[gmx-users] interrupting mdrun; queuing

[Martina Bertsch, PhD]

Dear gmx developers and users,

Our administrator is planning to bring down our Linux cluster for 
upgrades later this week. I have an mdrun job that is running on the 
cluster for about a month and a half, and will not be complete for at 
least another month. Is there a special command to terminate an mdrun 
job nicely? Is there a recommended kill command? I know how to restart 
jobs that crashed, but I do not want to destabilize a perfectly running 
job by intentionally crashing it.

 I was also wondering if anybody could share any experience using 
GROMACS on a cluster with the Condor queuing system?

Best regards,
Martina Bertsch, Ph.D.