Hi Marko,<br><br>The option disable-for-first-nbytes disables write behind for the first n bytes written, where n is the value of the option.<br><br>Also, Can you please send the benmark results for the tests you carried out?<br>
<br>regards,<br><div class="gmail_quote">On Wed, Apr 8, 2009 at 6:18 PM, Marko <span dir="ltr"><<a href="mailto:gluster@sopca.com">gluster@sopca.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Hello,<br>
<br>
To clearify:<br>
* im testing with gluisterfs-2.0.0rc7<br>
* all bricks are on same physical server(Xen guests). It's a testing environment.<br>
<br>
These are a few benchmarks I've done so far:<br>
* time make-many-files #(this is slightly modified version that I've found here:<a href="http://www.linuxinsight.com/files/make-many-files.c" target="_blank">http://www.linuxinsight.com/files/make-many-files.c</a>)<br>
* time dd if=/dev/zero bs=8 count=128000 of=file1MB.bin #(effectively creates lots of small consecutive fops)<br>
* time dd if=/dev/zero bs=4096 count=25000 of=file100MB.bin #(creates optimal transactions from HDDs physical point of view. I have best results here with all configurations)<br>
* time cp -a 0 1 2 /tmp #(/tmp is mounted as tmpfs; 0 1 2 are directories created by "make-many-files" )<br>
* time rm 0 1 2 -fr<br>
<br>
I wish GlusterFS team provided simmilar set of tests so one can measure his performance in a way that can be compared to results from others. I think it would be a great value to all GlusterFS users and developers. I think that to create basic set of these tests is a trivial task( maybe just use mine :D ).<br>
<br>
Below I attached my configuration. Without write-back translator I get better results in most of the tests.<br>
I can't understand why write-back has such a bad impact on performance(being a performance *booster*).<br>
I've also noticed that TCP packets are much lower than MTU in first benchmark. Meaning write-back doesn't optimize writes.<br>
Can you explain that?<br>
Can someone help me to get high performance with AFR?<br>
<br>
Regards,<br>
Marko<br>
<br>
<br>
#------------- configuration ---------------------<br>
########## server ###########################<br>
volume posix-brick<br>
type storage/posix<br>
option directory /srv/gluster<br>
end-volume<br>
<br>
volume lock-brick<br>
type features/posix-locks<br>
subvolumes posix-brick<br>
option mandatory-locks on<br>
end-volume<br>
<br>
volume server<br>
type protocol/server<br>
option transport-type tcp/server<br>
subvolumes lock-brick<br>
option auth.addr.lock-brick.allow *<br>
end-volume<br>
<br>
<br>
<br>
########## client ###########################<br>
<br>
volume brick1<br>
type protocol/client<br>
option transport-type tcp<br>
option remote-host gluster-host1<br>
option remote-subvolume lock-brick<br>
end-volume<br>
<br>
volume brick2<br>
type protocol/client<br>
option transport-type tcp<br>
option remote-host gluster-host2<br>
option remote-subvolume lock-brick<br>
end-volume<br>
<br>
volume AFR<br>
type cluster/replicate<br>
subvolumes brick1 brick2<br>
end-volume<br>
<br>
volume wb<br>
type performance/write-behind<br>
subvolumes AFR<br>
option flush-behind on<br>
option window-size 1MB<br>
option aggregate-size 512KB<br>
end-volume<div><div></div><div class="h5"><br>
<br>
<br>
_______________________________________________<br>
Gluster-users mailing list<br>
<a href="mailto:Gluster-users@gluster.org" target="_blank">Gluster-users@gluster.org</a><br>
<a href="http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users" target="_blank">http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users</a><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Raghavendra G<br><br>