Homework - tutorial 4

Prepare and execute a simple workflow on the aurora cluster:

1) Prepare the workflow: The workflow should consist of a series linux commands in the form of a bash script. The purpose of the workflow is to gather various information about the computer.  As a minimal information, the workflow should obtain the hostname of the machine, the "uptime" information, disk space usage and the number of entries in the /etc and /usr/bin directories. Also add a few minutes ”sleep”. Save the workflow as SLURM job descripton script (e.g. homework4.sh), execute it on the login or "frontend" and save the output to a file (e.g. hw4-script-frontend-output).

2) Execute the same workflow on the cluster worker nodes using the SLURM system:  Start with a fresh login to aurora then record your session using the logging feature of the screen utility: start up a screen session with logging enabled (screen –L), create a slurm submission script based on the homework4.sh script (call the slurm script as homework4-slurm.sh) and submit several jobs with sbatch, list the queue status, list the job status, cancel some of the jobs, play with the slurm commands, then exit screen

3) Place all homework files into a dedicated CLEAN directory within your home area (e.g. HW4-final). The directory should contain the workflow script, the slurm submission script, the execution output from the frontend run,  the screenlog.0 file and some of the SLURM session logs.

4) Once everything is done and the dedicated HW4 directory is available, send me the full path of that directory.

Submission and communications via Canvas only.