hello,
log monitoring revealed that a user processed crashed a few days
ago and left an ooops message along with a zombie user process
(in non-interruptible disk wait).
the machine was still running fine with +2 extra load.
ksymoops output attached along with other details.
regards,
--
Steinar Hauan, dept of ChemE -- [email protected]
Carnegie Mellon University, Pittsburgh PA, USA
>>>>> " " == Steinar Hauan <[email protected]> writes:
> hello,
> log monitoring revealed that a user processed crashed a few
> days ago and left an ooops message along with a zombie user
> process (in non-interruptible disk wait).
Hi,
Did you find any other NFS related messages in your syslog? Short of
a bug in truncate_inode_pages(), I really can't see how this could
happen. Certainly whatever is was should at least have triggered a
couple of NFS warnings...
Cheers,
Trond
On 20 Jan 2002, Trond Myklebust wrote:
> >>>>> " " == Steinar Hauan <[email protected]> writes:
>
> > hello,
> > log monitoring revealed that a user processed crashed a few
> > days ago and left an ooops message along with a zombie user
> > process (in non-interruptible disk wait).
>
> Did you find any other NFS related messages in your syslog? Short of
> a bug in truncate_inode_pages(), I really can't see how this could
> happen. Certainly whatever is was should at least have triggered a
> couple of NFS warnings...
there are no nfs error or warning messages at all on either the server
or the client for one week prior to the crash. this includes logs created
by kern.* and *.debug entries in syslog.conf.
regards,
--
Steinar Hauan, dept of ChemE -- [email protected]
Carnegie Mellon University, Pittsburgh PA, USA