Bacula-users

Re: [Bacula-users] Possible bug in database update code

2011-08-10 08:52:32
Subject: Re: [Bacula-users] Possible bug in database update code
From: Martin Simmons <martin AT lispworks DOT com>
To: bacula-users AT lists.sourceforge DOT net
Date: Wed, 10 Aug 2011 13:49:47 +0100
>>>>> On Tue, 9 Aug 2011 12:02:18 +0200, Frank Altpeter said:
> 
> Hi there,
> 
> It's possible that I just hit a bug with the routines that are
> updating the catalog database.
> 
> Following prerequisites:
> 
> Server OS: SUSE Linux Enterprise Server 11 (x86_64)
> Bacula Version: 5.0.3
> MySQL Version: 5.0.67
> Catalog Database size: 49G
> 
> The configuration states to do incremental backup daily except sunday
> night where the full backup is running. Full backup is done from
> client to disk via vchanger. After all full backups are done, monday
> afternoon a copy job is being run which brings the full backups from
> vchanger to an LTO-4 tape changer.
> Usually this setting is running fine.
> 
> Today (tuesday) one of my clients is doing a full backup, where it
> should run incremental. Looking in the log it says "Prior failed job
> found in catalog. Upgrading to Full.". I was wondering about that,
> because I know that the incremental backup from yesterday was running
> fine.
> 
> So I checked the database entries for this client. I've put the output of
> 
> select 
> JobId,PriorJobId,Job,Type,Level,JobStatus,SchedTime,StartTime,EndTime,RealEndTime
> from Job where Name like 'host.customer.com%' order by JobId;
> 
> into http://pastebin.com/4CVHzqaD for review.
> 
> The interesting line are the ones with JobId 113187 and 112500. I
> don't know why these two jobs have identical values for StartTime and
> EndTime. I think that some catalog update routine was going wrong here
> because it was confused by the running copy job. I see that database
> entry as reason that the incremental job of today thinks that the last
> job failed.

The identical values for StartTime and EndTime are normal for copy jobs, so
there is probably a bug in the code that finds the prior failed job.

The copy job 113295 was still running at 2011-08-09 00:42:01 (when 113450
found the prior failed job) so I suspect it found that copy job instead of the
real full job 113026.

I suggest you report it at bugs.bacula.org, with the full logs and the
pastebin contents.

__Martin

------------------------------------------------------------------------------
uberSVN's rich system and user administration capabilities and model 
configuration take the hassle out of deploying and managing Subversion and 
the tools developers use with it. Learn more about uberSVN and get a free 
download at:  http://p.sf.net/sfu/wandisco-dev2dev
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users

<Prev in Thread] Current Thread [Next in Thread>