Bug 475865
| Summary: | Periodic* race in JobRouter (and elsewhere) | ||
|---|---|---|---|
| Product: | Red Hat Enterprise MRG | Reporter: | Matthew Farrellee <matt> |
| Component: | grid | Assignee: | Matthew Farrellee <matt> |
| Status: | CLOSED ERRATA | QA Contact: | Jeff Needle <jneedle> |
| Severity: | high | Docs Contact: | |
| Priority: | high | ||
| Version: | 1.0 | CC: | dan, jsarenik, rrati |
| Target Milestone: | 1.1.1 | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2009-04-21 16:18:14 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
This will be addressed in 7.2.1-0.2
commit d1763c3bc25c2dfb511a61048eb872d5f28fd2da
Author: Dan Bradley <dan>
Date: Fri Dec 12 17:38:52 2008 -0600
Added protection against periodic expressions messing up the 'spooling' hold state.
This is a temporary solution for 7.2 only. In 7.3, we will get rid of
the spooling state.
I have reproduced the bug on RHEL5.3 i386, condor-7.2.0-3.el5 (from MRG 1.1) The bug is not present on condor-7.2.2-0.7.el5 (from MRG-candidate repo). An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHEA-2009-0434.html |
Specifically for the JobRouter, the configuration contains: set_PeriodicRemove = JobStatus == 5 || \ (JobStatus == 1 && (CurrentTime - QDate) > 3600*6); \ JobStatus 5 is the Hold state. When a job is submitted to Condor it is briefly put in the Hold state while its data is spooled. That brief hold can be long enough, or just timed poorly enough, for the Periodic expressions to be evaluated. In this example, the PeriodicRemove evaluates to true, JobStatus == 5, and the job is removed before it is completely spooled. Oops! A temporary workaround, which is fragile, is to test (JobStatus == 5 && HoldReason =!= "Spooling input data files").