Kernel panic when copying files under NFS

From: Fergus Wilde (fwilde@chethams.org.uk)
Date: Tue May 27 2003 - 04:54:36 EDT


Hello All,

I'm having a consistent problem with an ultra5 acting as an nfs server for a
SuSE linux client:

# uname -a
SunOS ultra5 5.9 Generic_112233-04 sun4u sparc SUNW,Ultra-5_1

# share
- /data rw=linux

This was set up a few weeks ago for regular manual copying of file trees from
the Linux system, largely using cp -Rpu /[directory]. A rough form of
syncing, I know, but adequate for the particular task, particularly as only
human-selected directories need doing. All was sweetness and light. Now,
however, such copy operations consistently cause a kernel panic on the Sun
box, which loops through panic after panic until the Linux client is killed
off.

The share mounts fine (it is set up to mount manually rather than in the
client's fstab) and produces ls outputs, touching of files, and other 'minor'
ops without demur, but goes over when sizeable copies are tried:

/var/adm/messages:May 18 15:18:23 ultra5 ^Mpanic[cpu0]/thread=3000134a580:
/var/adm/messages:May 18 15:20:44 ultra5 savecore: [ID 570001 auth.error]
reboot after panic: BAD TRAP: type=31 rp=2a100656f00 addr=8 mmu_fsr=0
occurred in module "ufs" due to a NULL pointer dereference
/var/adm/messages:May 18 15:21:21 ultra5 ^Mpanic[cpu0]/thread=3000120c7e0:
/var/adm/messages:May 18 15:23:55 ultra5 savecore: [ID 570001 auth.error]
reboot after panic: BAD TRAP: type=31 rp=2a100570f00 addr=8 mmu_fsr=0
occurred in module "ufs" due to a NULL pointer dereference

I think, but am not sure, that things were OK until this patch, part of the
recommended cluster, was loaded:

113278-01 nfs patch

but I freely admit I've no idea if it was that patch or one of the dozens of
others, or something else again. It has nfs in the name. The obvious ploy
would be to back out the patch, but the machine assures me this isn't
possible when a patchrm is tried. I was surprised by this, as I thought the
install_cluster script kept backups around.

Other major disk ops, e.g. big tar / untar operations or compressing, big sftp
copies from the same Linux client, cause no problems at all, and I ran vts
several times on the ultra 5 without any problems being identified. It has a
history of total reliablility. Anyone seen this kind of thing / know what to
do?

Thanks,

-- 
Fergus Wilde
Chetham's Library
Long Millgate
Manchester
M3  1SB
Tel: +44 161 834 7961
Fax: +44 161 839 5797
http://www.chethams.org.uk
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers


This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:28 EDT