Bacula-users

[Bacula-users] Fwd: Bacula went fubar after some weeks

2009-02-10 11:21:05
Subject: [Bacula-users] Fwd: Bacula went fubar after some weeks
From: John Drescher <drescherjm AT gmail DOT com>
To: bacula-users <Bacula-users AT lists.sourceforge DOT net>
Date: Tue, 10 Feb 2009 11:16:42 -0500
Hi John,

I did not post "list media" output because I manually tried to fix
thing using commands from bconsole to change the volume statuses to
get things fixed.
And failed in that....
So the current status is something to work from this point, not for
error/problem analysis I am afraid. Sorry for that.
Current output is:

*list media
Automatically selected Catalog: MyCatalog
Using Catalog "MyCatalog"
Pool: Default
No results to list.
Pool: Daily
+---------+-----------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
| MediaId | VolumeName                  | VolStatus | Enabled |
VolBytes      | VolFiles | VolRetention | Recycle | Slot | InChanger |
MediaType | LastWritten         |
+---------+-----------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
|       1 | Bacula_Daily-Dinsdag-Tape   | Used      |       1 |
     1 |        0 |      518,400 |       1 |    0 |         0 | DDS-4
   | 2009-02-04 00:06:15 |
|       2 | Bacula_Daily-Woensdag-Tape  | Used      |       1 |
9,911,688,192 |       11 |      518,400 |       1 |    0 |         0 |
DDS-4     | 2009-02-05 00:04:59 |
|       3 | Bacula_Daily-Donderdag-Tape | Used      |       1 |
9,929,106,432 |       11 |      518,400 |       1 |    0 |         0 |
DDS-4     | 2009-02-06 00:07:28 |
|       5 | Bacula_Daily-Maandag-Tape   | Recycle   |       1 |
     1 |        0 |      518,400 |       1 |    0 |         0 | DDS-4
   | 2009-02-03 00:04:08 |
+---------+-----------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
Pool: Weekly
+---------+--------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
| MediaId | VolumeName               | VolStatus | Enabled | VolBytes
    | VolFiles | VolRetention | Recycle | Slot | InChanger | MediaType
| LastWritten         |
+---------+--------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
|       4 | Bacula_Weekly-WeekTape-1 | Recycle   |       1 |
  1 |        0 |    1,641,600 |       1 |    0 |         0 | DDS-4
| 2009-01-24 00:05:15 |
|       6 | Bacula_Weekly-WeekTape-2 | Used      |       1 |
9,849,305,088 |       11 |    1,641,600 |       1 |    0 |         0 |
DDS-4     | 2009-01-31 00:04:56 |
+---------+--------------------------+-----------+---------+---------------+----------+--------------+---------+------+-----------+-----------+---------------------+
Pool: Scratch
No results to list.
You have messages.

So basically mediaID 2,3,6 seem to be ok. The rest is borked up by now.

Regards
Olaf


John Drescher wrote:

On Tue, Feb 10, 2009 at 10:13 AM, Olaf Zevenboom <olaf AT artefact DOT nl> 
wrote:


Dear list,

Bacula : strange things happening at night

After running successfully (besides weird verify issues)  Bacula (2.4.1
on Debian Etch) has gone fubar.

We have a single slot tapedrive with 2 schedules. Each session contains
3 jobs: full backup, verify, backup catalog
There is a daily schedule (run on Monday-Thursday) (4 tapes) and a
weekly schedule run on Fridays (3 tapes). All full backup using a single
tape per day/session.

A full cycle would consist of:
daily cycle (4 days/tapes)
weektape #1
daily cycle (recycling the 4 tapes)
weektape #2
daily cycle (recycling the 4 tapes)
weektape #3
And rerun the full cycle (recycling the 3 weektapes too)

And at that point things did go horribly wrong:
Bacula never created (auto label) weektape #3 but instead:
- volume "Bacula_Weekly-WeekTape-1" was set to recycle
- Bacula wanted to write to that tape and not the blank tape (never
used) in the drive (was meant to become "Bacula_Weekly-WeekTape-3")
This sounded strange to me as the retention time was 19 days and it was
only 14 days after the volume Bacula requested was written to.
Error report: Please mount Volume "Bacula_Weekly-WeekTape-1" or label a
new one

As it was a very inconvenient moment for me at that time and I did not
see a simple solution I deciced to kill all three running jobs.
At this point I was under the impression that I had to look into the
issue but that the other schedule would keep running. I was wrong.

At Monday night I came across the following situation:
- Three jobs started as scheduled
- error report: waiting. Cannot find any appendable volumes. Please use
the "label"  command to create a new Volume
The correct (Monday) tape was inserted and mounted but for some reason
the volume was not re-used although it this schedule cycle had already
had been run 3 times successfully!
Taking a closer look revealed: (running status all in bconsole)
- according to the scheduling overview the jobs were executed to use the
proper volumes
- according to the running job it was looking for a volume without a name
- the volume in use was the volume "Bacula_Weekly-WeekTape-3". How
strange! I canceled all jobs for the weekly (Friday) schedule on Sunday
as that volume was never created and Bacula wanted to use volume
"Bacula_Weekly-WeekTape-1" instead (see above).
After some manual pruning and updating volumes suddenly Bacula wanted
volume/tape Tuesday. At this time it was way beyond 00.00 so it actually
was Tuesday but the jobs were executed to use volume/tape Monday.
For some reason (it ran towards 03.00 am) I wanted some sleep and I
decided to kill the scheduled jobs. After(!) I did this for some reason
the verify job ran (although canceled) and reported many new file (my
verify issue reported earlier).

Either I am doing something terribly wrong or I ran into a serious bug I
suppose. The in use volume looked a bit like a ghost job which I (and
others as I saw on this list) have seen before. No idea how to get rid
of these things besides restarting the director. Any suggestions on that
subject please?

If I am doing something wrong, can it be related to the use of a counter
(which is not very well documented in the docs)? What am I doing wrong?

I am quite stuck here. Please help



Post the output of

list media

John




-- 
John M. Drescher

------------------------------------------------------------------------------
Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
software. With Adobe AIR, Ajax developers can use existing skills and code to
build responsive, highly engaging applications that combine the power of local
resources and data with the reach of the web. Download the Adobe AIR SDK and
Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
_______________________________________________
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>