Le Thursday 05 March 2009 17:41:51 Dirk Bartley, vous avez écrit : > On Thu, 2009-03-05 at 15:40 +0100, Eric Bollengier wrote: > > Hello Dirk, > > > > Bugs are more easy to track and to fix with mantis tool > > http://bugs.bacula.org > > All apologies for taking so long. Busy day at work today. > Unfortunately, I didn't understand yet that this was a bug. Are you > officially saying that it should be. I thought it may just be something > like "Read the document at http://xxx"; for the proper way to migrate to > what is going to be 3.0.
In fact, i don't see the problem. Without the complete job output, it's very hard to see what you say. > > Le Thursday 05 March 2009 15:00:15 Dirk Bartley, vous avez écrit : > > > Greeting > > > > > > I'm using the head development version in my live active work server to > > > get a better chance of solving some issues in bat. Now however I'm > > > having some issues that I hoping can be solved on the list. > > > > > > Revision: 8498 > > > > > > My recollection is that this particular issue started after updating > > > the storage daemon. My storage daemon runs on my workstation. > > > > > > So I have been getting lots of failures with the last couple of nights > > > backups. > > > > Failures with tape (hardware or software) can desynchronize File count > > between the volume and the catalog. In this case, bacula will put the > > Volume in Error state until you fix it (which is a good idea if the > > volume is dirty or broken). > > Maybe it's just dirt. It has been a bit since I ran my cleaning tape > over the heads. All do that now and try again. It would be a pain if > dirt affected my like this at the same time I updated software. I'm just trying to imagine what "lots of failures" means. > > > This morning I tried to reproduce in a simple way and > > > succeded. I started with a tape that was freshly purged and relabeled. > > > Started a job which would use the main server as the client as well. > > > It started along backing up to tape just fine. Then I started a job > > > which uses a different file daemon and backing up on the tape drive > > > just stopped. > > > > You describe the most common case, i don't see the problem. Is your first > > job OK ? or you've got an error somewhere ? The job ouput is welcome. > > That's curious. I would say starting a second job before the first job > is finished and that resulting in the stopping of backing up from > occurring to the tape entirely would be an indication of a problem. It's a very strange problem, regression scripts are doing this all the time, and it seems to work well. > > > All efforts to set the tape back to append, > > > > Which one ? What is the error message ? "update volume" command permits > > to change the volume status or the file count easily. > > Well I'm describing a common occurance to edit the tape properties with > bat like so: > > update volume="LTO_WDay_004_" volstatus="Append" recyclepool="" > New Volume status is: Append > New Volume status is: Append > New RecyclePool is: *None* > New RecyclePool is: *None* If you have a file count problem, you need also to update volfile parameter (not possible on command line i think) > > > umount and mount the drive > > > to get the jobs going to tape again have failed. > > > > > > > > > Here is the log from the second job. > > > > > > Log records for job 7910 > > > 2009-03-05 08:36:30 > > > srvalum3-dir > > > Using Device "LTO" > > > > > > Start Backup JobId 7910, Job=UPSJob.2009-03-05_08.36.27.05 > > > 2009-03-05 08:36:58 > > > srvalum2-sd > > > Volume "LTO_WDay_004_" previously written, moving to end of data. > > > 2009-03-05 08:37:14 > > > srvalum2-sd > > > Marking Volume "LTO_WDay_004_" in Error in Catalog. > > > > > > Error: Bacula cannot write on tape Volume "LTO_WDay_004_" because: > > > The number of files mismatch! Volume=1 Catalog=0 > > > 2009-03-05 08:37:28 > > > srvalum2-sd > > > Job UPSJob.2009-03-05_08.36.27.05 waiting. Cannot find any appendable > > > volumes. > > > Please use the "label" command to create a new Volume for: > > > Storage: "LTO" (/dev/nst1) > > > Pool: LTO_TueThu > > > Media type: LTO > > > > > > After that job is started, the tape in the medialist goes from > > > status=append to status=error. > > > > This is normal if bacula founds a difference between the catalog and the > > volume. For example, if you shutdown the storage daemon during a job, you > > will have this behavior > > In this case, I'm not shutting down the storage daemon, just starting a > second job. Something that in the past has not caused this. > > > > Here are my config files that I think my be pertainent. > > > http://www.schupan.com/baculaconfs/bacula-dir.conf > > > http://www.schupan.com/baculaconfs/bacula-sd.conf > > > > > > As always thank you very much in advance for any assistance. Let me > > > know if there is any information I can provide. > > > > The support page on www.bacula.org provides many information on how to > > report a problem. > > I'm thinking from the perspective of I'm not a common user reporting a > bug (that has yet to be established as a bug yet). I'm a developer that > has spent many hundreds of hours writing software for the project > interested in doing some fixing of my part of the project but finding a > relatively serious issue in the way of doing that. In this page, you will find also needed information for problem reporting - job output - configuration files - step to reproduce - etc... > All apologies. No worry, bat is wonderful and you are doing a great job, helping you to resolv your problem is the least i can do. > What I will probably want to do is set my live environment back to a > working version, it's been a couple of days since I got a good set of > backups. Will having updated the database cause any issues with going > back in versions??? Any suggesions as to what version I should revert > to?? Nobody has reporting this kind of serious problem, i can't advise you on a specific version. To revert your catalog to the previous version, you can just update the version field to 10, it should work like a charm. UPDATE Version SET VersionId=10; You will have to revert to 11 yourself later. Bye > I can get another environment set up to do the development I was > hoping to do and maybe I'll run in to this issue again in development > environment. > > Dirk > > > Bye > > > > > Dirk ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ Bacula-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/bacula-devel
