Sign every compute task with run ID to correlate response #7463
Conversation
Unit Test ResultsSee test report for an extended history of previous test failures. This is useful for diagnosing flaky tests. 24 files ± 0 24 suites ±0 10h 18m 43s ⏱️ + 5m 18s For more details on these failures, see this check. Results for commit 4c4663c. ± Comparison against base commit 6dd3c70. ♻️ This comment has been updated with latest results. |
| @@ -536,6 +536,16 @@ async def release_all_futures(): | |||
| (f3.key, "executing", "released", "cancelled", {}), | |||
| (f3.key, "cancelled", "fetch", "resumed", {}), | |||
| (f3.key, "resumed", "memory", "memory", {}), | |||
| ( | |||
There was a problem hiding this comment.
The resumed task should be rejected by the scheduler because its run ID is stale, which triggers the task to be released and recomputed.
…ging stimulus_task_finished
|
|
|
From what I understand, failures are likely unrelated and just general CI flakiness. |
| @@ -4646,6 +4676,24 @@ def stimulus_task_finished(self, key=None, worker=None, stimulus_id=None, **kwar | |||
| "stimulus_id": stimulus_id, | |||
| } | |||
| ] | |||
| elif ts.run_id != run_id: | |||
There was a problem hiding this comment.
The clauses in stimulus_task_finished could likely be improved, but that would also mean some breaking changes to the transition logic and should be done in a PR focussing on that.
Supersedes #7372
pre-commit run --all-files