Backup Central

Hi,

I have a bacual-dir with ~250 clients. And each client has its own catalog.

When bacula-dir start, bacula-dir checks each database (~10s per client = ~45 minutes !!!)
Meanwhile, the director is not accessible...

How can I bypass that or reduce that ?

Thanks in advance,
--
Julien

I have a bacual-dir with ~250 clients. And each client has its own catalog.

When bacula-dir start, bacula-dir checks each database (~10s per client =
~45 minutes !!!)
Meanwhile, the director is not accessible...

How can I bypass that or reduce that ?


I would try speeding up your database server perhaps by putting the
databases on an SSD or an array of them or at minimum a RAID10 of fast
harddrives.

John

------------------------------------------------------------------------------
For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
http://p.sf.net/sfu/Boundary-d2dvs2
_______________________________________________
Bacula-users mailing list
Bacula-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Le 6 avril 2012 18:10, John Drescher <drescherjm < at > gmail.com ([email]drescherjm < at > gmail.com[/email])> a écrit :
> I have a bacual-dir with ~250 clients. And each client has its own catalog.

When bacula-dir start, bacula-dir checks each database (~10s per client =
~45 minutes !!!)
Meanwhile, the director is not accessible...

How can I bypass that or reduce that ?



I would try speeding up your database server perhaps by putting the
databases on an SSD or an array of them or at minimum a RAID10 of fast
harddrives.

John


I have installed a Debian server Squeeze with a xen kernel (2.6.32-5-xen-amd64) and on a RAID10 (PERC H700NV with 6*600 SAS 15k).
Then, I have installed FreeBSD 9.0 RELEASE servers on 2 VM :
 - 1 VM for bacula-dir 5.2.6
 - 1 VM for PostgreSQL 8.4.11

And installed FreeBSD 9.0 RELEASE server on an independent server (bacula-sd 5.2.6) with zfs partitionning for the storage.

Do you think that the configuration may slow ?

Do you think if I tuning databases with :
http://www.newitperson.com/2011/01/reduce-bacula-dbcheck-time/
CREATE INDEX File_JobId_idx ON File(JobId);
CREATE INDEX File_PathId_idx ON File(PathId);
CREATE INDEX File_FilenameId_idx ON File(FilenameId);
CREATE INDEX Path_PathId_idx ON Path(PathId);
CREATE INDEX Job_FileSetId_idx ON Job(FileSetId);
CREATE INDEX Job_ClientId_idx ON Job(ClientId);
CREATE INDEX File_FilenameIdPathId_idx ON File(FilenameId,PathId);

Or with :
http://www.bacula.org/manuals/en/catalog/catalog/Catalog_Maintenance.html#SECTION00291000000000000000
CREATE INDEX file_jobid_idx on file (jobid);
CREATE INDEX file_jpf_idx on file (jobid, pathid, filenameid);

it will be faster?


Le 6 avril 2012 18:10, John Drescher <drescherjm < at > gmail.com ([email]drescherjm < at > gmail.com[/email])> a écrit :
> I have a bacual-dir with ~250 clients. And each client has its own catalog.

When bacula-dir start, bacula-dir checks each database (~10s per client =
~45 minutes !!!)
Meanwhile, the director is not accessible...

How can I bypass that or reduce that ?



I would try speeding up your database server perhaps by putting the
databases on an SSD or an array of them or at minimum a RAID10 of fast
harddrives.

John

On Mon, Apr 9, 2012 at 4:39 PM, Julien S <jusouschi < at > free.fr> wrote:
Do you think if I tuning databases with :
http://www.newitperson.com/2011/01/reduce-bacula-dbcheck-time/
CREATE INDEX File_JobId_idx ON File(JobId);
CREATE INDEX File_PathId_idx ON File(PathId);
CREATE INDEX File_FilenameId_idx ON File(FilenameId);
CREATE INDEX Path_PathId_idx ON Path(PathId);
CREATE INDEX Job_FileSetId_idx ON Job(FileSetId);
CREATE INDEX Job_ClientId_idx ON Job(ClientId);
CREATE INDEX File_FilenameIdPathId_idx ON File(FilenameId,PathId);

Or with :
http://www.bacula.org/manuals/en/catalog/catalog/Catalog_Maintenance.html#SECTION00291000000000000000
CREATE INDEX file_jobid_idx on file (jobid);
CREATE INDEX file_jpf_idx on file (jobid, pathid, filenameid);

it will be faster?


Neither of these will effect the problem you are trying to solve.

John

------------------------------------------------------------------------------
For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
http://p.sf.net/sfu/Boundary-d2dvs2
_______________________________________________
Bacula-users mailing list
Bacula-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Perhaps by distributing the load over several director ?

Le 9 avril 2012 22:42, John Drescher <drescherjm < at > gmail.com ([email]drescherjm < at > gmail.com[/email])> a écrit :


Neither of these will effect the problem you are trying to solve.

John