Introduction to Condor for users¶

Condor is a job submission system that creates a high throughput computing environments.

Getting started¶

After gaining ssh accesh to a condor frontend, a user can see what resources are available by running

condor_status

and check the status of the queue by running

condor_q

In order to submit jobs, one needs to create a submit script that defines the requirements and a few options. A simple example to get started would be

Example job submission script helloworld.submit¶

executable     = helloworld.sh
universe       = vanilla 

should_transfer_files = YES
when_to_transfer_output = ON_EXIT
stream_output = true

transfer_input_files = helloworld.dat

request_cpus  = 1
request_memory  = 8000

requirements   = (target.Arch == "X86_64")

input          = /dev/null
output         = out                
error          = error             
log            = log                                                    

queue                  

Example additional file helloworld.dat¶

Hello World!

Example executable script helloworld.sh¶

#!/bin/bash
echo "----------------------" 
hostname
echo "----------------------" 
date
echo "----------------------" 
echo "Sleeping 20s" 
sleep 20
echo "----------------------" 
cat helloworld.dat

After creating these 3 files, the job would be submitted by running

condor_submit helloworld.submit

Checking status¶

To see details about the status of each job in the queue, one would run

condor_q -b

When a job is running, it is also possible to access the node it is using, by identifying the jobid with condor_q and then running

condor_ssh_to_job jobid

More Information¶

http://research.cs.wisc.edu/condor/manual/v7.4/ref.html