Project

General

Profile

Bug #24057

jobsub_q --better-analyze often asks wrong schedd for job info

Added by Dennis Box 5 months ago. Updated 3 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
-
Target version:
Start date:
02/17/2020
Due date:
% Done:

0%

Estimated time:
First Occurred:
Occurs In:
Stakeholders:
Duration:

Description

See INC000001092616.

The jobsub_q command, for example:
jobsub_q -G mu2e --better-analyze --jobid

Is trying to ask schedd jobsub03.fnal.gov to analyze why job 2898203.0 has not started.

There is intermediate code between the above jobsub_q command and the resulting condor_q command that finds the least busy schedd (there are currently 3) and directs the query to that one. In the above example I found this in the log:

condor_q -allusers -nobatch -name jobsub02.fnal.gov -better-analyze -constraint 'True && True && regexp("jobsub03.fnal.gov#2898203\.0#.*",GlobalJobId) && True'.

-name jobsub02.fnal.gov should be -name jobsub03.fnal.gov. This got through CI as we only have one schedd on our test setup.

Associated revisions

Revision a7cd2f2b (diff)
Added by Dennis Box 2 months ago

test for issue #24057

History

#1 Updated by Dennis Box 3 months ago

  • Assignee changed from Dennis Box to Shreyas Bhat
  • Status changed from New to Feedback

see branch 24057 for review. You may diff against bd5ac0c58bedbf02d31d4ac54b00f7acde79b58c to see relevant changes

#2 Updated by Shreyas Bhat 3 months ago

So this looks good, but I'm wondering if lines 267-279 can be more cleanly and clearly accomplished with regex matching, since we're trying to extract the schedd there. Perhaps something like:

    if a_filter:
        jobid_cnst_regex = re.compile('JobsubJobId==\"\d+(?:\.\d+)?@(.+)\"')
        match = jobid_cnst_regex.search(a_filter)
        if match is not None and match.groups():
            s_list = list(match.groups(1))

What are your thoughts on that?

#3 Updated by Shreyas Bhat 3 months ago

  • Assignee changed from Shreyas Bhat to Dennis Box
  • Status changed from Feedback to Under Discussion

#4 Updated by Dennis Box 3 months ago

  • Status changed from Under Discussion to Resolved

I like the review suggestion. Merged a slightly modified 24057 to master.



Also available in: Atom PDF