Using the 244 Lab

The Computer Science department Linux lab in IACC 244 is a 40 computer lab (lab00-lab39) running Ubuntu 24.04 x86_64 GNU/Linux with the MATE Desktop Environment. The lab is for learning Linux, running programs designed to be run in a Linux environment, and course projects. The lab can be accessed remotely using an SSH client (PuTTY for Windows or simply 'ssh' from Mac, Linux or BSD) or sitting down at the workstations in QBB 244. The lab is open 24/7 but does require key-card access via your NDSU ID card. Contact your instructor or advisor if you need access.

• Hadoop is offline for security reasons.
• Spark is up.
• Lab13 is down.
• Lan21 is down.

The door to the lab is locked 24 hours a day. Your NDSU Bison Card grants you access if you're part of the following groups:

• Faculty/Staff
• Enrolled in a course that requires use of the lab
• Request lab access from the CS Department.

You can also access the computers in the 244 lab remotely via SSH.

Your username and password is that of your Bison Account (formally your NDSU Account) and NOT that of your NDUS account. It's format is first.last(.and maybe a number).

If you have forgotten your password, you can reset it here. If ITS moves the link (they do rather often), ask at the help desk in QBB 150. I think the current rules are 14 chars, 1C, 1N, &, 1SC.

You can use SSH to log in to the computers in the lab. The computers are at lab00.cs.ndsu.nodak.edu, lab01.cs.ndsu.nodak.edu, and so on up through lab20.cs.ndsu.nodak.edu.

• In Linux/Mac, you can just use the 'ssh' command from a terminal.
• For Windows, you may want to use PuTTY as an SSH client.
  • An even better Windows SSH client is MobaXterm, although the free version does have some annoyance features to encourage you to upgrade to the paid version.
  • Windows 10 (after the April 2018 update) now has a SSH Client, although it's not enabled by default on older releases.
  • Windows Power Shell also supports SSH but its cut & paste functionality is not… optimal!

If you wish to upload/download files to your home directory (assignments, notes, etc.), you can do so using SCP. On Windows, WinSCP works well. You simply connect, and drag your files over to copy them to the lab. A command-line SCP application comes default with most Linux distributions. Tutorial on using SCP from the Command Line. Also FileZilla is quite common.

The lab has a number of programs installed for you to use. Additional programs can be requested by emailing support@cs.ndsu.edu with details on the program and licensing that you'd like installed. Everything you need for your coursework is already installed.

• Obviously you can't have SUDO privlidges to install software yourself.
• If using Python be sure to use Python Virtual Environments so you can install needed packages. Here's a tutorial. The system admin will NOT install packages for you.

The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using simple programming models. The Hadoop head node is zoidberg.cs.ndsu.nodak.edu

Apache Spark™ is a multi-language engine for executing data engineering, data science, and machine learning on single-node machines or clusters. The spark head node is spark.cs.ndsu.edu

Labs 02-05 contain NVIDIA 24GB Tesla M40 GPU's (These are past current Tensorflow version support) and Labs 06-09 contain NVIDIA 16GB Tesla P100 GPU's.

These machines have Nvidia Cuda 11.5.1 and cudnn8 installed. For Python users, see https://wiki.cs.ndsu.nodak.edu/doku.php?id=deptlab:python_virtual_env

Use responsibly.

MPICH is installed in the lab. MPICH is a high-performance and widely portable implementation of the MPI-3.1 standard from the Argonne National Laboratory. It supports different computation and communication platforms including commodity clusters, SMPs, massively parallel systems, and high-speed networks.

Your home directory (/home/<username>) is stored on a Network File Server (NFS), and is mounted on all 40 machines, so your files are available everywhere. Quotas are not enforced, just try to be responsible with space usage.

If your program or research experiment requires high levels of file IO, we ask that you do not use the NFS server for this (you will destroy performance across the lab, plus it will be slow for your experiment). If you require faster storage, contact support@cs.ndsu.edu and we can grant you access to local storage on each box.

We're happy to provide the 244 lab for student use; lots of creative things can come out of having access. But, we have some restrictions:

1. Don't do anything illegal. Just don't.
2. Log off when you're done. Don't just lock the screen unless you're returning shortly. Too many people doing that can waste a lot of resources and in-lab classes have priority access.
3. Don't power off the workstations. Other users may be logged into the machines or want remote access.
4. Don't unplug monitors, workstations, or other equipment. Students need the lab to function properly when you leave, so please leave the environment like you found it. Especially remember that students may be working on assignments remotely, so unplugging a machine will result in them losing their work.
5. Don't use up all the resources. Again, the lab is shared, so if you're running an experiment that uses up all the CPU/memory on the machines, you are making it difficult for students to complete their work. If this becomes a problem, the Administrator will lock your account. If you are a student and have large CPU/memory/disk requirements for your research, coordinate with your adviser and department Systems Administrator.
6. Don't prop open the lab door unless you have permission (You'd know if you do).

In the lab, you can put files in your public_html directory, they will be served out at http://students.cs.ndsu.nodak.edu/~yourusername/.

You will use this directory for certain courses. You're also welcome to use it for small scale private hosting. Don't post Personal Identifiable Information (PII)!

x2go connections are available to lab01…lab20. Using x2go for Windows

Be sure to log out of your session or you may generate multiple sessions and get locked out. To fix this you will need to log in via SSH and delete old x2go session information from your home directory.

Maximum session idle time is three days, before sessions will be purged.

Note: Some graphical applications do not like to work via x2go. GoogleEarth for example. Also GPU compute applications often run faster via a SSH text connection.

Note: X2go usually works. If it doesn't…

• Try connecting via the command line using SSH to see if you can actually reach the server to which you are trying to connect.
• Be sure your client version of x2go is at least 4.1.0. It if gets too out of date there are handshake issues.
• Check your firewall settings. Sometimes it's set automatically by your anti-virus/malware software.
• You may need to get yourself a proxy service if connecting from outside of the USA. China, Russia, and parts of Africa and India all have had blocked connections.
• You can run x2go in debug mode to see a log file that may provide some clues as to what is wrong on your end. If you don't see debug mode, reinstall x2go.

X2go install on a Mac is a bit different. It requires Xquartz to be installed. https://wiki.cs.ndsu.nodak.edu/doku.php?id=helpdocs:x2go:windows&s[]=x2go#x2go_on_mac It usually works but sometimes it doesn't and neither Google or I have been able to determine why.

x2go is also available for Linux and can sometimes be made to work on a Chromebook.

If you need help using the 244 lab. Contact your instructor or academic advisor.