runner now
waits for the new process (never connected, generation 0), connects it to CMO
and then waits for the rejoin. A wait reports an error only while it is the
latest answer: the failure named an early timeout as its cause.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Version 0.7.92 (restart-reconnect) c11c61b18e

vinny manually merged commit c11c61b18e into master 2026-10-17 11:16:32 -04:00

vinny referenced this pull request 2026-10-17 19:58:38 -04:00

Build the laboratory corpus with the Scenario Builder #7

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels

No items

No labels

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

vinny/CMOnet!149

No description provided.