Connect to a Cluster

  • Article

Applies To: Windows Compute Cluster Server 2003

The Compute Cluster Job Manager can be installed on any 32-bit or 64-bit workstation that meets the operating system requirements of Compute Cluster Server 2003 for remote workstation computers. The Job Manager is installed by default on cluster head nodes.

Using the Job Manager, a cluster user can connect to any cluster to which the user has network access and permissions. A CCS cluster is identified by the name of the head node.

If the Job Manager is run locally on the head node itself, the Job Manager will connect by default to the CCS Job Scheduler on the local host.

To connect to a cluster

  1. In the Start menu, point to All Programs, click Microsoft Compute Cluster Pack, and then click Compute Cluster Job Manager.

  2. In Job Manager, click File, then Connect to a Cluster. Specify the cluster head node to which you want to connect:

    1. If you are using Job Manager from a workstation, select Another computer and enter the name of the cluster head node you want to use.

    2. If you are running Job Manager on the head node itself, select Local computer.

  3. Click OK.

Note

By default, in subsequent sessions, the Job Manager attempts to connect to the most recent cluster specified.

You may specify a non-default cluster at the command prompt in one of two ways:

  • Specify it using the /scheduler:<host> option for any CLI command, where host is the name of the head node of the cluster that you want to use. For example: job submit /scheduler:myhost myapp.exe

    Note

    This option works only on a command-by-command basis.

  • On your client node (not the head node, unless it is also the client node), create the environment variable CCP_SCHEDULER and set it to the specified cluster. This will direct all CLI commands to the specified cluster. For example:
    set CCP_SCHEDULER=myhost