[Beowulf] diskless cluster nfs

Chris Samuel csamuel at vpac.org
Tue Dec 7 15:23:00 PST 2004


On Wed, 8 Dec 2004 09:23 am, Josh Kayse wrote:

> Master node: system drive is 1 36GB SCSI drive
>                      /home raid5 5x 36GB SCSI drives

What distro, kernel and filesystem are you using ?

> all of the nodes are diskless and get their system from the master
> node over gigabit ethernet.

Is your network infrastructure capable of jumbo frames ?

> The engineers use files over nfs for message passing, and no, they
> will not change their code to mpi even though it would be an
> improvement in terms of manageability and probably performance.

You have my considerable sympathies. :-)

> Basically, my question is:  what are some ways of testing the
> performance of nfs ande then, how can I improve the performance?

A good start is the NFS Performance HOWTO at:

 http://nfs.sourceforge.net/nfs-howto/performance.html

A lot of folks report that XFS makes a superior underlying FS for NFS than 
ext3 and we're planning to use that on our next NFS server we build.

> Thanks for any help in advance.
>
> PS: nfs mount options: async,rsize=8192,wsize=8192,hard
>        file sizes: approx 2MB

Out of interest, what does /proc/mounts say when you do and when you don't 
specify the rsize and wsize when mounting the filesystem ?

cheers,
Chris
-- 
 Christopher Samuel - (03)9925 4751 - VPAC Systems & Network Admin
 Victorian Partnership for Advanced Computing http://www.vpac.org/
 Bldg 91, 110 Victoria Street, Carlton South, VIC 3053, Australia

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.scyld.com/pipermail/beowulf/attachments/20041208/cc738f1b/attachment.bin


More information about the Beowulf mailing list