On Thu, Dec 02, 2004 at 04:54:10PM -0000 or thereabouts, Peter Gronbech wrote:
> Since I upgraded to 2_2_0 my MDS service on the ce has crashed twice, at
>
> about 2:45am and then again at 1400 today.
> Is it known to be less stable than the previous release??
> Pete
I don't know that is known but certainly from observation this is
the case though this may be just that gstat is doing a much better
job of giving a global picture of the problem. It appears to have
got very bad in the last 2-3 weeks....
People are currently examining core dumps to understand the conditions.
A simple noddy cron along the lines of
/sbin/service globus-mds status >& /dev/null
if [ $? = "0" ] ; then
/sbin/service globus-mds restart
fi
ran every 5 minutes may be the best for now.
Steve
>
> ----------------------------------------------------------------------
> Peter Gronbech Unix Systems Manager Tel No. : 01865 273389
> Department of Particle Physics Fax No. : 01865 273418
> University of Oxford,
> Keble Road, Oxford OX1 3RH, UK E-mail : [log in to unmask]
> ----------------------------------------------------------------------
--
Steve Traylen
[log in to unmask]
http://www.gridpp.ac.uk/
|