[Gluster-devel] Infiniband slow throughput

[Anand Avati]

>
> Gluster:
>
> [root at cluster0 src]# glusterfs -f /usr/local/etc/glusterfs/client.vol -n
> brick1 /share
> [root at cluster0 share]# sync ; time (dd if=/dev/zero of=/share/foo.bar
> bs=4k count=512K ; sync )
> 524288+0 records in
> 524288+0 records out
> 2147483648 bytes (2.1 GB) copied, 222.344 seconds, 9.7 MB/s
>
> real    3m42.384s
> user    0m0.168s
> sys     0m1.016s
>
>
>
Which version of glusterfs are you using? Please load write-behind in the
client volume file (over unify). Have you tried ib-verbs transport (instead
of ib-sdp) ?

avati