Re: kernel OOPS: 2.4.18, nscd, nfsd

[Richard Ems <r.ems.home@gmx.net>]

Neil Brown wrote:

> On Wednesday June 19, r.ems.home@gmx.net wrote:
> > Hi all!
> >
> > Two kernel Oopses in short time (22:35:59 and 22:50:00). But the computer was still alive until 00:00:00, where the daily cron jobs are started and then ... kernel panic, LED's where blinking   :(
> >
> > kernel is 2.4.18, from SuSE's k_deflt-2.4.18-174 package (2.4.19-pre10aa2)
> >
> > Please CC to r.ems@gmx.net, I'm not on the linux-kernel mailing
> > list.
>
> Would I be right is surmising that you are exporting an ISO filesystem
> over NFS??  That would be the second Oops in as many days with that
> scenario.
>
> If that is the case, then I'm afraid that I cannot point you to any
> fix, though exporting with "no_subtree_check" may reduce the incidence.
>
> NeilBrown

No, no ISO fs exported.

My /etc/exports:

# cat /etc/exports
/home \
 diablo(rw,no_root_squash) @linux(rw,root_squash) @unix(rw,root_squash) mtgvaio1(ro,root_squash) @cluster01_hosts(rw,root_squash)
/tmp \
 diablo(rw,no_root_squash) @linux(rw,root_squash) @unix(ro,root_squash) mtgvaio1(ro,root_squash) @cluster01_hosts(rw,root_squash)
/usr/local \
 diablo(ro,no_root_squash) @linux(ro,root_squash) @unix(ro,root_squash) mtgvaio1(ro,root_squash) @cluster01_hosts(ro,root_squash)

and

# egrep " /home | /tmp | /usr " /proc/mounts
/dev/vg01/home /home ext3 rw 0 0
/dev/vg01/tmp /tmp ext3 rw 0 0
/dev/vg01/usr /usr ext3 rw 0 0

all exported filesystems are ext3 over LVM.

(I stopped a while ago exporting ISO fs's over nfs, I'm now copying the distro DVD (SuSE) on a HD ... Any "good" solution for exporting ISO fs's over NFS ? the problem is always when you need the
cdrom/dvd drive and want to umount the NFS exported cdrom/dvd, then ...)

What about the first Oops ?

Thanks, Richard



--
Richard Ems
... e-mail: r.ems@gmx.net
... Computer Science, University of Hamburg

Unix IS user friendly. It's just selective about who its friends are.