Job stuck with message "Jobs waiting to reserve a drive"

[Thomas Plant]

Hello. First post on this group 

I’ve a problem with a job in Bareos 13.2.2 .

A single job is stuck since this night with the following message:

Jobs waiting to reserve a drive:
3608 JobId=808 wants Pool="Incr" but have Pool="Full" nreserve=0 on drive "FileStorage" (/bareos).

There is a second job which is running normally and doing a full backup.

Only thing what happened this night was that the ‚Incr‘ Pool recycled one of the volumes:

23-Jun 23:10 nfs2-sd JobId 800: User defined maximum volume capacity 10,737,418,240 exceeded on device "FileStorage" (/bareos).
23-Jun 23:10 nfs2-sd JobId 800: End of medium on Volume "Incr-0034" Bytes=10,737,389,924 Blocks=166,445 at 23-Jun-2014 23:10.
23-Jun 23:10 nfs2-sd JobId 800: Job vs29.2014-06-23_23.05.00_14 is waiting. Cannot find any appendable volumes.
Please use the "label" command to create a new Volume for:
Storage: "FileStorage" (/bareos)
Pool: Incr
Media type: File
23-Jun 23:25 nfs2-dir JobId 800: There are no more Jobs associated with Volume "Incr-0036". Marking it purged.
23-Jun 23:25 nfs2-dir JobId 800: All records pruned from Volume "Incr-0036"; marking it "Purged"
23-Jun 23:25 nfs2-dir JobId 800: Recycled volume "Incr-0036"
23-Jun 23:25 nfs2-sd JobId 800: Recycled volume "Incr-0036" on device "FileStorage" (/bareos), all previous data lost.
23-Jun 23:25 nfs2-sd JobId 800: New volume "Incr-0036" mounted on device "FileStorage" (/bareos) at 23-Jun-2014 23:25.
23-Jun 23:31 nfs2-sd JobId 803: Elapsed time=00:26:29, Transfer rate=1.855 M Bytes/second
23-Jun 23:31 nfs2-dir JobId 803: Bareos nfs2-dir 13.2.2 (12Nov13):

But all other backups continued after this normally. Only a single job is sitting there and waiting …….

If you need any more logs or config files just let me know..

Any hints how to get the job working would be very appreciated. Many thanks.
Thomas Plant
Limitis GmbH

[Thomas Plant]

Sorry, some information missing I think:

We are doing the backup to disk (no tape involved). The stuck job should do an incremental backup and the other job I mentioned above is doing a Full Backup.
First job is a Linux FD and second is on Windows.

Concurrent jobs is setted to 20. And all jobs run when specified. Don't think this is the problem.

Thanks,
Thomas

[Thomas Plant]

Update: after the 'Full' backup job terminated the waiting 'Incremental' job started and finished successfully.

Could it been that because the 'Full' job was upgraded from incremental as it was it's first run and blocked the other?

[lst_...@kwsoft.de]

Zitat von Thomas Plant <blues...@gmail.com>:

Priorities are checked at starttime of the jobs and only run at the
same time if the priorities are equal. If you have different
priorities all running jobs must finish first to get the rest
evaluated again. So if your full has different priority the
incremental have to wait for it to finish.

Regards

Andreas

[Thomas Plant]

Hi.
Did not set the priority parameter in the job definition. I presume so they have all the same priority?

[Thomas Plant]

Here I'm again. Excluded some directories from the backupset vs21. It is doing againg (obviously) a full backup. But the one job I mentioned yesterday is again waiting on completing this one job.

I look at the configuration, all jobs have the same priority (10), in a JobDef.

What I see and it seems strange to me:

Running Jobs:
Writing: Full Backup job vs21 JobId=823 Volume="Full-0101"
pool="Full" device="FileStorage" (/bareos)
spooling=0 despooling=0 despool_wait=0
Files=426,002 Bytes=122,305,730,536 AveBytes/sec=5,906,020 LastBytes/sec=6,322,656
FDReadSeqNo=6,299,959 in_msg=5073117 out_msg=5 fd=6
Writing: Incremental Backup job vs23 JobId=824 Volume="Full-0101"
pool="Incr" device="FileStorage" (/bareos)
spooling=0 despooling=0 despool_wait=0
Files=0 Bytes=0 AveBytes/sec=0 LastBytes/sec=0
FDSocket closed

Job vs23 has 'Volume=Full-1010"', but it should do an incremental and thereby using an 'Diff-xxxx' Volume. And what is even stranger, the vs23 job starts at 23:00 and the vs21 much later 02:35 in the morning.....

Thanks for any hints how to resolve this.

Thomas P.

[Marco van Wieringen]

The reason is simple and overlooked by most people but a device in
Bacula/Bareos can only have one volume at a time. So the file storage
can only have "loaded" either a volume in the "Full" or "Incr" pool.
On disk you could have multiple volumes open but for that to work you
have several options:

- diskchanger script.
- vchanger
- internal file base autochanger via changerdevice /dev/null

The idea is that you then define multiple devices and the code (which
has a tapechanger background) will do the right thing and load one volume
in the virtual device 0 of the diskchanger and the other into virtual
drive 1 of the diskchanger. The whole autochanger setup is not ideal
as it has some problems when a certain volume is loaded in a certain
"drive" and the code want it to be in an other "drive". That is something
we eventually hope to fix but that is still on the to do list.

--
Marco van Wieringen marco.van...@bareos.com
Bareos GmbH & Co. KG Phone: +49-221-63069389
http://www.bareos.com

Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646
Komplementär: Bareos Verwaltungs-GmbH
Geschäftsführer: Stephan Dühr, M. Außendorf, J. Steffens,
P. Storz, M. v. Wieringen