[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#325117: NFS problem(s) with kernel 2.6



tag 325117 +upstream
thanks

On Fri, Oct 07, 2005 at 09:11:56AM +0200, Ruediger Oberhage wrote:
> > Hi,
> 
> Hello,
> 
> many thanks for your kind reply.
> 
> > is it possible for you to test the 2.6.12 kernel package
> > that has been produced for Sarge. Its available at the
> > following URL as 2.6.12-5.99
> 
> Well, I did try some packages mentioned to be available on
> http://packages.vergenet.net/testing/linux-2.6/
> [linux-image-2.6.12-1-686_2.6.12-5.99.sarge1_i386.deb and
>  dependancies] and they didn't work, either, or more precisely showed
> the same symptom.
> [This and a patch I found for the MSB-problem of the 32bit
>  cookies (or even 64bit cookies without export-option) for which
>  SGI IRIX 6(.5) is notorius for and which I applied to earlier 2.6er
>  kernels seem to indicate, that the problem hasn't vanished in
>  between and is not related to (only) the 32bit nfs-cookie thing.
>  I'm not sure if I mentioned that in the original message.]
> 
> > It would be good to know if the problem was fixed between
> > 2.6.8 and 2.6.12.
> 
> I don't think so (see above).

Yes I agree

> > If not I would recommend starting a dialog with the upstream NFS
> > maintainers, I can point you to the right place.
> 
> That would be nice thank you. I'm willing to try everything that
> I'm carefully guided to :-), as long as my resources allow.
> Since it is important to me for this to work, I'd like to help
> where I can.

As I understand your problem seems to be with the NFS client,
not the NFS server portion of the kernel. The contact for
that is Trond Myklebust <trond.myklebust@fys.uio.no>, you
should also CC linux-kernel@vger.kernel.org. 

If you see anything related to this message in dmsg, send that too.

On the Debian side, it would be good to CC 325117@bugs.debian.org,
to keep this bug up to date. Upstream lives on CC, so it
probably won't drop off in a hurry.

> > If so, we have a starting point to try and isolate the change
> > that resolve the problem. Though it may prove too extensive
> > to be appropriate for backporting to 2.6.8.
> 
> Yes, I do understand this, and I would gladly be willing to
> switch to a newer kernel. 2.6.8 is a non-optimal choice anyway
> in my eyes, being the last kernel which has practically no
> useful (udev) classes but the most general (e.g. the 'dvb' class
> is still missing from its modules/drivers).
> 
> Thus it wouldn't be that hard for me to part with 2.6.8, but
> a transition beyond 2.6.12 (e.g. 2.6.13) with 'sarge' might
> be hard (or impossible?), too, regarding its 'tools' dependancies.
> 
> The most important thing would be, to learn what's going wrong
> with 'nfs', though, I think. At least to me and may be to you, too.

-- 
Horms



Reply to: