On Mon, May 01, 2006 at 10:43:11PM +0100, Alan Brown wrote: > On Mon, 1 May 2006, John Kodis wrote: > > >You have more CPU than I do, but only about half the memory. Since > >you say that the disks are thrashing, I'd guess that the lack of > >memory is more likely to be the culprit than the difference in > >databases. There have, however, been some messages posted here > >indicating that MySQL needs more attention to tuning when backing up > >this many files than a similar Postgres installation does. > > By way of comparison my systems were thrashing badly even with 2Gb of ram > because mySQL wasn't using enough memory. Allowing it to grow to 1Gb ram > solved it almost entirely.
I just have 512 MB of RAM but you made me think about my server parameters. > Adjusting /etc/my.cnf is pretty much a necessity on large database > systems, or mysql will make extensive use of temporary files and things > can grind to a halt quickly. The default settings are only ok for > small systems or for testing. While I used settings from the "large" my.cnf example configuration file and had a slight speedup it still took longer than a few hours so I interrupted. Following to hints from the helpful inhabitants of #bacula on irc.freenode.net I ran a "dbcheck" searching for orphaned rows. Voila, my row count decreased from 21 million to 3 million. I even needed two runs because dbcheck seemed to have an upper limit of 10 million entries to delete. :) I have no idea why my catalog grew like this. I have been restoring it once with bscan because the catalog was lost and I had to recover the information from 20 DDS-3 tapes (not the most interesting tale of my life). I still don't have an idea why the restore took that long on my main server. The very same catalog was imported during an hour on a workstation. Pretty strange. What's also still a bit weird: I have 450,000 distinct files in my backup. Still the `File` table counts 3 million rows. I was told that the number of rows there is independent from the number of times a certain file was backed up. So it's not like 450,000 files x 10 backups = 4,500,000 million rows. Kindly Christoph ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
