nfsrvstats.srvrpc_errs rapidly increasing

[Heinrich Rebehn]

Mohan Srinivasan wrote:
>>>Are you using NFS/TCP ? Can you force the mount to NFS/UDP ?
>>
>>Yes, we use TCP. It is strongly recommended for multispeed networks and 
>>we did have problems with retransmissions using UDP.
> 
> 
> I'm not disputing the merits of NFS/TCP. I suggested this merely as
> a workaround, and to see if the bug in question is causing the hangs.

I will try it on a single machine when the problem occurs again. ATM, 
everything runs flawlessly :-)
> 
> 
>>Wouldn't it be better to fix the bug? Is there a problem report on this?
> 
> 
> Of course it would be better to fix the bug. But until someone fixes it...
> 
> There's no problem report on this. 

I wonder if it will ever get fixed then..

Since i am neither familiar with kernel sources nor with the internals 
of NFS: Could you show me, where the bug can be traced, i.e. at which 
point one can detect that the stream is out of sync? I would like to 
insert another log() in order to see, if
- the error occurs at all
- is in any way related to our problems.

--Heinrich