Can you try with write-behind loaded on client side? ----- Original Message ----- From: "Paolo Pisati" <paolo.pisati at tomato.it> To: "Raghavendra G" <raghavendra at zresearch.com> Cc: "Paolo Pisati" <paolo.pisati at contactlab.com>, gluster-users at gluster.org Sent: Monday, May 25, 2009 1:50:39 PM GMT +04:00 Abu Dhabi / Muscat Subject: Re: gfs vs nfs: abysmal performance difference Raghavendra G wrote: > Hi Paolo, > > what is the client and server configuration you are using? the simplest you could think: client: volume client type protocol/client option transport-type tcp option remote-host 10.0.0.1 # IP address of the remote brick option remote-subvolume brick # name of the remote volume end-volume server: volume posix type storage/posix option directory /usr/src end-volume volume locks type features/locks option mandatory-locks on subvolumes posix end-volume volume brick type performance/io-threads option thread-count 8 subvolumes locks end-volume volume server type protocol/server option transport-type tcp option auth.addr.brick.allow 10.0.0.* subvolumes brick end-volume -- Paolo Pisati Reparto Sistemi --------- ContactLab Via Natale Battaglia, 10 20127 Milano (Italy) Tel. +39 02 28 31 181 Fax +39 02 70 03 02 69. http://www.contactlab.com
