SSH & Login Node Usage

The most common method of logging into the cluster is using the command-line interface (CLI). If you're using any Linux or Linux-like system (e.g., MacOS, BSD) you probably already have a terminal installed by default. Newer versions of Windows also have a new Linux sub-system so there are also native options to bring up a local terminal.

Logging into klone#

ssh UWNetID@klone.hyak.uw.edu

You log into the klone.hyak.uw.edu cluster above at the terminal using your netID. You will be prompted for your password and 2FA (DUO) authentication. We don't allow ssh keys to the login node since it would be bypassing one of the factors (of 2-factor authentication).

Monitoring & Warnings

Login-node CPU and memory usage are closely monitored by automation. Any egregious use will result in immediate performance throttling, the severity of which escalates exponentially while the offending activity persists. Warnings describing the crossed thresholds & the extent of the throttling will be sent by our monitoring system, Arbiter2, to the email associated with your UW NetID.

Acceptable Uses of the Login Node#

The klone login node (i.e. klone-login01 or klone.hyak.uw.edu) is a resource shared by all cluster users. As such, acceptable uses of this system are very limited:

  1. Downloading to or uploading from the cluster.
  2. File management like moving, copying, or renaming files and directories.
  3. Light text editing with vim or a worse text editor.
  4. Interacting with the Slurm scheduler, i.e. to submit jobs.
  5. Running Hyak-built commands like hyakalloc or hyakstorage.
tip

If your command would run a little slow on an inexpensive laptop from 2002, you shouldn't run it on the login node! Instead, request an interactive job on a compute node.