Accessing

• Connecting to the Cluster

  The only direct access to the cluster is through the submit node, brood00.umiacs.umd.edu. The only means to access this node is through SSH. If you're accessing the cluster from a UMIACS or CS workstation, then you already have SSH installed.

  Indirectly, jobs might get run on the cluster if they are submitted through Condor. Our Condor pool includes some other computational clusters, and many workstations as well(bug cluster currently is not included in condor). If you do not have a need to run parallel jobs on the hive cluster, then the preferred means for accessing the cluster is through Condor. You do not need an account on the Hive cluster, in this case -- just an account on a host where you can submit Condor jobs (such as the openlab cluster).

• Getting an Account

  If you need to run parallel jobs on the Hive cluster, you need an account there. To get one, go to our account request form. In the "PI's email:" field you should put down the Primary Investigator that you are associated with, or your advisor. In the "Comment:" field you should indicate that you need a hive cluster account. This will get you an account in our OpenLab, which includes generic UMIACS resources, as well as the Hive cluster.