Re: XFS NFS Fileserver, kernel: nfs: server server not responding, still trying

[Juri Haberland <juri@xxxxxxxxxxxxxx>]

Russell Cattelan wrote:
> On Sat, 2003-11-08 at 02:49, Juergen Sauer wrote:
>> Am Freitag, 7. November 2003 21:12 schrieb Juri Haberland:
>> > Juergen Sauer wrote:
>>  
>> > I think this is completely a NFS problem so you better might ask on the
>> > NFS ML.
>> > Or what reasons do you have to believe that it's a XFS problem?
>> 
>> On the NFS List they belive it's a XFS-related problem.
>> I posted this to the NFS List also.
>> Because the problem was arising on a new kernel
>> on the server (2.4.22-xfs is now back on 2.4.20-xfs)
>> and a new kernel on client 2.4.22-xfs, which is hardware
>> tied to .22.

> Do the nfs people have any reason to claim this is an XFS
> bug? or is it just a case of "blame a bug".

I had a look at the nfs list and actually no-one said it isn't a NFS
problem. One said that there was an unusual high number of NFS
retransmissons and he suggested to change some NFS parameters. The other
suggested using NFS over TCP...

In the mean time Jürgen seems to have found the real culprit:
The NVIDIA network driver...

Juri