[Gluster-users] GlusterFS Preformance

[Anand Avati]

>
> I ran a basic test by writing 1G to an NFS server and this gluster pair:
> [root at glust1 ~]# time dd if=/dev/zero of=/mnt/glust2_nfs/nfs_test
> bs=65536 count=15625
> 15625+0 records in
> 15625+0 records out
> 1024000000 bytes (1.0 GB) copied, 1718.16 seconds, 596 kB/s
>
> real    28m38.278s
> user    0m0.010s
> sys     0m0.650s
> [root at glust1 ~]# time dd if=/dev/zero of=/mnt/glust/glust_test bs=65536
> count=15625
> 15625+0 records in
> 15625+0 records out
> 1024000000 bytes (1.0 GB) copied, 3572.31 seconds, 287 kB/s
>
> real    59m32.745s
> user    0m0.010s
> sys     0m0.010s
>

The default block size in which dd operates is 512 bytes, which used
to be fine for local disks long ago, and has always been disastrous
for network filesystems. Please use a bigger block size (like 64KB or
higher) and rerun your tests. We would also suggest you to try our
latest 2.0.3 release.

Avati