The Fox Cluster's Interactive Machines

Fox has four machines for interactive use. These can be used to run calculations interactively, outside the queue system. It is also possible to run interactive jobs via the queue system, either from a login node or an interactive machine.

The four machines each have 128 CPU cores, 1024 GiB of memory, 6 TiB NVMe scratch disk and InfiniBand interconnect. They mount the same project and home areas as login and compute nodes. See more about Fox hardware.

The machines are named int-1, int-2, int-3 and int-4 and are only available from a Fox login node or another interactive machine. In the future, we hope to make them available directly from outside Fox.

To log in to one of them, first log in to Fox (ssh ec-<username>@fox.educloud.no or similar), and then use

ssh int-<N>

Once logged in, you can run terminal-based computations interactively on the command line, start programs in the background with &, or run programs with a graphical user interface (GUI). If you want to run GUI programs, you will want to use the -X or -Y switch when logging in (both to Fox itself and to int-<N>). The machines also have the queue system commands installed, so it is possible to submit jobs from them.

The same scientific software available on the rest of Fox is available on the interactive machines. See our sections on building scientific software, running scientific software, Easybuild, Python and R.

The machines have a fast local disk /localscratch, which can be used as a scratch area. Everyone can write to the area, but you can only delete your own files and directories. We encourage everyone to create a directory with your username and keep your files there to avoid file name collisions with other users:

mkdir /localscratch/ec-bhm
cd /localscratch/ec-bhm
chmod go-rwx  # To keep your files private

Usage Policy

Please delete files in /localscratch that you no longer need, otherwise the disk will eventually become full, and no-one will be able to use it any more.

Since the machines are a share resource, it is possible to overload them with work, making them unusable or even crashing them. So please do not start too many processes or use too much memory on them. Use top or a similar command to see how much is used and what else is running on the machine.

It is a good idea to select the machine with the least resource load; see the message on screen when you log in (or do cat /etc/motd).

We currently have no disk quotas or auto-deletion set up for this disk -- and would like to keep it that way if possible -- but that depends on users "cleaning up" after themselves. Be considerate to other users and their needs. Please practice Kardemommeloven (English: Law of Cardamom).