Menu
▾
▴
[Bacula-bugs] [bacula 0001507]: cancelled duplicate job shows up as failed
|
From: Mantis B. T. <no...@bu...> - 2010-02-26 13:21:51
|
The following issue has been ACKNOWLEDGED. ====================================================================== http://bugs.bacula.org/view.php?id=1507 ====================================================================== Reported By: atrox Assigned To: ====================================================================== Project: bacula Issue ID: 1507 Category: Director Reproducibility: always Severity: minor Priority: normal Status: acknowledged ====================================================================== Date Submitted: 2010-02-17 07:39 UTC Last Modified: 2010-02-26 13:03 UTC ====================================================================== Summary: cancelled duplicate job shows up as failed Description: I have these directives for a job: Allow Higher Duplicates = no Rerun Failed Levels = yes When a duplicate of the job is ran and it's cancelled (as it should), the next higher-level job (eg. incremental) will be upgraded to full again, because: "Prior failed job found in catalog. Upgrading to Full." I don't see any reason for a cancelled duplicate job to cause the same level to be ran again. Is there any? ====================================================================== ---------------------------------------------------------------------- (0005118) kern (administrator) - 2010-02-17 10:28 http://bugs.bacula.org/view.php?id=1507#c5118 ---------------------------------------------------------------------- Unfortunately, the Allow Duplicate Job code has a number of bugs, particularly with Allow Higher Duplicates. The feature has been significantly re-written for version 5.0.1 (Allow Higher Duplicates is no longer available, and there is a new directive). Since so much has changed, looking at this particular bug does not make sense. However, if you want to test the new version, I recommend that you pull it from the git repo from branch "Branch-5.0" (not master). We will be releasing it this week or during the weekend. There may be an issue that you bring up with Rerun Failed Jobs, but at the moment I don't completely understand it and it doesn't make sense to address it with the old code. We'll need to look at Rerun Failed Jobs when you are using Bacula version 5.0.1. ---------------------------------------------------------------------- (0005142) atrox (reporter) - 2010-02-26 12:29 http://bugs.bacula.org/view.php?id=1507#c5142 ---------------------------------------------------------------------- I tested this with 5.0.1 now (replacing "Allow Higher Duplicates = no" with "Allow Duplicate Jobs = no"). It still works the same way, ie. the duplicate (and cancelled) job is seen as failed and the level is thus ran again. ---------------------------------------------------------------------- (0005143) kern (administrator) - 2010-02-26 13:03 http://bugs.bacula.org/view.php?id=1507#c5143 ---------------------------------------------------------------------- Unfortunately, when you first submitted this, the situation was confused with the problems of Allow Higher Duplicates, and I didn't fully appreciate or understand the problem of a subsequent upgrade due to a canceled job. I agree with you I don't think that it is appropriate to upgrade the next job after a canceled one. However, we need to figure out a solution since that is the normal behavior, and I think this is one of the few or only cases where we really don't want the upgrade to apply. Issue History Date Modified Username Field Change ====================================================================== 2010-02-17 07:39 atrox New Issue 2010-02-17 10:28 kern Note Added: 0005118 2010-02-17 10:28 kern Status new => closed 2010-02-17 10:28 kern Resolution open => won't fix 2010-02-26 12:29 atrox Note Added: 0005142 2010-02-26 12:29 atrox Status closed => feedback 2010-02-26 12:29 atrox Resolution won't fix => reopened 2010-02-26 13:03 kern Note Added: 0005143 2010-02-26 13:03 kern Status feedback => acknowledged ====================================================================== |