Am 20.01.2010 19:02, schrieb Jiann-Ming Su: > On Wed, Jan 20, 2010 at 10:23 AM, pkoelle<pkoelle at gmail.com> wrote: >> Am 19.01.2010 18:30, schrieb Jiann-Ming Su: >>> >>> I ran a simple bonnie++ benchmark on 3.0.0. The read performance >>> through FUSE is horrible. Are there some parameters that can be tuned >>> to improve read performance? >> >> Hi Jiann, >> >> You didn't provide information for anyone to help you. You should at least >> post the exact bonnie++ command line you used and some information about >> your setup (config files, transport, etc.). >> > > Bonnie++ command: ./bonnie++ -u 99:99 -d /var/boot -s 2000 > > The system I'm testing on has 1GB RAM. /var/boot is the glusterfs > mount via fuse. > > The transport is tcp but all on localhost. Here's the glusterfsd.vol: > > volume posix1 > type storage/posix > option directory /mnt/gluster/gfs2 This looks suspicious. Is this a gfs2 mount? what kind of performance do you get wen running bonnie++ directly against /mnt/gluster/gfs2? My setup is vastly different so I can't give direct advice. It should be possible to watch bonnie++ and the glusterfs processes via strace to find out whats happening (especially for the ~8sec read latency). Can you leave out some of the performance/* translators and also try a "real" network interface with lower MTU values to see of that helps? I had terrible results with bigger MTU (6500) and bigger values for net.core.(r/w)mem_max in /etc/sysctl.conf. IOPS where in the range of 5/sec. hth Paul > end-volume > > volume locks1 > type features/locks > subvolumes posix1 > end-volume > > volume brick1 > type performance/io-threads > option thread-count 8 > subvolumes locks1 > end-volume > > volume server-tcp > type protocol/server > option transport-type tcp > option auth.addr.brick1.allow 127.0.0.1 > option transport.socket.listen-port 6996 > option transport.socket.nodelay on > subvolumes brick1 > end-volume > > > Here's the client config: > > # TRANSPORT-TYPE tcp > volume localhost-1 > type protocol/client > option transport-type tcp > option remote-host localhost > option transport.socket.nodelay on > option transport.remote-port 6996 > option remote-subvolume brick1 > end-volume > > volume writebehind > type performance/write-behind > option cache-size 4MB > subvolumes localhost-1 > end-volume > > volume readahead > type performance/read-ahead > option page-count 4 > subvolumes writebehind > end-volume > > volume iocache > type performance/io-cache > option page-size 1MB > option cache-size 1GB > option cache-timeout 1 > subvolumes readahead > end-volume > > volume quickread > type performance/quick-read > option cache-timeout 1 > option max-file-size 64kB > subvolumes iocache > end-volume > > volume statprefetch > type performance/stat-prefetch > subvolumes quickread > end-volume > >
