Bacula-users

Re: [Bacula-users] Fwd: SD Bug "User specified spool size reached"

2012-05-28 18:52:10
Subject: Re: [Bacula-users] Fwd: SD Bug "User specified spool size reached"
From: Tilman Schmidt <t.schmidt AT phoenixsoftware DOT de>
To: bacula-users <bacula-users AT lists.sourceforge DOT net>
Date: Tue, 29 May 2012 00:49:33 +0200
Am 28.05.2012 19:21, schrieb John Drescher:
> On Mon, May 28, 2012 at 9:38 AM, Sean Cardus <scardus AT zebrahosts DOT net> 
> wrote:

>>> On a Bacula installation with a CentOS 6.2 (64 bit) server
>>> running Bacula 5.2.6 director and SD with an HP LTO3 autoloader
>>> and backing up some twenty-odd machines of various platforms,
>>> all the jobs report "User specified spool size reached"
>>> much too early, currently after 1.45 GB instead of the really
>>> specified size of 50 GB. Example log excerpt:
>> [..]
>>> Scanning the log it looks like the storage daemon decreases its
>>> spool size limit every time a backup job crashes. The last time
>>> it used the full 50 GB was here:
>>
>> I've been seeing exactly the same behaviour here (CentOS 4.9, Bacula SD 
>> 5.0.3) for quite some time.  However, until I saw your email I didn't 
>> correlate the max spool size reducing with each time a job crashes.  I can 
>> also confirm that whenever a job crashes, the max spool size drops.
>>
>> The only workaround I can see at the moment is restarting the SD...
> 
> I would be interested in seeing why jobs are crashing and possibly
> submit bug reports for that. I have run 30 thousand or so jobs at work
> and I do not see very many job crashes.

That one is easy, and not very interesting actually. The two most
frequent causes are:
a) network outages (we do some backups over WAN VPN connections)
b) clients crashing during backup, either coincidentally or because the
additional load from the backup pushes an already flakey machine over
the brink.
Neither is Bacula's fault, and therefore something that would warrant a
Bacula bug report I think.
(Although I do sometimes wish Bacula would handle clients disappearing
in the middle of a backup more gracefully.)

To sum it up: there can be legitimate reasons for a job not terminating
successfully, but that shouldn't lead to a diminishing maximum spool
size IMHO.

T.


Attachment: signature.asc
Description: OpenPGP digital signature

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users