Is the RDMA support going to be fixed with 3.4 ?
Regards,
--
Bartek Krawczyk
network and system administrator
___
Gluster-users mailing list
Gluster-users@gluster.org
http://supercolony.gluster.org/mailman/listinfo/gluster-users
-brick1 new-brick2
> 4、gluster volume rebalance VOLNAME fix-layout start
> 5、gluster volume rebalance VOLNAME migrate-data start
> Is there something I missed? Thanks a lot.
I haven't tried that but it seems OK. Just keep in mind that your new
volume won't be replicated but distribut
ing transport rdma on 3.3.0 but
there wasn't any speed difference. I'm not sure there is any
difference in 3.3.1.
Regards,
--
Bartek Krawczyk
network and system administrator
___
Gluster-users mailing list
Gluster-users@gluster.org
http://supercolony.gluster.org/mailman/listinfo/gluster-users
On 19 October 2012 15:00, Bartek Krawczyk wrote:
> in /etc/rdma/rcma.conf:
> IPOIB_LOAD=yes
/etc/rdma/rdma.conf of course
--
Bartek Krawczyk
network and system administrator
___
Gluster-users mailing list
Gluster-users@gluster.or
rfs.
Be careful when you'll be detatching existing volume from GbE peers to
those new IPoIB peers.
And for sure you have to do some reading on InfiniBand in Linux. The
hints above are just a Quick Quick Quick start.
Regards,
--
Bartek Krawczyk
network and
name.
Due to the lack of rdma support in 3.3.x we decided to stick with plain IPoIB.
Regards
--
Bartek Krawczyk
network and system administrator
___
Gluster-users mailing list
Gluster-users@gluster.org
http://supercolony.gluster.org/mailman/listinfo/gluster-users
10,8389 s, 484 MB/s
I did a "sync" between those tests. Funny isn't it? RDMA is much
slower than IPoIB on writing and faster on reading.
I think we'll stick to the IPoIB until something's fixed in glusterfs.
And still - why the results are so similar to Gigabi
