Gluster users - Jul 2011 - Optimal config for 2 replicated webservers with read performence

Hi!

With my setup, when I copy a 500MB file from the Gluster mount to local 
disk it takes about 6-10 times longer than disk to disk. When I 
investigate it further, I see huge amounts of data between my two 
gluster servers. Is this the way gluster should work? I thought it just 
checked meta-data on the other server, it should not copy the entire 
file from the other server?

My client config is the same on both servers:

volume global1
     type protocol/client
     option transport-type tcp
     option remote-host 10.10.10.20
     option transport.socket.nodelay on
     option transport.remote-port 6996
     option remote-subvolume brick2
end-volume

volume global2
     type protocol/client
     option transport-type tcp
     option remote-host 10.10.10.21
     option transport.socket.nodelay on
     option transport.remote-port 6996
     option remote-subvolume brick2
end-volume

volume mirror-0
     type cluster/replicate
     subvolumes global1 global2
end-volume

volume writebehind
     type performance/write-behind
     option cache-size 10MB
     subvolumes mirror-0
end-volume

volume iocache
     type performance/io-cache
     option cache-size 1GB
     option cache-timeout 5
     subvolumes writebehind
end-volume

volume quickread
     type performance/quick-read
     option cache-timeout 5
     option max-file-size 1024kB
     subvolumes iocache
end-volume

volume global
     type performance/stat-prefetch
     subvolumes quickread
end-volume


- Marius