[HTCondor-users] setting up dedicated pool for parallel universe

[Kodanda Ram Mangipudi <kodanda@xxxxxxxxx>]

Hi,

I am a newbie for htcondor set-up and administration though I was a user in the past.Â

We are trying to set up a pool of 2 machines each with dual CPUs with 20 cores each. I made an i5 6 core machine as the master, and the other 2 HPC workstations as the nodes.All are running Ubuntu 16.04. The condor installation is from the default Ubuntu repositories installed using apt-get.

We have to run MPI jobs. So, I made the masterÂalso as the DedicatedScheduler. When I submit this simple script (below) the job stays Idle forever.

universe = parallel

executable = /bin/sleep

arguments = 3000

machine_count = 4

log = log

should_transfer_files = IF_NEEDED

when_to_transfer_output = ON_EXIT

queue

I have configured the file following the manual. Below are the config files on master and node01. configuration files of node02 are identical to node01

I am not getting where the problem is and forums and google could not tell me how to get it done so far.Â

Any help is highly appreciated as we are stuck here for a couple of weeks.

master

/etc/condor/condor_config (package manager file UNCHANGED)

-------------------------------------Begin file ----------------------------------

RELEASE_DIR = /usr

LOCAL_DIR = /var

LOCAL_CONFIG_FILE = /etc/condor/condor_config.local

#LOCAL_CONFIG_FILE = $(RELEASE_DIR)/etc/$(HOSTNAME).local

REQUIRE_LOCAL_CONFIG_FILE = false

LOCAL_CONFIG_DIR = /etc/condor/config.d

#LOCAL_CONFIG_DIR_EXCLUDE_REGEXP = ^((\..*)|(.*~)|(#.*)|(.*\.rpmsave)|(.*\.rpmnew))$

use SECURITY : HOST_BASED

#ALLOW_WRITE = *.cs.wisc.edu

##Â FLOCK_FROM defines the machines that grant access to your pool via flocking. (i.e. these machines can join your pool).

#FLOCK_FROM =

##Â FLOCK_TO defines the central managers that your schedd will advertise itself to (i.e. these pools will give matches to your schedd).

#FLOCK_TO = condor.cs.wisc.edu, cm.example.edu

#SEC_PASSWORD_FILE = $(LOCAL_DIR)/lib/condor/pool_password

##Â Pathnames

RUNÂ Â Â= $(LOCAL_DIR)/run/condor

LOGÂ Â Â= $(LOCAL_DIR)/log/condor

LOCKÂ Â = $(LOCAL_DIR)/lock/condor

SPOOLÂ Â= $(LOCAL_DIR)/spool/condor

EXECUTE = $(LOCAL_DIR)/lib/condor/execute

BINÂ Â Â= $(RELEASE_DIR)/bin

LIBÂ Â Â= $(RELEASE_DIR)/lib/condor

INCLUDE = $(RELEASE_DIR)/include/condor

SBINÂ Â = $(RELEASE_DIR)/sbin

LIBEXEC = $(RELEASE_DIR)/lib/condor/libexec

SHAREÂ Â= $(RELEASE_DIR)/share/condor

PROCD_ADDRESS = $(RUN)/procd_pipe

CONDOR_HOST = $(FULL_HOSTNAME)

DAEMON_LIST = COLLECTOR, MASTER, NEGOTIATOR, SCHEDD, STARTD

----------------------------------------------- end of file ----------------------------------------------------

master

File: /etc/condor/config.d/00debconf Â(Edited for configuration)

-------------------------------------Begin file ----------------------------------ÂÂ

DAEMON_LIST = COLLECTOR, MASTER, NEGOTIATOR, SCHEDD

CONDOR_ADMIN = root@localhost

RESERVED_MEMORY =

FILESYSTEM_DOMAIN =

UID_DOMAIN = XXXXXXXXXX

CONDOR_HOST = 10.10.48.81

ALLOW_WRITE = 10.10.48.81, 10.10.48.90, 10.10.48.86

ALLOW_NEGOTIATOR = 10.10.48.81

-----------------------------------end of file ----------------------------------------------

node01

File: /etc/condor/condor_confic (Package manager's copy; Identical to that of master node)ÂÂ

/etc/condor/config.d/00debconf (Edited for configuration)

-------------------------------------Begin file ----------------------------------ÂÂ

DAEMON_LIST = STARTD, SCHEDD, MASTER

CONDOR_ADMIN = root@localhost

RESERVED_MEMORY =

FILESYSTEM_DOMAIN =

UID_DOMAIN = XXXXXXXXX

CONDOR_HOST = 10.10.48.81

ALLOW_WRITE = 10.10.48.90, 10.10.48.81, 10.10.48.86

ALLOW_NEGOTIATOR = 10.10.48.81

# Added: by system admin: For Dedicated scheduler for parallel universe

DedicatedScheduler = "DedicatedScheduler@xxxxxxxxxxx"

STARTD_ATTRS = $(STARTD_ATTRS), DedicatedScheduler

# 2) Always run jobs, but prefer dedicated ones

STARTÂ Â Â Â Â Â= True

SUSPENDÂ Â Â Â Â= False

CONTINUEÂ Â Â Â = True

PREEMPTÂ Â Â Â Â= False

KILLÂ Â Â Â Â Â = False

WANT_SUSPENDÂ Â = False

WANT_VACATEÂ Â Â= False

RANKÂ Â Â Â Â Â = Scheduler =?= $(DedicatedScheduler)

MPI_CONDOR_RSH_PATH = $(LIBEXEC)

CONDOR_SSHD = /usr/sbin/sshd

CONDOR_SSH_KEYGEN = /usr/bin/ssh-keygen

--------------------end of file ------------------------------------------------------------Â

Â

Best regards,

Ram