Hi,
I'm administering a condor pool with multiple workstations and two dedicated machines (32 cores each). I wanted to jobs to be executed on the dedicated servers instead of workstations so I added this to the servers configs
IsServer = True
STARTD_ATTRS = $(STARTD_ATTRS) IsServer
and this to the negotiator config
NEGOTIATOR_PRE_JOB_RANK = (IsServer =?= True && RemoteOwner =?= UNDEFINED) + \
RemoteOwner =?= UNDEFINED
However when I send dummy sleep jobs to the cluster they get executed on the servers and workstations. How can I debug this? Does it matter that the machines use Dynamic slots? Does LoadAvg is taken into account in choosing slots?
Thanks in advance!
Best,
Chris