This repository contains scripts and configuration for periodically submitting HTCondor sleep jobs from an AWS-based access point (AP) that flocks to CHTC. The jobs serve as a lightweight, ongoing test to verify that job flocking between the AWS AP and CHTC is functioning correctly.
sleep.sh— Executable bash script that sleeps for 6 seconds (the job payload)sleep.sub— HTCondor submit file that queues 100 instances ofsleep.shcron/aws-flocking-ap-test— Cron job file that submits the sleep jobs once per hourcron/htcondor-nightly-update— Cron job file (root) that reinstalls HTCondor nightly at 0:30 AM
On the AWS access point (logged in as a user with sudo privileges), clone this
repository into the ec2-user home directory and install the cron files:
# Clone the repository
[ec2-user@ec2 ~] $ git clone https://github.com/CHTC/aws-flocking-ap-test-scripts.git
# Install the cron job for periodic sleep job submissions
[ec2-user@ec2 ~] $ crontab aws-flocking-ap-test-scripts/cron/aws-flocking-ap-test
# Install the root cron job for nightly HTCondor updates
[ec2-user@ec2 ~] $ sudo crontab -e
# Confirm cron jobs were installed
[ec2-user@ec2 ~] $ crontab -l
[ec2-user@ec2 ~] $ sudo crontab -lThe first cron job will run condor_submit once per hour as ec2-user, submitting
100 sleep jobs to the HTCondor pool. The second cron job will run as root at 0:30 AM daily,
updating HTCondor using dnf --assumeyes update condor.
For instructions on configuring an EC2 instance as an HTCondor access point that flocks to CHTC, see: https://github.com/CHTC/htcondor-access-point-config