共用方式為

Log On to a Compute Node to Interact with Your Job in Windows HPC Server 2008 R2

  • 發行項

Updated: February 2011

Applies To: Windows HPC Server 2008 R2

This topic describes how a job owner can log on to a compute node in a Windows® HPC Server 2008 R2 cluster to view and interact with the application display. To view the display, the application must be running in a remote or console session (not in session 0, which is the default). You can set a job-level environment variable to have your job run in an existing remote or console session, or you can have a console session automatically created for your application when the job starts. For additional information, see Job or task environment variables for console or remote desktop sessions.

In this topic:

  • Requirements

  • Start a remote session, and then submit your job

  • Submit your job, and then start a remote session

  • Additional considerations and known issues

Requirements

To view the application display for a job that is running on a compute node, you must have the following:

  • You must be added as a user on the cluster.

  • The cluster network must be set up so that you can remotely log on to the compute nodes, either directly or through a gateway server.

  • You must have remote log on permissions on the compute nodes.

    Note
    Remote log on permissions are not automatically granted to cluster users. Contact your cluster administrator to acquire these permissions.

Start a remote session, and then submit your job

The following procedure describes the general steps for running a job in an existing remote session.

To start a remote session, and then submit your job

  1. Establish a Remote Desktop Connection to one or more nodes.

  2. Create your job, and specify the following job properties:

    • Required nodes. Specify the nodes to use for the job. Ensure that you include the nodes to which you have established a remote session in this list.

    • Environment variables. Add the following environment variable:

      HPC_ATTACHTOSESSION=True

      You can use one of the following methods:

      • In HPC Job Manager, in the New Job or Modify Job dialog box, click the Environment Variables tab. Click Add. In Name, type HPC_ATTACHTOSESSION. In Value, type True.

      • In HPC PowerShell, type: set-hpcjob –id <yourJobID> -jobenv HPC_ATTACHTOSESSION=True

      • In a Command-prompt window, type: job modify <yourJobID> /jobenv:HPC_ATTACHTOSESSION=True

  3. Submit your job, and then monitor and interact with the application through the Remote Desktop Connection.

Important
If you log on to the compute node with keyboard, video, and mouse (KVM), or KVM over IP, you must attach your job to the console session, not to a remote session. To attach a job to an existing console session, set the job-level environment variable HPC_ATTACHTOCONSOLE instead of HPC_ATTACHTOSESSION.

Submit your job, and then start a remote session

The following procedure describes the general steps for submitting a job and then starting a remote session to the allocated nodes. When you set the HPC_CREATECONSOLE environment variable, the HPC Job Scheduler Service automatically creates a console sessions for your job when the job starts. When the job ends, the HPC Job Scheduler Service closes the console session.

Important
The automated create console functionality must be enabled on the cluster by the cluster administrator. For more information, see Enable HPC_CREATECONSOLE for GPGPU Jobs in Windows HPC Server 2008 R2.

To submit your job, and then start a remote session

  1. Create your job, and add the following environment variable to the environment variable job property:

    HPC_CREATECONSOLE=True

    You can use one of the following methods:

    • In HPC Job Manager, in the New Job or Modify Job dialog box, click the Environment Variables tab. Click Add. In Name, type HPC_CREATECONSOLE. In Value, type True.

    • In HPC PowerShell, type: set-hpcjob –id <yourJobID> -jobenv HPC_CREATECONSOLE=True

    • In a Command-prompt window, type: job modify <yourJobID> /jobenv: HPC_CREATECONSOLE=True

  2. Submit the job.

  3. After the job starts, view the job to see the list of allocated nodes.

    You can use one of the following methods:

    • In HPC Job Manager, double-click your job in the views pane. In the View Job dialog box, click the Allocated Nodes tab.

    • In HPC PowerShell, type: get-hpcjob –id <yourJobID>|select –property allocatedNodes

    • In a Command-prompt window, type: job view <yourJobID>|find “Allocated Nodes”

  4. Establish a Remote Desktop Connection to one or more of the allocated nodes.

  5. Monitor and interact with the application through the Remote Desktop Connection.

Additional considerations and known issues

Running Excel workbook offloading jobs in a console session

Issue: If a node is restarted while running a workbook offloading job, the auto recovery mechanism might try to restart the Excel instances (independently of a cluster job) if the job owner logs back on to the node. This only happens if the job owner set the HPC_ATTACHTOSESSION job environment variable to TRUE or TRY and the job successfully attached to the job owner’s log on session.

Cause: A job owner can log on to a node and attach their job to their console session in order to view the Excel processes that are created for the workbook offloading job. If the computer gets restarted, the Excel processes might be automatically restarted when the user logs back into the computer. Windows Update service often causes a computer to restart.

Solution: To avoid this behavior, use HPC_ATTACHTOSESSION only if necessary. For workbook offloading jobs, the Mayn usage is for troubleshooting.  Additionally, do not allow Windows Update to conflict with cluster jobs. This can be done by performing manual updates or scheduled updates which correspond with scheduled cluster or node downtime.

See Also

Job or task environment variables for console or remote desktop sessions Creating and Submitting Jobs [Help link] Job Submission in Windows HPC Server 2008 Quick Reference