[Condor-users] Problems setting up a dedicated cluster

[Rui Peixoto <ruimiguel1988@xxxxxxxxx>]

Hi.

I'm having some problems setting up a dedicated condor cluster to run parallel jobs.

I have 2 machines: condorA(to submit and schedule jobs) and condorB (to execute jobs). The config in condorB is:

DedicatedScheduler = "DedicatedScheduler@xxxxxxxxxxxxxx"

SUSPEND         = Scheduler =!= $(DedicatedScheduler) && ($(SUSPEND))

PREEMPT         = Scheduler =!= $(DedicatedScheduler) && ($(PREEMPT))

RANK_FACTOR     = 1000000

RANK            = (Scheduler =?= $(DedicatedScheduler) * \

                  $(RANK_FACTOR)) + $(RANK)

START           = (Scheduler =?= $(DedicatedScheduler)) || ($(START))

MPI_CONDOR_RSH_PATH = $(LIBEXEC)

CONDOR_SSHD = /usr/sbin/sshd

CONDOR_SSH_KEYGEN = /usr/bin/ssh-keygen

STARTD_EXPRS = $(STARTD_EXPRS), DedicatedScheduler

The problem is, if i submit a simple task like this:

universe = parallel

executable = /bin/sleep

arguments = 5

machine_count = 1

queue

It stays in Idle state forever. condor_q -analyze reports this:

-- Submitter: condorA : <192.168.10.100:34766> : condorA

---

071.000:  Request has not yet been considered by the matchmaker.

 

What am i doing wrong?