Computing Cluster FAQ

Computing Cluster FAQ

This page is intended to address common questions and issues users have and encounter while using UAlbany’s High Performance Computing (HPC) cluster.

 

Who can access the cluster?

UAlbany students, faculty, and staff can access the computing cluster. Any faculty members working with outside individuals (e.g. a professor from another institution) will need to request a guest account for that outside individual.

How much storage do I get for the cluster?

An individual user’s home directory gets 10GB of storage. Users who are a member of a lab directory get 10TB to share among all lab directory members. Only UAlbany faculty and staff are allowed to request a lab directory and update directory membership; students cannot request a lab directory or access to their PI’s existing lab directory on their own initiative.

Why can’t I access the cluster?

If you have been added as a user to the cluster and are unable to log in, try checking the following:

  1. Are you entering your alphanumeric UAlbany NetID? It would look something like ab123456

    1. If you are entering your NetID, are you entering it in in all lowercase, as above?

  2. If you are not physically on the UAlbany campus, are you using our VPN? Instructions for the VPN can be found here: Virtual Private Network (VPN)

If you are a student and require access, please contact your PI or supervisor and have them submit a ticket to ITS.

I want to run a script but it is not working. Why is this?

Please check the following:

  1. Is your home directory full? If so, are there any files that you can remove to clear some space?

Best practice for any analysis or computing work is to store any files relevant to your work (data or inputs, scripts) within and that your outputs are written to be stored within a lab directory, since that affords you much more storage to work with than your individual home directory.

  1. If you have submitted a job through our SLURM job scheduler, have you:

    1. allocated enough computing resources for your job?

    2. checked that your script was able to finish in time? The maximum time limit by default is 14 days. If you require more than this, please contact ITS

More information about SLURM can be found on our wiki here: How-to: Scheduling via SLURM

A side note on the /tmp directory

Please do not store any files in the /tmp directory, such as training scripts or outputs. As mentioned above, these are much better utilized within a lab directory