Data Compression (Archiving Data)

The amount of time to schedule is problem dependent. The sbatch flag -t 240 specifying 240 minutes of scheduled time was pedagogically supplied and is problem dependent. Your own job may need significantly less time, or potentially more.

Note: the htc partition has a max walltime of 4 hours (240 minutes) to increase the walltime beyond 4 hours add the -p general flag example: sbatch -p general -t 300

Overview

How to tarball data using SBATCH with and without compression depending on the data type.

Without compression (good for binary data)

sbatch -t 240 --wrap="tar cvf mytarball.tar paths/to/be/tarred/"

The above command will submit a command to a compute node in the htc partition that requests 1 core for 4 hours (240 minutes). The command will create the uncompressed archive mytarball.tar that contains the contents of the paths specified.

With compression (good for ASCII data)

sbatch -t 240 --wrap="tar czvf mytarball.tgz paths/to/be/tarred/"

The above command will submit a command to a compute node in the htc partition that requests 1 core for 4 hours (240 minutes). The command will create the gzipped compressed archive mytarball.tgz that contains the contents of the paths specified.

Additional Help