SUMMARY Further Information: NFS problem

From: <Andrew.Raine_at_mrc-dunn.cam.ac.uk>
Date: Tue, 17 Jul 2001 11:32:06 +0100 (BST)

Dear Managers,

I think I have resolved the NFS problem that I reported here earlier.
Trond Myklebust on the NFS list (and one of the authors/maintainers of
the Linux NFS code) asked me a question which caused me to look at the
configuration of the switch port that the client machine was connected
to.

The client machine was plugged in to an 8-port Netgear switch, which
was in turn plugged into a port on one of our main Cisco Catalyst 5000
switches. The port on the Cisco was configured thus:

        set spantree portfast 2/3 enable

        Warning: Spantree port fast start should only be enabled on
        ports connected to a single host. Connecting hubs,
        concentrators, switches, bridges, etc. to a fast start port can
        cause temporary spanning tree loops. Use with caution.

I've corrected this, and also moved the client machine to its own port,
and the problem seems to have disappeared.

Original question appended.

Andrew

--
Dr. Andrew Raine, Head of IT, MRC Dunn Human Nutrition Unit, 
Wellcome Trust/MRC Building, Hills Road, Cambridge, CB2 2XY, UK
phone: +44 (0)1223 252830   fax: +44 (0)1223 252835
web: www.mrc-dunn.cam.ac.uk email: Andrew.Raine_at_mrc-dunn.cam.ac.uk
> Dear Managers,
> 
> Further to my earlier posting about problems mounting a Tru64 disk on a
> RedHat Linux workstation:  Larye Parkins pointed out that Linux NFS
> used to be limited to version 2, which might be expected to gave
> problems on files/partitions larger than 2 GB.
> 
> However, the documentation with my user's RH7.1 machine claims that
> NFSv3 is supported, and tests on a 2.6 GB text file show that his
> machine can indeed read the whole file, while a 6.2 machine with an
> older version of mount truncates it at 2 GB.
> 
> So I'm still stuck!  Any further suggestions?  One further symptom is
> that once once the RedHat machine has triggered the lock-up, other
> (e.g. IRIX, normally unaffected) machines also freeze when trying to
> access the same partition.
> 
> Original question appended:
> 
> Many thanks,
> 
> Andrew
> 
> --
> Dr. Andrew Raine, Head of IT, MRC Dunn Human Nutrition Unit, 
> Wellcome Trust/MRC Building, Hills Road, Cambridge, CB2 2XY, UK
> phone: +44 (0)1223 252830   fax: +44 (0)1223 252835
> web: www.mrc-dunn.cam.ac.uk email: Andrew.Raine_at_mrc-dunn.cam.ac.uk
> 
> > Dear All,
> > 
> > I have a user, running their own Linux (7.1) workstation, who is having
> > problems NFS-mounting from my ES40.  Other clients in the Unit, running
> > IRIX or Tru64 have had no problems mounting from the same server, but
> > they probably don't put such a heavy load on the system.
> > 
> > Server:
> > 
> > 4-processor ES40, Tru64 UNIX V5.1 (Rev. 732), also running TruCluster
> > Server V5.1 (Rev. 389), but currently the only machine in the "cluster".
> > 
> > Exported disks are a single 180GB RAID5 volume in an HSZ80, attached by
> > SCSI (not fibre)
> > 
> > Client:
> > 
> > 800MHz PIII running RedHat 7.1
> > 
> > Symptom on Client:
> > 
> > During intensive use, access to the NFS volume freezes.  In
> > /var/log/messages I see things like -
> > 
> > Jun 28 08:32:48 bioinf kernel: nfs_notify_change: attr=1134896, fattr=1146880??
> > Jun 28 08:37:59 bioinf kernel: nfs: server cluster.mrc-dunn.cam.ac.uk not responding, still trying
> > 
> > Symptom on server:
> > 
> > in /var/adm/messages I see things like -
> > 
> > Jun 27 19:31:50 beta vmunix: NFS server: stale file handle fs(2633,839795) file 433277 gen 32771
> > Jun 27 19:31:50 beta vmunix:  RFS3_WRITE, client address = 193.60.81.170, errno 70
> > 
> > A reboot clears the problem, but it recurs pretty quickly once the user
> > starts work again.  The same code works fine on files NFS mounted from
> > other makes of server (as far as I can gather - we can't easily test
> > this as we don't have a non-Tru64 server with enough disk space to test
> > this).  I don't think the individual files are disappearing from under
> > the user.
> > 
> > Does anyone have any clues as to what's going on, or where I can look
> > next?  I'm new to Tru64 (I'm more familiar with IRIX).
> > 
> > I will, of course, summarize.
> > 
> > Many thanks,
> > 
> > Andrew
> 

Received on Tue Jul 17 2001 - 11:31:42 NZST