This assignment will provide practice for various Slurm commands and features such as job arrays and job dependencies. You will write 3 small scripts for this assignment:

  • energize.sh
  • analyze.sh
  • submit.sh

At a high level, submit.sh will be run on a login node (not in a job) and will submit energize.sh and analyze.sh for you. The energize.sh job will generate data, and the analyze.sh job will sort through the new data.

energize.sh

The first script, energize.sh, will be a bash script that gets submitted as a job array in Slurm. It generates data for you. It will print the following information to a file named $SLURM_ARRAY_JOB_ID.$SLURM_ARRAY_TASK_ID.nrg: the $SLURM_ARRAY_TASK_ID, followed by a space, followed by a random integer (which can be obtained with $RANDOM), hence referred to as that task’s “energy.”

Note that Slurm has an option to control the name of the output file–you are not required to use it, but it may be helpful. Search the sbatch man page for --output and filename pattern; you’ll want the symbols %A and %a for this job script.

analyze.sh

The second script, analyze.sh, will be a bash script that gets submitted as a job dependency in Slurm. It will be dependent on the job id from submitting the energize.sh script. It analyzes the data generated by energize.sh. This script will read all the files from a given job array and pipe the output to awk. The output from awk should be a single task id and value pair. This should be the pair that has the smallest “energy” value from all tasks of the array. The analyze.sh script must have no other output.

submit.sh

The third script, submit.sh, will be run on the login nodes (with bash, not sbatch) and submit a job array of size 10 that runs energize.sh, once per array task. It will then parse the job id from running sbatch (search the sbatch man page for “parsable”) and submit a second job that is dependent on the first completing successfully; if the first job fails, the second job should not run. The second job will run analyze.sh. You will need to pass the job id from the first job to the analyze.sh either as a command line argument or an environment variable.

Make sure not to run submit.sh from a node-local directory like /tmp or /dev/shm–if you do so, no outfile will show up since the job will try to write it to the compute node’s storage rather than the storage of the node you’re on.

Submission

A tar file (tape archive), also referred to as a tarball, bundles many files and directories into one archive without compression, preserving folder structure, timestamps, and Unix permissions. Researchers use tar to ship datasets and code as a single, reproducible package, move projects between servers or storage systems, and avoid metadata loss that can break analyses. It also streams well over SSH and fits cleanly into backup workflows, whereas having multiple, smaller files does not. A tar file is characterized by the .tar extension.gzip is a separate compression tool that’s applied on top to give it a .tar.gz or .tgz ending. It’s the same tar bundle, just compressed to save space and transfer time.

To tar and compress with gzip, a command might look like this: tar -czf my_first_tarball.tar.gz file_1.txt file_2.sh img_3.png.

Run man tar to see what those flags mean and how to decompress and extract files.

Turn your files in as a single [lastname].tar.gz on Canvas.

Grading

This assignment is worth 15 points. The following deductions, up to 15 points total, will apply:

Deficiency Points
energize.sh fails to print the correct information to the right file 3
analyze.sh fails to run awk on the correct (and only the correct) *.nrg files 3
submit.sh doesn’t submit energize.sh and analyze.sh jobs appropriately 3
The job ID for the energize.sh job isn’t passed to analyze.sh 3
Incorrect results 3
Your tarball isn’t organized correctly or doesn’t contain each file needed to successfully run 3