On Wed, 2012-02-08 at 11:23 -0500, J. Bruce Fields wrote:
> On Mon, Feb 06, 2012 at 07:57:16PM -0500, Trond Myklebust wrote:
> > It is perfectly legal to negotiate up to 2^32-1 slots in the protocol,
> > and with 10GigE, we are already seeing that 255 slots is far too limiting.
>
> Seems like an entirely reasonable change, but just out of curiosity,
> what workload are you using to hit that limit?
>
> (If it's just bulk IO over a local 10GigE network, then surely the
> problem is elsewhere? Even on 10GigE I think you'd need fairly high
> latencies (a little over 200ms, if my math is right?) to keep that much
> data in flight. Or are you testing much smaller operations?)
10GigE + high latencies is exactly where we're seeing the value. Andy
has been working with the high energy physics community doing NFS
traffic between the US and CERN...
--
Trond Myklebust
Linux NFS client maintainer
NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
