[Condor-users] What is necessary to make regular checkpoints in VM universe (using VMware)?

[Rob <spamrefuse@xxxxxxxxx>]

Hi,

I have Linux Master and WindowsXP Pool PCs, all running condor 7.4.2.

I want regular checkpoints to be created, say once very 30 or 60 minutes.

I have now a Condor VM job running for hours and no checkpoint has
yet been created. I monitor this on the master in the directories
   /var/lib/condor/spool/cluster161.proc0.subproc0/
and 
  /var/lib/condor/spool/cluster161.proc0.subproc0.tmp/



On the Linux Master I have following lines in the local config file:

# Checkpoint every 30 minutes on average, with a +-10 minute random
# factor to avoid having many jobs hit the checkpoint server at
# the same time.
PERIODIC_CHECKPOINT = $(LastCkpt) > ( 30 * $(MINUTE) + \
                      ($RANDOM_INTEGER(-10,10,1)) * $(MINUTE) )


In the job submission file I have these lines:
vm_type = vmware
vm_networking = false
vm_checkpoint = true

Nothing particular is set on the WindowsXP pool PC.
Should I also add something there for VMware checkpointing?


Thank you!

Rob.