After your job has been submitted, you may want to see how fast it is moving up in the queue. And once it starts running, you might then like to peek at its output to confirm that it is doing what you expect it to do. In this section we describe how to do these sorts of things on Frontera through the use of Slurm commands and job attributes.

When we say "output", we are referring mainly to the standard output and error streams. Of course, output files written by the application can be examined as well.

We discussed this in Filesystems, but it is worth reiterating that large or numerous output files should be written to $WORK or $SCRATCH, where it will be easy to inspect them. And in some cases you will get superior performance by writing files temporarily to the local SSD which is mounted as /tmp on each compute node. Files in /tmp can be viewed by ssh'ing to the appropriate node while the job is still running.