[Gluster-users] Usage Case: just not getting the performance I was hoping for

[David Coulson]

On 3/15/12 9:46 AM, Sean Fulton wrote:
> In a case where four client nodes need equal read/write access to the 
> data, is it better to have four Gluster nodes in a replicated 
> configuration with each mounting the gluster volume locally, or having 
> TWO Gluster server nodes with the four clients mounting the volume 
> from the two servers? In other words, the replication would only touch 
> two nodes instead of four; would that improve performance.
Yes. two replicated nodes is faster for writes than four.
>
> Also, would NFS be better in this case, mounting from just ONE of the 
> server nodes, or using Gluster native client to mount from either of 
> the two server nodes. Or can NFS mount from any/all of the gluster 
> nodes using the gluster NFS server.
Never tested it, but in theory running the native gluster client should 
be faster than NFS - The client will connect to the two nodes 
individually, rather than sending everything through the NFS server, 
then it distributing reads and paralleling writes. Remember, when you 
use the native client it connects to whichever node you specify to get 
volume information, then connects to each brick in the volume directly - 
If there is not a brick for the volume on the server you tell it to 
connect to, it won't send any IO to it. That is a major distinction 
between native client and NFS.

A better comparison would be NFS to remote node vs. NFS to local node 
(but with no bricks on local node). I use the latter extensively, since 
it's less work than trying to deal with NFS redundancy.