fix prolog/epilog timeout handling #6677
Merged
+36
−52
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Problem: A comment in the perilog jobtap plugin is outdated. Fix the comment to describe actual plugin behavior.
Problem: When a prolog/epilog timeout is configured, the perilog.so plugin delays the start of the timeout timer until the bulk-exec `on_start` callback is called. But this could result in the timer never being started if one or more tasks get stuck starting. Start the timeout timer immediately on launch of the prolog or epilog tasks. This removes the need for the start_cb(), so remove that function. Fixes flux-framework#6644
Problem: The perilog plugin raises an error if the kill-timeout is set for the epilog, but there is no reason any longer to disallow this, in fact it may be useful in the future. Remove the code that prevented kill-timeout from being set for the epilog. Update one test.
chu11
approved these changes
Mar 3, 2025
doc/man5/flux-config-job-manager.rst
Outdated
| seconds to wait until any nodes with prolog tasks that are still | ||
| active will be drained. The drain reason will include the string | ||
| "canceled then timed out". The default is 60. | ||
| (optional, float) If a the prolog times out, or a job exception is raised |
There was a problem hiding this comment.
both here and in epilog, "If a the" -> "If the"?
doc/man5/flux-config-job-manager.rst
Outdated
Comment on lines
135
to
136
| (optional, bool) A boolean indicating whether a fatal job exception is | ||
| raised while the prolog is active terminates the prolog. The default |
There was a problem hiding this comment.
both here and in epilog, first sentence seems off to me with the "while the prolog is active terminates the prolog". Perhaps reorder the sentence, "A boolean indicating if a prolog should be terminated if a fatal job exception is raised while the prolog is active."?
There was a problem hiding this comment.
Oops, this was an edit issue - two versions of the sentence remain grafted together here. Anyway, I like your suggestion so let's use that.
Problem: The documentation of the prolog and epilog configuration values in flux-config-job-manager(5) is unclear in many ways and incorrect in some ways. Amend the documentation for clarity and correctness.
|
Thanks! I've fixed the wording as suggested and will set MWP. |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #6677 +/- ##
==========================================
- Coverage 83.86% 83.85% -0.01%
==========================================
Files 534 534
Lines 88939 88931 -8
==========================================
- Hits 74588 74576 -12
- Misses 14351 14355 +4
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds a probable fix for #6644 (epilog doesn't time out), plus some other bits of cleanup.