[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Trying to be too clever

On Thu, Jun 09, 2005 at 12:13:46PM +0100, Dominik Kasprzyk wrote:
> Hello all
> I'm trying to upgrade a cluster that's running Redhat/Condor6.6 to Debian/Condor6.7.6.  Everything is nearly ready for the reinstall and I just had a question that I started to wonder about when I learnt the head node (schedd,negot,collec) could be powered down for a kernel upgrade without actually losing any jobs.  So, I was wondering could I completely reinstall the head node and have it come back up and automatically carry on picking up what all the nodes are doing?  If not, is there anything (simple?) I can do to achieve this effect?  If there is, would it work between the latest stable release and 6.7.6?

No. With 6.6, any interruption to the condor_schedd causes running jobs to
be interrupted.

With 6.7, jobs submitted with JOB_LEASE_DURATION can be reattached
to, but for this upgrade you're out of luck.