This would be very safe to do. You can most likely update to 2.3.3 with no problems as well. Running jobs should continue merrily along in either case. Only bug fixes have been done from 2.2.4 to 2.2.7 so in this case you shouldn't need to wait for a drained node to update but if you feel safer doing that a 2.2.4 slurmd will work just fine with a 2.2.7 slurmctld.
Danny On 03/13/12 07:55, Loris Bennett wrote: > Hi, > > We are planning to update from version 2.2.4 to version 2.2.7. On > several nodes we have jobs which have been running for over 30 days and > which cannot be restarted from checkpoints. We were thinking to update > our two master nodes, put all the compute nodes on drain, and update > each compute node once it is drained. > > Is it possible to update in this way without interruption? > > Cheers > > Loris
