I will run whatever specific test you would like with Bonnie++, just give me the command line arguements you would like to see. i have each filesystem mounted to /test$filesystem so you can include that if you like. I have never used bonnie++ before.
let me know what you want to have run and i will try to get some results today or tomorrow. thanks David Rees wrote: > On Wed, Mar 19, 2008 at 11:07 PM, dan <[EMAIL PROTECTED]> wrote: > >> CPU e8400 3Ghz Dual Core. >> single 7200rpm 16MB cache 200GB maxtor drive. >> ubuntu 7.10 >> > > You don't mention how much memory you have in the machine... > > >> FILE COUNT >> 138581 634MB average of 4.68KB per file(coped the /etc directory 20 times) >> > > This doesn't look like a large enough data set, unless you are > dropping all caches in between each test. See > http://linux-mm.org/Drop_Caches > You should run `sync; echo 3 > /proc/sys/vm/drop_caches` before each test. > > >> find all files and run 'wc -l'(access speed) (wow zfs-fuse is slow here) >> zfs compression=gzip 9.688 sec >> zfs compression=off 10.734 sec >> *ext3 .3218 sec >> *reiserfs .431 sec >> jfs 36.18 sec >> *xfs .310 sec >> > > I've used jfs before and would have noticed that it performed an order > of magnitude worse than the other filesystems - I have to think that > there is something peculiar with your benchmark. > > >> copy from RAM to disk(/dev/shm -> partition w/ filesystem. bus speed not a >> factor) >> > > Why read from /dev/shm ? Something like this would be better: > > time dd if=/dev/zero of=/tmp/bigfile count=10000 bs=1M > > Adjust count as necessary to ensure that you are writing out > significantly more data than you have available RAM. > > >> issues:jfs and xfs both did write caching and then spent periods catching >> up. >> ext3 1m13s 8.68MB/s >> jfs 3m21s 3.15MB/s >> *reiserfs 20s 31.7MB/s (WOW!) >> xfs 2m56s 3.60MB/s >> zfs (CPU bound) 2m22.76s 4.44MB/s >> > > All of your numbers seem to be very slow. I would expect at least > 25MB/s, probably 50MB/s for ext3, jfs, reiserfs and xfs. > > Could you try running an established disk IO benchmark tool like bonnie++? > > -Dave > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ BackupPC-users mailing list [email protected] List: https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki: http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
