Updated 2019-04-30

Log on to CUI Cluster

What is the CUI Cluster?

  • CUI stands for Controlled Unclassified Information
  • The CUI headnode is a secure headnode that runs RHEL7


Before logging into the CUI cluster, you must be connected to the PACE VPN. Use pace.vpn.gatech.edu for the VPN address

Important Differences with CUI Cluster

  • The system will log you out after 1 hour of inactivity.


  • Output doesn't count as activity, there must be input
  • If you try 5 times and fail to log in (shouldn't happen with SSH keys), you are locked out for 30 min
  • The PACE VPN doesn't do split tunneling, so once you connect you cannot RDP into the machine connected to the VPN

Log on to CUI Cluster - Linux / Mac

  • Reminder: You must connect to the PACE VPN, using pace.vpn.gatech.edu for the VPN address
  • For more information on SSH, check out the Logging on with SSH guide
  • The headnode address is login-c1.pace.gatech.edu
  • To login, open up a new terminal and simply use this template:
ssh someuser3@login-c1.pace.gatech.edu
# Use your GT username in place of "someuser3"
# Must be on GT campus internet or connected to Gatech VPN
  • You should be logged into your account on a headnode, you can now interact with the cluster, submit jobs, make files and do anything else you would on a standard headnode

Log on to CUI Cluster - Windows

  • You can still use ssh on windows, but you will need a client such as PuTTy
  • Again, You must connect to the PACE VPN before logging on to the CUI cluster. Use pace.vpn.gatech.edu for the VPN address
  • For more information on loggin in with PuTTy, check ou the guide on Loggin in to the cluster
  • In the HostName bar, enter the address for the CUI Cluster: someuser3@login-c1.pace.gatech.edu. Use your GT username instead of "someuser3"
  • Click open and in the terminal type in you GT password when prompted for password.

Storage on the CUI Cluster

  • The way storage is set up is slightly different on the CUI Cluster


  • Intead of a data directory there is the cui_data directory
  • There is no scratch directory

Differences in RHEL7

  • The main difference to notice in RHEL7 is the modules that are available
  • On RHEL7, some modules may not yet be installed on RHEL7, but the flipside is RHEL7 has newer versions of some modules
  • Therefore, PBS scripts you have made on RHEL6 will remain the same for RHEL7 except for modules
  • Solution : While logged in to the RHEL7 headnode, use module avail <software name> to make sure the modules you are loading in your PBS script exist and are the right version on RHEL7. If you find that more up to date versions of your modules exist on RHEL7, feel free to update the modules in your PBS script
  • Adjust the module portions of your PBS scripts accordingly to what modules are currently available.