Notes on Accessing the IDEA Cluster

If you have been granted access to the IDEA Cluster. If you do not have access and would like to make a request, fill out our form.

 

  1. You will be using your RCS credentials
  2. To quickly test that you have access, go to https://lp01.idea.rpi.edu/rstudio-ose/ while on the RPI network.
  3. To access RStudio Server only, use one of the following links
    http://lp01.idea.rpi.edu/rstudio-ose/  (Classes and Research)
    http://lp01.idea.rpi.edu/rstudio-ose-3/ (Research)
  4. To access JupyterHub (EXPERIMENTAL!): http://lp01.idea.rpi.edu/jupyter-gpu
  5. To access a node via the shell (e.g. ssh in Linux and Mac or Putty in Windoze), there are two steps:
    1. First access the "landing pad" (Linux commands shown):
      $ ssh username@lp01.idea.rpi.edu
    2. Now, access one of the compute nodes; for example, idea-node-02
      $ ssh username@idea-node-02

       

    3. You will now be in your home directory on idea-node-02 (which happens to have the GPUs, etc.). R and Python are installed on all of the nodes, but most people doing heavy Py work do it on Node 2 so they can also use the GPUs.

FYI:

  • The cluster uses a storage server, so you will access your files in the same way across all the machines.
  • idea-node-01 is used heavily by students in various IDEA courses, including IDM and DARL, using RStudio Server
  • idea-node-03 is just like idea-node-1, but is "reserved" for use by researchers
  • idea-node-02 doesn't provide access to RStudio, but does have R, Py, and some extra libraries to support GPU work
  • idea-node-04 supports a number of virtual servers
  • Note that sftp is also available for transferring files from e.g. your laptop onto the Cluster.