The portal cluster

The portal cluster is a general purpose cluster of servers which is the gateway into the CS Dept. computing environment. Logins are load balanced among the servers. Upon an 'ssh' to, your login will be placed on the least loaded of the cluster's servers (portal[01,02, etc.]). Each system is running CentOS Linux and has plenty of memory, CPU cores, and scratch (/localtmp) space. Your home directory is automatically mounted and available when you login.

You must use your CS userid (identical to your UVA userid) and password to 'ssh' to portal. For example:

ssh -l abc1de Alternatively, ssh

(Note that if you are opening a terminal on your Mac or PC, your username on the Mac or PC may be different from your CS domain userid. So be sure to include your CS/UVA userid on the 'ssh' command line).

Most popular software is already available on portal, so you do not need to install it. See: Software Modules

You can use portal to perform many interactive tasks, like editing files, compiling code, testing code, etc. If you intend to run long, compute intensive programs, you should use other Computing Resources which are available for general login or through the job scheduler (see Scheduling a Job using the SLURM job scheduler).

We also have a cluster of servers running Ubuntu Server called uportal. These are also in a load balanced cluster, so upon ssh login, you will be placed on the least loaded server. You can login to these identically to the portal cluster. For example:


See Linux Server Access for more information about accessing CS Dept servers.

  • compute_portal.txt
  • Last modified: 2023/04/10 21:52
  • (external edit)