Just complementing... As Kern said, let Bacula handle this for you is the best way.
Best regards, Ana On Thu, Oct 1, 2015 at 11:17 AM, Ana Emília M. Arruda < emiliaarr...@gmail.com> wrote: > Hello Georg, > > An unmount will not solve this for you. Maybe a Run After Job directive > that in the case of a Job fail a script could delete the volume from > catalog and the volume file in your archive device (if you're using tapes, > this will be a little complicated...). This way another volume for the next > job will be labeled correctly, but I'm affraid you will have a very large > MediaId number in the case you have a large number of job failures... > > You can also have a few pools and your volume labels without > client/date/job information. Bconsole have a few commands very usefull to > find out where is specific data for a job/client or whatever you are > looking for. The restore command lets you choose some ways to find out what > you want to restore. There is also a "query" command with some useful SQL > queries and you can have your own queries too. You can formulate them and > include in the query.sql file. > > Best regards, > Ana > > On Thu, Oct 1, 2015 at 10:52 AM, Georg Altmann <geo...@george-net.de> > wrote: > >> Hello Ana, >> >> that makes perfect sense. I overlooked this. Thanks for clearing it up! >> >> My intention here is to ease the search for a recent backup in a >> emergency / recovery situation with bacula command line tools. Maybe I >> can force a volume umount with a Run Before Job directive. >> >> Creating a separate pool for each client or job seems to cause a lot of >> maintenance work and complicates the setup a lot. I don't think this is >> a good solution. >> >> Regards, >> Georg >> >> Am 30.09.2015 um 18:45 schrieb Ana Emília M. Arruda: >> > Hello Georg, >> > >> > IMHO, this is not a bug, but the expected bacula behaviour. >> > >> > Before running a job, bacula tries to reserve a volume (according to its >> > volume recycling algorithm). If it did not find one, it labels one. If >> > the job fails, the volume will be in the pool available for other jobs >> > that uses this pool for its backups. >> > >> > I particularly do not like base a volume label in a job or date if it >> > can be used/recycled by other jobs in any other moments. If you want to >> > have specific volumes to be used by specific jobs/clients, I would >> > recommend you to have specific pools for these clients/jobs. >> > >> > Best regards, >> > Ana >> > >> > On Wed, Sep 30, 2015 at 1:14 PM, Georg Altmann <geo...@george-net.de >> > <mailto:geo...@george-net.de>> wrote: >> > >> > Hi, >> > >> > >> > there seems to be a problem with variable expansion in >> Pool/LabelFormat >> > after failed jobs. I randomly get expansions that do not have the >> > expected value. >> > >> > Pool { >> > Name = Disk3Pool >> > Pool Type = Backup >> > Recycle = no >> > AutoPrune = Yes >> > Volume Retention = 500 years >> > Maximum Volume Jobs = 1 >> > Label Format = >> > >> >> "disk3-${Disk3VolNum+:p/6/0/r}-${Year}-${Month:p/2/0/r}-${Day:p/2/0/r}-${Job}-${Level}" >> > } >> > >> > >From the log, with expected volume name: >> > >> > Build OS: x86_64-unknown-linux-gnu archlinux >> > Job: gurke-arch-linux.2015-09-30_17.17.01_08 >> > Backup Level: Incremental, since=2015-09-15 08:55:26 >> > Client: "gurke-arch-linux-fd" 7.0.5 (28Jul14) >> > x86_64-unknown-linux-gnu,archlinux, >> > Pool: "Disk3Pool" (From Job resource) >> > Volume name(s): >> > disk3-000616-2015-09-30-gurke-arch-linux-Incremental >> > >> > But then the label is often incorrect / unexpected: >> > >> > Job: BackupCatalog.2015-09-30_17.17.01_16 >> > Backup Level: Full >> > Client: "gurke-arch-linux-fd" 7.0.5 (28Jul14) >> > x86_64-unknown-linux-gnu,archlinux, >> > FileSet: "Catalog" 2011-11-13 17:21:08 >> > Pool: "Disk3Pool" (From Job resource) >> > Volume name(s): >> > disk3-000618-2015-09-30-georgtp-arch-Incremental >> > >> > Another incarnation of the same job with correct volume label: >> > >> > Job: BackupCatalog.2015-09-15_08.55.24_39 >> > Backup Level: Full >> > Client: "gurke-arch-linux-fd" 7.0.5 (28Jul14) >> > x86_64-unknown-linux-gnu,archlinux, >> > FileSet: "Catalog" 2011-11-13 17:21:08 >> > Pool: "Disk3Pool" (From Job resource) >> > Volume name(s): >> > disk3-000615-2015-09-15-BackupCatalog-Full >> > >> > >> > This problem seems to occur when a previous run job fails. In fact >> the >> > unexpected label >> > disk3-000618-2015-09-30-georgtp-arch-Incremental >> > for the job >> > BackupCatalog.2015-09-30_17.17.01_16 >> > would have been the correct label for the first of two failed jobs >> that >> > were scheduled before BackupCatalog: >> > >> > Job: georgtp-arch.2015-09-30_17.17.01_12 >> > Backup Level: Incremental, since=2015-09-15 09:12:20 >> > Client: "georgtp" 7.0.5 (28Jul14) >> > x86_64-unknown-linux-gnu,archlinux, >> > Volume name(s): >> > Volume Session Id: 3 >> > Volume Session Time: 1443626168 >> > Last Volume Bytes: 0 (0 B) >> > Non-fatal FD errors: 1 >> > SD Errors: 0 >> > FD termination status: Error >> > SD termination status: Waiting on FD >> > Termination: *** Backup Error *** >> > >> > Notice the empty Volume name field. >> > >> > So it seems as if the volume label for a failed job remains in some >> kind >> > of queue of buffer that is then incorrectly used for a following >> job. >> > >> > Can anybody confirm this? >> > I did not find a bug report in mantis. May this have been fixed with >> > 7.2? >> > >> > I am running bacula 7.0.5 on Arch Linux: >> > % pacman -Q |grep bacula >> > bacula-bat 7.0.5-1 >> > bacula-common 7.0.5-1 >> > bacula-console 7.0.5-1 >> > bacula-dir 7.0.5-1 >> > bacula-dir-postgresql 7.0.5-1 >> > bacula-fd 7.0.5-1 >> > bacula-sd 7.0.5-1 >> > >> > >> > Regards, >> > Georg >> > >> > >> > -- >> > PGP-Key: 0x1E320E65 >> > D150 7783 A0D1 7507 1266 C5B3 BBF1 9C42 1E32 0E65 >> > >> > I don't like the idea of secret agencies to analyse and archive >> > personal communication. GnuPG is available as open source, free as >> as in >> > freedom, as a countermeasure. I use http://www.enigmail.net/ for >> Mozilla >> > Thunderbird. If you can, please use a frontend of your choice to >> send me >> > encrypted e-mail. See http://www.gnupg.org/ for an overview. >> > >> > >> > >> > >> >> ------------------------------------------------------------------------------ >> > >> > _______________________________________________ >> > Bacula-devel mailing list >> > Bacula-devel@lists.sourceforge.net >> > <mailto:Bacula-devel@lists.sourceforge.net> >> > https://lists.sourceforge.net/lists/listinfo/bacula-devel >> > >> > >> >> -- >> PGP-Key: 0x1E320E65 >> D150 7783 A0D1 7507 1266 C5B3 BBF1 9C42 1E32 0E65 >> >> I don't like the idea of secret agencies to analyse and archive >> personal communication. GnuPG is available as open source, free as as in >> freedom, as a countermeasure. I use http://www.enigmail.net/ for Mozilla >> Thunderbird. If you can, please use a frontend of your choice to send me >> encrypted e-mail. See http://www.gnupg.org/ for an overview. >> >> >
------------------------------------------------------------------------------
_______________________________________________ Bacula-devel mailing list Bacula-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-devel
