[Gluster-users] GlusterFS performance

[Ben Turner]

----- Original Message -----
> From: "Явор Маринов" <ymarinov at neterra.net>
> To: "Nux!" <nux at li.nux.ro>
> Cc: gluster-users at gluster.org
> Sent: Thursday, May 23, 2013 4:41:23 AM
> Subject: Re: [Gluster-users] GlusterFS performance
> 
> Thanks for your reply.
> 
> No matter how many nodes (currently the volume is only with its own node) the
> speed is really slow. For testing purposes, i made a volume with only one
> node, without any replication - however the speed is still ~500kb/s. The
> cloud servers are limited to 30Gbit/s but still the traffic when writing to
> the node is ~500kb/s

I agree that there is a performance problem on FUSE mounts on RHEL 5 clients.  A while back I opened https://bugzilla.redhat.com/show_bug.cgi?id=896314.  Check out the DD numbers where I mount the same volume on a RHEL 5 and a RHEL 6 client:

https://bugzilla.redhat.com/show_bug.cgi?id=896314#c5

Just wildly shooting from the hip I would guess that the problem is in the version of FUSE packaged with RHEL 5, but I don't have any evidence to confirm or deny that.  I have tested with 5.6, 5.8, and 5.9 and I saw the problem across the board.  Is there any way you could switch your clients to RHEL 6 until this gets resolved?  If not maybe try mounting NFS?

-b
 
> i'm using 3.3.1 glusterfsd with kernel 2.6.18-348.el5xen and i need to know
> if the the problem is within the kernel.
> 
> ---
> Find out about our new Cloud service - Cloudware.bg
> Access anywhere. Manage it yourself. Pay as you go.
> 
> Yavor Marinov
> System Administrator
> 
> Neterra Ltd.
> Telephone: +359 2 975 16 16
> Fax: +359 2 975 34 36
> Mobile: +359 888 610 048
> www.neterra.net
> 
> 
> On 05/23/2013 11:36 AM, Nux! wrote:
> 
> 
> On 23.05.2013 08:38, Явор Маринов wrote:
> 
> 
> Hello,
> 
> I'm trying to setup GlusterFS with 2 nodes, but it seems that speed
> is really low. The 3 nodes are based on a cloud service, i'm using 2
> for bricks, and 1 for a client, where the volume is mounted. The
> operating system is CentOS 5.9 x64, and actual problem is that the
> write speed when a copy of file is performed on the client side
> mounted dir, is not going over 500kbytes/sec. Anyone have any
> suggestion where to dig further in order to locate the problem?
> 
> Hello Yavor,
> 
> What kind of setup? Replicated?
> With 2 node replicated setups the speed from a 3rd mount point would
> essentially be 2x slower than with only 1 node, so the question is what kind
> of speed do you get between your nodes?
> Is any of them limited to 10 Mbps by any chance?
> 
> 
> 
> 
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users