Frequently Asked Question
This article covers essential PBS commands used to submit, monitor, and manage jobs in an HPC environment.
1. Submitting a Job: qsub
The `qsub` command is used to submit a job to the PBS scheduler. This command takes the PBS script, which contains job instructions, and sends it to the queue for execution.
qsub myjob.pbs
In this script:
- Job resources (nodes, memory, walltime) are specified using PBS directives.
- PBS processes the script and assigns the job an ID for tracking purposes.
2. Checking Job Status: qstat
Once a job is submitted, you can monitor its status with the `qstat` command. This command provides a list of all jobs along with their status, such as running (R), queued (Q), or completed (C).
qstat
3. Displaying Job Information: qstat -f
The `qstat -f` command provides detailed information about a specific job, such as resource usage, submission time, and job attributes. This is useful for debugging or tracking resource consumption.
qstat -f
This provides verbose output with resource usage (CPU, memory) and other metadata.
4. Deleting a Job: qdel
The `qdel` command is used to cancel a job from the queue or terminate a running job. You’ll need the job ID that was assigned when the job was submitted.
qdel job_id
This stops the job immediately and removes it from the queue. It is useful when a job is running too long or consuming excessive resources.
5. Example: Submit and Monitor a Job
Please note that you need to submit the *.pbs file, not your actual code file (e.g., *.py). For more details on preparing your PBS script, refer to the Introduction to PBS.