Login Nodes

  1. Usage
  2. Access
  3. Authentication
  4. Nodes
  5. Transfering Data

The login nodes are the front end to the cluster for all our users. They are accessible from on and off campus. They are a shared resource and, as such, it is expected that users do not monopolize them.

Usage

Appropriate uses of the login nodes:

  • Transferring files to and from the clusters
  • Creating, modifying, and compiling code and submission scripts
  • Submitting and monitoring the status of jobs
  • Testing executables to ensure they will run on the cluster’s infrastructure (limited to a maximum of 5 minutes of cpu-time).

Access

The login nodes only allow connections from within the umich network, connections from the public internet are not allowed. To login from off campus you will need to use either the ITS Login Service or the ITS VPN Service.

Authentication

In order to login to the cluster you need a valid UMich uniqname with kerberos password. This is the same account you use to access your email and or wolverineaccess.umich.edu.

Nodes

Primary login node pool biostat-login.sph.umich.edu will log you into one of the login nodes randomly. If you need to use the same node, for instance to use an existing screen session, here is the list of current nodes that make up the login pool.

Hostname Architecture
biostat-login.sph.umich.edu Intel
   
bajor.sph.umich.edu Intel
idran.sph.umich.edu Intel

Transfering Data

To transfer data, or code, to and from the cluster you must use scp or sftp. There are several clients options depending on the platform of your client.

  • scp/sftp are available by default on all linux distributions
  • WinSCP for Windows provides an Explorer like interface that support drag-n-drop
  • Putty for windows is a suite of ssh related apps, most notably the terminal appilication. It also includes utilities for scp/sftp.
  • CyberDuck for MacOSX