BUG?: NFS client problems with 2.4.22

BUG?: NFS client problems with 2.4.22

[Joshua Weage]

I have just compiled 2.4.22 with the NFS inode deadlock patch applied
from the 2.4.23-pre tree.

Configuration is SMP, athlon, bonding, eepro100, e1000, NFS v2 and 3
client.  Pretty basic.

Two nodes use only the eepro100 adapters, two nodes have bonded
eepro100 and e1000 adapters, both show the same issue.

This kernel is now producing "kernel: nfs: server not responding"
"kernel: nfs: server OK" messages at a rate of 5-10 per second.  It
looks like this kernel is ignoring the timeo option specified in
/etc/fstab and setting the timeout very small.  I currently have timeo
set to 40.

I'm using the following options on my nfs mounts:
rsize=8192,wsize=8192,hard,intr,timeo=40

Previous Redhat kernels 2.4.18 to 2.4.20-20 have not displayed this
problem.

Is this a known problem with the 2.4.22 kernel?

Thanks,

Joshua Weage

=====


__________________________________
Do you Yahoo!?
Exclusive Video Premiere - Britney Spears
http://launch.yahoo.com/promos/britneyspears/

Re: BUG?: NFS client problems with 2.4.22

[Bernd Schubert]

> Previous Redhat kernels 2.4.18 to 2.4.20-20 have not displayed this
> problem.
>
> Is this a known problem with the 2.4.22 kernel?
>

Yeah it is and I think its already fixed in 2.4.23-pre8.

If you want to use 2.4.22 (2.4.21 also doesn't have this problem) :

> There should be. Try applying patches number 02, 03 and 04 (in
> that order) from
>
> http://www.fys.uio.no/~trondmy/src/2.4.23-pre7
(citation of a mail Trond Myklebust sent to the NFS-Maillinglist)


Cheers,
	Bernd