Il 12 giu 2017 3:11 PM, "Jon SCHEWE"
<[email protected]<mailto:[email protected]>> ha scritto:
However when the job needs to switch tapes and the changer isn't full, the job
gets stuck waiting on the storage.
Please, can you elaborate on this? "gets stuck waiting" for? An empty volume? A
missing, partially used volume? A scratch volume, missing in the library?
The new job has gone past the schedule stage and says "Waiting for storage".
The problem appears to be that the storage daemon doesn't know where to put the
tape that is currently in the
Also, what exactly are reported by "status dir" and "status storage" during
such "stuckness"?
I did not keep the screen shots of it. The storage daemon believed that a tape
was in use. The director said "waiting for storage <storage name>" on the job.
tried just filling the changer and running "update slots".
Filling with what? Empty volumes properly labeled and initialized
(scratch/pools/etc.)? Partially used tapes? Other?
This morning my offsite tapes showed up and I was able to put those back in the
changer.
Here's a timeline:
1. 2 weeks ago I ran an offsite backup. That used 2 tapes.
2. I ejected those 2 tapes from the library manually and sent them off. At
this point I had 1 tape in the drive and 3 empty slots.
3. Sometime later "update slots" was run
4. This past weekend my next offsite backup started.
5. The tape that was in the drive was not for my offsite pool, so bareos
needed to take the tape out of the drive.
6. This morning I came into the office and found bareos stuck with the
offsite job waiting on storage.
7. I tried running "mount" and that didn't change things
8. I tried running "unmount" and the storage daemon reported that the tape
was in use
9. I manually put the two tapes I had just received into the changer
10. I ran "update slots" and bareos noted that 3 tapes had their changer slot
updated (including the one in the drive)
11. I tried running "mount" and no change
12. I restarted the storage daemon and that caused the job to fail, but then
the next offsite job started as it should and switched tapes without problems.
However that didn't get the job started, I needed to restart the storage daemon
and then everything started working again.
It would be nice if you provide bareos log file (DIR log, I mean) cought BEFORE
and AFTER such a restart. They surely contains important details.
11-Jun 23:05 mn-server-dir JobId 33615: Start Backup JobId 33615,
Job=windigo-offsite.2017-06-11_04.00.00_50
11-Jun 23:05 mn-server-dir JobId 33622: 2001 Job windigo.2017-06-10_23.05.00_43
marked to be canceled.
11-Jun 23:05 mn-server-dir JobId 33622: 3000 JobId=33608
Job="windigo.2017-06-10_23.05.00_43" marked to be canceled.
11-Jun 23:05 mn-server-dir JobId 33623: Cancelling duplicate JobId=33609.
11-Jun 23:05 mn-server-dir JobId 33623: JobId 33609, Job
superior.2017-06-10_23.05.00_44 marked to be canceled.
11-Jun 23:05 mn-server-dir JobId 33609: Bareos mn-server-dir 16.2.4 (01Jul16):
11-Jun 23:05 mn-server-dir JobId 33624: Cancelling duplicate JobId=33610.
11-Jun 23:05 mn-server-dir JobId 33624: JobId 33610, Job
shoretel.2017-06-10_23.05.00_45 marked to be canceled.
11-Jun 23:05 mn-server-dir JobId 33610: Bareos mn-server-dir 16.2.4 (01Jul16):
... More cancellations of duplicate jobs that are waiting because the offsite
is stuck.
This shows up after I restarted the storage daemon.
12-Jun 08:08 mn-server-dir JobId 33615: Fatal error:
Storage daemon didn't accept Device "LTO-changer" command.
12-Jun 08:08 mn-server-dir JobId 33615: Error: Bareos mn-server-dir 16.2.4
(01Jul16):
Have others seen this?
Is this expected behavior for bareos?
- Yes (if BareOS need a tape, and in the library there is no "available" tape);
That should have generated an email for manual intervention required. This
generated no such email. The only way I knew it was stuck was by looking at the
director.
- No (storage restart should not be needed at all)
Bye
DV
--
You received this message because you are subscribed to the Google Groups
"bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.
