GUI - remote desktop
Remote desktop protocol (RDP) provides a remote fully graphical linux environment on the cluster
SUMMARY
Graphical user interfaces (GUIs) enable users to compute on the cluster using little or no command line tools. These graphical access options come in two forms:
Remote desktop sessions
FastX desktop client
ThinLinc desktop client
Web interface
FastX web client
Using web applications such as
JupyterHub
Jupyter Notebooks,
WebMO and
other web gateways
The steps below outline how users can get remote desktop access to our HPC cluster. Please note that you would need to use a VPN to access the RDP services from off-campus networks.
RDP using StarNet FastX
A new remote desktop service (StarNet's FastX3) is available to provide a fully graphical Linux environment on the HPC. StarNet's FastX3 is a tool for remote desktop access to Linux systems. We have a FastX server running on the cluster and users can connect to the cluster using a
FastX desktop client or
Web client
A FastX3 interactive session can be disconnected from without closing the session, allowing users to leave applications running and resume their sessions from other locations. More information on this utility can be found on the StarNet website: https://www.starnet.com/fastx
Desktop Client
Here are steps as adapted from University of Utah's HPC center
Download the latest FastX client for your OS.
Start the FastX client on your desktop (Note: If you are connecting using a VPN, be sure to start the VPN prior to opening the FastX client.)
Use the "+" found in the upper right corner to define a server.
You need to provide a name, the host and a user (your HPC user name) data.
Servers can be edited (clicking on the right "pencil") or deleted by clicking on the X-icon on the right edge.
You can create several entries (i.e. different hosts & user names)
Once you have servers defined you can select one of them and login by clicking on them.
When you do so, two windows will open. In the top window, you will be asked to enter your password. As soon as you are logged, the password window will disappear and the main window will appear. To open a new session on this server, click on the "+" symbol in the top window.
A new window will pop up. You need to either take one of the options given in the window or add a command. When you do this the session becomes defined in this window, along with the start and last connect time, and a new window opens for your session. Your choice depends on your needs. There are both full desktop (XFCE desktops) and different xterm/terminal options.
You can open a terminal or run any graphical applications from the graphical desktop session.
If you want to finish the session, you can close this window. Please do this for any session that you are done with to ensure there are sufficient resources available to meet the active user needs. If you want to keep the session in order to return to it, right click on the session name in the list and select the pause ("||" button) You can close the FastX window listing the sessions at this point, and this closes your connection on the server. To reconnect, either from the same desktop or another, start FastX, select and login to the same server, and you should see your saved session. Right click on the session, and select the "play" button. All windows you had open when you disconnected will appear.
Please be sure to close any sessions you no longer need to keep the number of sessions at a minimum.
Web Client
The steps to access the cluster using a web client are the same as those outlined for the desktop client
Open a web browser on your local computer and point it to
hpc.cofc.edu:3300orhttps://hpc.cofc.edu/rdpFill out the proceeding pages with your user name, server name and login credentials
While the FastX web client is very convenient, it does pose security concerns. To prevent these security issues from being exploited, users should make sure
they are using it on a secure personal that only they have access to. Please DO NOT use the client on a public or shared computer.
do not save any login credentials after they are done with a session.
Much like other services, you would need to connected to the CofC VPN before accessing these remote desktop services.
RDP using Cendio ThinLinc
We have Cendio Thinlinc remote desktop services running on the cluster to provide users access to a graphical Linux environment. We only have 25 concurrent licenses FastX and 5 concurrent Thinlinc licenses, so please close the remote desktop session and exit as soon as you are finished. Users would need to download and install a Thinlinc Client from the Cendio site. There are Thinlinc Clients for Windows, MacOS and Linux.
We only have 5 concurrent licenses of Cendio Thinlinc, so please close the remote desktop session and exit as soon as you are finished.
Desktop Client
We have a Cendio Thinlinc remote desktop server running on the cluster to provide users access to a graphical Linux environment.
Download and install a Thinlinc Client. There are Thinlinc Clients for Windows, MacOS and Linux. https://www.cendio.com/thinlinc/download
After installing the Thinlinc Client, you can start the application and provide the necessary information to start the remote desktop session.
To ensure optimal usage without consuming a lot of resources on the client as well as server side, we recommend that you make the following changes to the under 'Options'.
After you provide all the necessary information to log in, you will be asked to pick a 'Profile' or desktop manager. To prevent these remote desktop from taking too many resources in the login node, we suggest that you use a clean and lightweight desktop like XFCE.
Once you have picked a 'profile' or desktop manager, you should see a Linux remote desktop environment.
Note: Because we only have 5 concurrent licenses, please close the remote desktop session and exit as soon as you are finished.
Web Interfaces
Some web interfaces enable setting up, running and analyzing HPC from a web browser. Two of those approaches will be discussed below.
JupyterHub
JupyterHub+JupyterLab provide a complete multi-user web interface to the HPC for interactive computing
Please see the page below.
Jupyter Notebooks
While you are encouraged to use JupyterHub for interactive computing because of its convenience, you are welcome to run Jupyter Notebooks as prescribed in the page below
WebMO
Please see the page below
Last updated
