Re: [Bacula-users] performance problems backing up ocfs2 clusters
Hi folks, thanks for your suggestions, I tried your tar suggestion and indeed it turns out that transfer rates drop to the dozens of kb/sec in one special directory stored on the ocfs2 filesystem. I'm now in contact with the ocfs2 devs on the users list to see if they have any suggestions. All the best, Uwe -- uwe.schuerk...@nionex.net phone: [+49] 5242.91- 4740 fax:-9722 Hauptsitz: Avenwedder Str. 55, D-33311 Guetersloh, Germany Registergericht Guetersloh HRB 4196, Geschaeftsfuehrer: Horst Gosewehr NIONEX ist ein Unternehmen der DirectGroup Germany www.directgroupgermany.de -- Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are powering Web 2.0 with engaging, cross-platform capabilities. Quickly and easily build your RIAs with Flex Builder, the Eclipse(TM)based development software that enables intelligent coding and step-through debugging. Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] performance problems backing up ocfs2 clusters
Hello folks, we're experiencing massive problems backing up an ocfs2 cluster filesystem mounted on SLES 10 SP2 machines located on a shared SAN storage). The cluster has 8 members, and we've already tried certain mount options (noatime et al.) in an attempt to improve performance, however bacula's transfer speeds drop down into the double kb / sec digits when it encounters directories which contain many small files (say about 20,000 per dir or so). The backup server in question is a Compaq dl360 (CentOS 5.2 on x86_64) which provides usually excellent performance when backing up other non-ocfs2 fileystems, so I was wondering if anyone has developed a successful strategy for backing up ocfs2 clusters (we're talking 160gb here, no it's not really a huge amount of data). Bacula Version used is 2.2.8 compiled from source, the file daemon on the target machine claims it's version Version: 2.4.3 (10 October 2008) x86_64-unknown-linux-gnu suse 10 Thanks in advance for any ideas or suggestions on this matter. All the best, Uwe -- uwe.schuerk...@nionex.net phone: [+49] 5242.91- 4740 fax:-9722 Hauptsitz: Avenwedder Str. 55, D-33311 Guetersloh, Germany Registergericht Guetersloh HRB 4196, Geschaeftsfuehrer: Horst Gosewehr NIONEX ist ein Unternehmen der DirectGroup Germany www.directgroupgermany.de -- Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] performance problems backing up ocfs2 clusters
On Wed, 04 Mar 2009 11:24:22 +0100, Uwe Schuerkamp wrote: Hello folks, we're experiencing massive problems backing up an ocfs2 cluster filesystem mounted on SLES 10 SP2 machines located on a shared SAN storage). The cluster has 8 members, and we've already tried certain mount options (noatime et al.) in an attempt to improve performance, however bacula's transfer speeds drop down into the double kb / sec digits when it encounters directories which contain many small files (say about 20,000 per dir or so). look at Spool Attributes = yes (or even Spool data http:// www.bacula.org/en/rel-manual/Configuring_Director.html). you can also try to make little test how long it takes to tar the same files. something like tar -cf /dev/null path (ok i'm not an advanced tar user ...) . maybe ocfs2 is the bottleneck as it (maybe) has to communicate the reads with the other nodes? - Thomas -- Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users