Hi,

27.11.2008 11:31, Boris Kunstleben wrote:
> Hi,
> 
> i am doing exactly that since last Thursday. 
> I have about 1.6TB in Maildirs and an huge number of small files. I have to 
> say it is awfull slow. Backing up a directory with about 190GB of Maildirs 
> took "Elapsed time: 1 day 14 hours 49 mins 34 secs".
> On the other hand i have a server with Documents and images (about 700GB) 
> took much less time.
> All the Servers are virtuall Enviroments (Virtuozzo).
> 
> Any Ideas would be appreciated.

Looks like the catalog database is the bottle-neck here. All the files 
need to be added to it.

There are two rather simple solutions:
- Don't keep file information for this job in the catalog. This makes 
restoring single mails difficult.
- Tune your catalog database for faster inserts. That can mean moving 
it to a faster machine, assigning more memory for it, or dropping some 
indexes (during inserts). If you're not yet using batch inserts, try 
to recompile Bacula with batch-inserts enabled.

Arno

> Kind regards Boris Kunstleben
> 
> 
> Silver Salonen schrieb:
>> On Thursday 27 November 2008 11:07:41 James Cort wrote:
>>   
>>> Silver Salonen wrote:
>>>     
>>>> On Thursday 27 November 2008 09:50:14 Proskurin Kirill wrote:
>>>>       
>>>>> Hello all!
>>>>>
>>>>> Soon I will deploy a large email server - it will use maildirs and will 
>>>>> be about 1Tb of mail with really many small files.
>>>>>
>>>>> It is any hints to make a backup via bacula of this?
>>>>>
>>>>>         
>>>> I think Bacula is quite good for backing up maildirs as they constist of 
>>>> separate files as e-mail messages. I don't think small files are a 
>>>>       
>> problem.
>>   
>>> I don't think they're a problem either and I also backup a maildir-based
>>> mail server.
>>>
>>> However, one thing you may want to be aware of - unless you take
>>> specific steps to avoid it, the maildirs on tape won't necessarily be in
>>> a consistent state.  Obviously this won't affect your IMAP server - but
>>> it does mean that when you restore, metadata like whether or not emails
>>> have been read or replied to and recently received/sent email won't be a
>>> perfect snapshot of how the mailserver looked at any given point in time.
>>>     
>> And when you have many incrementals in a row while restoring, you end up 
>> seeing many duplicate messages, that have been deleted or moved during these 
>> incrementals.
>>
>> It'll be better in Bacula 3.0, I guess :)
>>
>> --
>> Silver
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Bacula-users mailing list
>> Bacula-users AT lists.sourceforge DOT net
>> https://lists.sourceforge.net/lists/listinfo/bacula-users
>>   
> 
> 

-- 
Arno Lehmann
IT-Service Lehmann
Sandstr. 6, 49080 Osnabrück
www.its-lehmann.de

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users