New Feature Evaluation Guide for Windows HPC Server 2008 R2

Goal

Use the Add Software Updates Wizard to add an Apply Updates task to a node template.

Steps

The Mayntenance phase of a node template can include an Apply Updates task, with settings that you configure for which updates to apply. When you run the Mayntain action on nodes, the Apply Updates task downloads updates to the compute nodes from the Microsoft Update website or the WSUS server in your enterprise, and then installs the updates.

Note
The Apply Updates task in the node template cannot install updates that include software license terms (also known as an End User License Agreement or EULA). This type of update requires the administrator of each node to accept the software license terms before the update can be installed. In this scenario, you have to install updates that include software license terms manually.

The following procedure describes how to add the Apply Updates task to a node template. The node template must have already been used to deploy one or more nodes.

1. In HPC Cluster Manager, click Configuration, and then click Node Templates.
   
   
2. Right-click a node template, and then click Add Software Updates.
   
   
3. Follow the steps in the Add Software Updates Wizards to select the update level and the specific updates to add. In the final step, you can select to install the updates now or later.
   
   
4. To make changes to the Apply Updates task, you can run the wizard again, or right-click the node template and click Edit.
   
   

Related Resources

For more information about applying updates by using and enterprise WSUS server or by using a node template, see the Best Practices topic in the updating nodes step-by-step guide (https://go.microsoft.com/fwlink/?LinkId=194794).

Goal

Add workstation nodes to your HPC cluster and define a weekly availability policy to control when these nodes are brought online.

Steps

1. In HPC Cluster Manager, create a node template for the workstations:
   
   
   1. Use the Create Node Template Wizard to create a Workstation node template.
      
      
   2. On the Configure Availability Policy page, select how and when you want workstation nodes to be brought online and offline.
      
      
2. Install HPC Pack 2008 R2 on the workstations and select the Join an existing HPC cluster by creating a new workstation node option.
   
   
3. The nodes appear in Node Management as Unapproved.
   
   
4. Assign the workstation template that you created to the nodes.
   
   

Related Resources

Adding Workstation Nodes in Windows HPC Server 2008 R2 Step-by-Step Guide

Goal

Create one or more new tabs in Node Management.

Steps

1. In HPC Cluster Manager, click Node Management.
   
   
2. In the Navigation Pane, click Nodes.
   
   
3. In the view pane, click the blank tab, and then click Customize Tab.
   
   
4. Type a name for the tab, and then click List View or Heat Map View.
   
   
5. Add one or more metrics.
   
   
6. Click Apply at any time to see the current tab configuration.
   
   
7. Click OK to save your changes to the tab.
   
   
8. In the view pane, use the slide bar to adjust the heat map zoom, or click the Fit to window icon to automatically adjust the tile size for the best fit.
   
   
9. You can organize the heat map view by location by clicking the Group by location icon. Remove the location grouping by clicking the Group by name icon. (You can specify the node location property in the node XML or by selecting a node and clicking Edit.)
   
   

To change the settings on a tab, right-click the tab name, then click Customize Tab.

If you are creating a Heat Map tab, you can customize the following display options:

• Color scale: The minimum value for a metric is associated with a color, for example, white, and the maximum value for that metric is associated with another color, for example, blue. In this case, lower values for that metric appear as lighter shades of blue, and higher values appear as darker shades of blue. For each metric, you can customize the maximum and minimum values and associated colors. You can also flip the scale so that the minimum values are darker, and the maximum values are lighter.
  
  
• Linear or logarithmic color banding: The color bands that are used to display metric values can be displayed in linear or logarithmic scales. In a linear scale, the color bands are equally sized across the defined value range. In a logarithmic scale, the color bands are logarithmically sized across the value range. Logarithmic scale is useful when you want to visually distinguish values at one end of the value range.
  
  
• Stacking or overlaying metrics: You can view multiple metrics in Stacking or Overlaying view. Stacking displays a color bar for each metric. Overlaying displays only the most significant metric for each node. Significance is based on the order in which the metrics are defined in the Customize Tab dialog box. The first metric is displayed by default. If a metric value reaches the darkest color band, that is the metric that is displayed. If more than one metric reaches the darkest color band, the first one listed is the one that is displayed.
  
  
• Metric aggregation period: Aggregate metrics over a short time period by increasing the number of seconds for the metric value display.
  
  

Related Resources

HPC R2 Demo: New heat map and location-based node management features – video (7 min.)

Goal

Save the fsutil volume diskfree command as a diagnostic test that checks for free disk space on your nodes and then run the test and view results from HPC Cluster Manager.

Steps

Step 1: Define the test

1. On the head node, open a text editor such as Notepad, and paste in the following XML code. Optionally, change the Company name to your name.
   
   

   <DiagnosticTests>
     <DiagnosticTest
       Name="Free Disk Space"
       Description="Checks the amount of disk space free on the specified drive."
       Company="Beta Participant"
       Suite="Sample"
       Alias="diskspace">
       <Parameters>
         <Parameter Name="Drive" SwitchName="Drive" Type="String" Visibility="True" 
           Description="The letter name of the 
           drive for which you want to check the 
           amount of free disk space, without a 
           colon (:) or backslash (\). The default 
           is C." 
           DefaultValue="C" Format="{1}:"/>
       </Parameters>
       <RunStep Command="fsutil volume diskfree" />
     </DiagnosticTest>
   </DiagnosticTests>

2. Save the file as C:\SampleTests\DiskSpaceTest.xml.
   
   Ensure that in Save as type, you select All files (*.*).
   
   

Step 2: Add the test to the cluster

1. Run HPC PowerShell as an administrator. and type the following cmdlet:
   
   Add-HpcTest -File C:\SampleTests\DiskSpaceTest.xml
   
   
2. To see information about the test you added, type the following cmdlet:
   
   Get-HpcTestDetail -Alias diskspace
   
   
3. To see all cmdlets that relate to HPC tests, type Get-help *hpctest*.
   
   

Step 3: Run the test and view results

1. In HPC Cluster Manager, click Diagnostics.
   
   
2. In the Navigation Pane, under Tests, select the new node named Beta Participant.
   
   
3. In the view pane, right-click Free Disk Space, then click Run.
   
   
4. In the Run Diagnostics Tests dialog box, click Run.
   
   
5. In the Navigation Pane, click Test Results, select the new node named Beta Participant.
   
   
6. In the Details Pane you can see the progress of the test. Click the Results tab to see the results.
   
   

Related Resources

• Diagnostics Extensibility in Windows HPC Server 2008 R2 CTP2 Step-by-Step Guide on MSDN (https://go.microsoft.com/fwlink/?LinkId=177604)
  
  
• HPC R2 Demo: New diagnostics extensibility features – video (11 min.)
  
  

Goal

Change the scheduling mode from Queued to Balanced.

Steps

1. Use one of the following methods to change the Scheduling Mode from Queued to Balanced:
   
   
   • In HPC Cluster Manager, click Options, and then click Job Scheduler Configuration. Scheduling Mode and the associated settings can be configured on the Policy Configuration tab.
     
     
   • Run HPC PowerShell as an administrator, and then type:
     
     Set-HpcClusterProperty –schedulingMode Balanced
     
     
   • Open a Command Prompt window as an administrator, and then type:
     
     cluscfg setparams SchedulingMode=Balanced
     
     
2. Submit several jobs to the cluster.
   
   

After you have set the Balanced mode, you can adjust how additional resources are allocated with the PriorityBias setting, and how often the scheduler rebalances with the ReBalancingInterval setting.

PriorityBias controls how additional resources are allocated to jobs. In terms of Balanced mode, “additional resources” refers to cluster resource above the total minimum resources for all running jobs. Tasks that are running on additional resources can be canceled with immediate preemption to accommodate new jobs or to converge on the desired allocation pattern. You can choose from the following three options:

• HighBias: All additional resources are allocated to higher priority jobs.
  
  
• MediumBias (Default): Each priority band is given a higher proportion of additional resources than the band below it. The priority bands are Highest, Above Normal, Normal, Below Normal, and Lowest.
  
  
• NoBias: Resources are allocated equally regardless of priority.
  
  

ReBalancingInterval represents the time, in seconds, between scheduler rebalancing passes.

You can use one of the following methods to change Priority Bias and ReBalancingInterval:

• Run HPC PowerShell as an administrator, and then type:
  
  Set-HpcClusterProperty –PriorityBias HighBias –ReBalancingInterval 20
  
  
• Open a Command Prompt window as an administrator, and then type:
  
  cluscfg setparams PriorityBias=HighBias ReBalancingInteral=20
  
  

Related Resources

Queued mode is priority-based, first come first served scheduling like in Windows HPC Server 2008. For information, see Understanding Job Scheduling Policies (https://go.microsoft.com/fwlink/?LinkId=177866).

Goal

Add a service on the cluster and manage the service configuration settings from HPC Cluster Manager.

Steps

1. Copy your service .dll file to a folder named C:\ServicesR2 on each compute node.
   
   
2. On the head node, copy the service configuration file to the C:\Program Files\Microsoft HPC Pack 2008 R2\ServiceRegistration folder.
   
   
3. Click Start, point to All Programs, click Microsoft HPC Pack 2008 R2, and then click HPC Cluster Manager.
   
   
4. In HPC Cluster Manager, click Configuration, and then click Services.
   
   
5. The view pane displays a list all services that have configuration files in the ServiceRegistration folder. Verify that the service that you just added appears in the list.
   
   

6. Right-click your service, then click Edit Configuration File. The configuration file for your service opens in the default XML editor.
   
   Ensure that the assembly attribute of the service element points to the location of your service .dll (C:\ServicesR2\<yourServiceName>.dll). For example:
   
   

   <microsoft.Hpc.Session.ServiceRegistration>
     <service assembly="C:\ServicesR2\<yourServiceName>.dll">
   </microsoft.Hpc.Session.ServiceRegistration>

   Save the changes, if you made any, and then close the text editor.
   
   

7. To verify that the service can be loaded, right-click the service, and then click Run SOA Service Loading Diagnostic Test.
   
   
8. The Run Diagnostic Tests dialog box appears, and the service that you selected is automatically specified in the parameter for the test. Click Run.
   
   
9. To view test results: In Diagnostics, in the Navigation Pane, click Test Results.
   
   

Related Resources

• Enable and collect trace logs to troubleshoot SOA sessions
  
  
• Download the HPC Pack 2008 R2 Software Development Kit (https://go.microsoft.com/fwlink/?LinkId=198809).
  
  
• See the Session APIs under HPC Class Library (https://go.microsoft.com/fwlink/?LinkID=167944).
  
  
• SOA Applications and Windows HPC 2008 R2 (https://go.microsoft.com/fwlink/?LinkId=193167)
  
  

Goal

Enable tracing on the head node and collect the trace logs from each node that was used during the session.

Steps

When you enable tracing in the service configuration file, the trace information is logged to a file on the compute nodes. The log files trace steps from the service call and the intermediate results on cluster. You can collect and remove traces by using the Job Management view or the HPC PowerShell cmdlets. You can view the trace log files with the WCF Service Trace Viewer (SvcTraceViewer.exe).

1. Click Start, point to All Programs, click Microsoft HPC Pack 2008 R2, and then click HPC Cluster Manager.
   
   
2. In HPC Cluster Manager, click Configuration, and then click Services.
   
   
3. Right-click the service that you want to troubleshoot, and then click Set Event Logging Level. In the dialog box, select the desired trace level and then click OK.
   
   
4. Start a session with that service.
   
   
5. Click Job Management, and then click All Jobs.
   
   
6. In the job list, find the job that is associated with the session that you are debugging. The job ID is the same as the session ID.
   
   
7. Right-click the job, and then click Collect Trace.
   
   
8. In the Collect Trace dialog box, specify the shared folder where you would like to collect the trace logs. The folder must be accessible from the compute nodes.
   
   
9. Verify that the trace logs appear in the specified folder.
   
   
10. Right-click the job, and then click Delete Trace to delete the trace logs from the compute nodes.
    
    

Important
Event logging is not generally recommended for production environments. After collecting the trace logs, ensure that you delete them from the compute nodes to avoid consuming disk space.

Related Resources

• You can also use the C# Cluster-SOA Debugger for Windows HPC 2008 R2 to debug SOA applications that are running on a Windows HPC cluster.
  
  
• Download the HPC Pack 2008 R2 Software Development Kit.
  
  
• See the Session APIs under HPC Class Library (https://go.microsoft.com/fwlink/?LinkID=167944).
  
  
• SOA Applications and Windows HPC 2008 R2 (https://go.microsoft.com/fwlink/?LinkId=193167)
  
  

Goal

Users submit jobs with numerical priority levels.

When necessary, manually adjust priority levels on submitted jobs.

Steps

In HPC Pack 2008 R2, the job priority can have a value between 0-4000. Users can specify priority in terms of a priority band, a priority number, or a combination of the two. The priority bands and their corresponding numerical values are as follows:

• Lowest (0)
  
  
• BelowNormal (1000)
  
  
• Normal (2000)
  
  
• AboveNormal (3000)
  
  
• Highest (4000)
  
  

The numerical priority can have a value between 0 (Lowest) and 4000 (Highest). If you enter a value numerically, it is displayed as the corresponding priority band, or as a combination. For example, if you specify a value of 2500, the priority is displayed as Normal+500.

Monitor and adjust the job queue

Cluster administrators and job owners can modify the Priority job property for any active job (Queued or Running).

1. Run HPC PowerShell.
   
   
2. Type the following cmdlets to see a list of active jobs in job queue order (descending priority and ascending submit time):
   
   get-hpcjob|sort @{expression=”priority”; descending=$true}, @{expression=”submittime”; ascending=$true}
   
   
3. Modify priority levels on submitted jobs as appropriate. You can use the Set-HpcJob cmdlet to modify the priority of a job. For example, the following cmdlet sets a priority level of 2550 for job 122:
   
   set-hpcjob –id 122 -priority 2550
   
   

Related Resources

Check for license availability before a job is started

Goal

Build and try the Activation Filter sample that is included in the HPC 2008 R2 SDK samples (HPC2008R2.SampleCode.zip). The sample is a Visual Studio 2008 project named FlexLM.sln that is in the HPC2008R2.SampleCode \Scheduler\Activation Filter folder.

Steps

FlexLM.sln includes a sample activation filter that checks for license availability and the FlexLM.exe.config file that you can use to specify the location of the FlexLM utilites and license file. In the FlexLM projects properties, there are custom pre-build event commands and post-build event commands. The pre-build commands are used to create the files needed to create event log entries. The post-build commands unregister any old version of the FlexLM activation filter and then register the new version so that it can create events and the event viewer can display them. The commands assume that Visual Studio is creating the files in c:\Program Files\Microsoft HPC Pack 2008 R2\Bin\. The cluscfg command tells the HPC Job Scheduler to use the new Activation Filter.

The following steps describe how to configure and build the solution:

1. On the head node, run Visual Studio 2008.
   
   
2. Open FlexLM.sln.
   
   
3. Open the Jobs.cs file and update the reference to the job xml schema (from HPCS2008 to HPCS2008R2) in the ParseJobXml() method as follows:
   
   Change from:
   
   nsmgr.AddNamespace("ab", @"https://schemas.microsoft.com/HPCS2008/scheduler/");
   
   To:
   
   nsmgr.AddNamespace("ab", @"https://schemas.microsoft.com/HPCS2008R2/scheduler/");
   
   

   Important
   If you do not perform this step, the sample filter will not be able to parse the job XML file.

4. In the FlexLM.exe.config file, specify the paths to the FlexLM utility (in PollCommandName) and the FlexLM license file (in PollCommandArguments). The following XML code example shows how these paths are specified:
   
   

   <?xml version="1.0" encoding="utf-8" ?>
    <configuration>
      <appSettings>
        <add key="PollCommandName" value="C:\Program Files\Microsoft HPC Pack 2008 R2\Bin\lmutil.exe" />
        <add key="PollCommandArguments" value="lmstat -c C:\license-fluent.dat -a" />
        <add key="CacheFile" value="FlexLM.xml" />
        <add key="ApplicationStartupTimeInSeconds" value="30"/>
      </appSettings>
   </configuration>

5. Build the solution. The program deploys the DLL and .config files on the head node, creates event log files, and adds the activation filter to the Job Scheduler parameters.
   
   

To test the filter, submit jobs to the cluster that require licenses.

Related Resources

None.

Goal

Force cancel a job or task.

Steps

1. Submit a job with at least two tasks including:
   
   
   1. A task that runs an application that responds to the CTRL_BREAK event.
      
      
   2. A Node Release task.
      
      
2. To force cancel a task, use one of the following methods. Include the –force parameter and specify the ID of your job and task, and optionally, the sub-task.
   
   
   1. In HPC PowerShell, use the following cmdlet: Stop-HpcTask –JobId <yourJobID> -TaskID <yourTaskID> [-subTaskID <yourSubTaskID>] -force
      
      
   2. At a command prompt, use the following command: task cancel <yourJobID>.<yourTaskID>[.<yourSubTask>] /force
      
      
3. To force cancel a job, use one of the following methods. Include the –force parameter, and specify the ID of your job.
   
   
   1. In HPC PowerShell use the following cmdlet: Stop-HpcJob <yourJobID> -force
      
      
   2. At a command prompt use the following command: job cancel <yourJobID> /force
      
      

Related Resources

• Allow canceled tasks time to save state information or clean up before exiting
  
  
• Submitting a Job with Node Preparation and Node Release Tasks in Windows HPC Server 2008 R2 Step-by-Step Guide
  
  

Goal

Add one or more nodes to the Excluded Nodes job property.

See all excluded nodes on the cluster (Administrator).

Steps

Defining excluded nodes for a job

For any active job, you can add or remove nodes in the Excluded Nodes jobs property, or clear the list. The following lists the commands to modify and view the Excluded Nodes list using HPC PowerShell or a command prompt.

In HPC PowerShell, use the following cmdlets:

• Set-HpcJob –JobId <yourJobID> /addExludedNodes <nodeName>, <nodename>
  
  
• Set-HpcJob –JobId <yourJobID> /removeExcludedNodes <nodeName>, <nodename>
  
  
• Set-HpcJob –JobId <yourJobID> /clearExcludedNodes
  
  
• (Get-HpcJob –JobId <yourJobID>).ExcludedNodes
  
  
• Or to view all job properties, Get-HpcJob –JobId <yourJobID>|fl
  
  

At a command prompt, use the following commands:

• job modify <yourJobID> /addExludedNodes <nodeName>, <nodename>
  
  
• job modify <yourJobID> /removeExcludedNodes <nodeName>, <nodename>
  
  
• job modify <yourJobID> /clearExcludedNodes
  
  
• job view <yourJobID> /detailed|find “excludednodes” /i
  
  
• Or to view all job properties, job view <yourJobID> /detailed
  
  

Monitoring excluded nodes on the cluster

To see all excluded nodes on a cluster, use the Get-HpcJob PowerShell cmdlet. The following example shows how to list all of the excluded nodes for jobs that were submitted today. The script also lists the job template that was used for the job that excluded the node. In the following cmdlet, <today’s date> is specified in a date format such as mm/dd/yyyy:

Get-HpcJob –beginSubmitDate <today’s date>|select ExcludedNodes, Job Template|sort

If the cluster administrator detects and resolves the issue on one or more nodes, the administrator can remove the fixed node from any node exclusion list in which it appears. The following cmdlet gets all active jobs and removes the fixed nodes from the node exclusion lists (this has no effect on jobs that do not list the specified nodes):

Get-HpcJob|Set-HpcJob –removeExcludedNodes <fixedNodeName>,<fixedNodeName>

Related Resources

Job Submission in Windows HPC Server 2008 Quick Reference

Goal

Enable eMayl notification on the cluster and submit a job that requests notification on job completion.

Steps

Enable eMayl notification on the cluster:

1. In HPC Cluster Manager, click Options, and then click Job Scheduler Configuration.
   
   
2. On the E-Mayl notifications tab, specify the SMTP server, authentication, and originating address. Click the “More about eMayl notifications” link on the tab for important considerations.
   
   

Submit a job that requests notification on completion:

1. In HPC Cluster Manager, in Job Management, click New Job.
   
   
2. In Job run options, in Send a notification, select the Completes check box.
   
   
3. Add one or more tasks to the job, and then click Submit.
   
   

Related Resources

None.

Goal

Submit a job with Node Preparation and Node Release tasks.

Steps

For detailed step-by-step instructions, see Submitting a Job with Node Preparation and Node Release Tasks in Windows HPC Server 2008 R2 Step-by-Step Guide.

1. Create a new job.
   
   
2. Add a Node Preparation task.
   
   

   Note
   If you are using HPC PowerShell or a Command Prompt window, use the –Type property to designate a Node Preparation task, for example: Add-HpcTask –jobID <ID> –Type NodePrep Job add <ID> -Type:”NodePrep”

3. Add one or more primary tasks (Basic or Parametric Sweep) to the job.
   
   
4. Add a Node Release task.
   
   

   Note
   If you are using HPC PowerShell or a Command Prompt window, specify a task with Type set to NodeRelease.

5. Submit the job.
   
   

Now try to cancel a Running job that includes a Node Release task.

Related Resources

• Understanding Task Types
  
  
• Submitting a Job with Node Preparation and Node Release Tasks in Windows HPC Server 2008 R2 Step-by-Step Guide
  
  

Goal

Set and Mayntain values for job Progress and Progress Message from an application or script.

Steps

Include commands to set Progress and Progress Message in your scripts or applications. For example, if your application includes a loop that performs some work, you can update the progress properties at each iteration.

To set the Progress and Progress Message properties in a batch (.bat) file, an HPC PowerShell script (.ps1), or an application, you can use the %CCP_JOBID% environment variable to get the job ID of the current job, as follows:

• In a .bat file, use the job modify command, for example:
  
  Job modify %CCP_JOBID% /progress:50 /progressmsg:”Hello”
  
  
• In a .ps1 file, use the Set-HpcJob cmdlet, for example:
  
  set-HpcJob –id %CCP_JOBID% -progress 50 –progressmessage ”Hello”
  
  
• For an example of how to set job progress through the APIs, see Setting Custom Job Progress Information in Windows HPC Server 2008 R2 Step-by-Step Guide.
  
  

You can use one of the following methods to see the progress information for a running job:

• In HPC Job Manager, double-click a running job to open the View Job dialog box.
  
  
• At a command prompt, type the following command, where <jobID> is the ID of your job:
  
  Job view <jobID>
  
  
• In HPC PowerShell, type the following cmdlet, where <jobID> is the ID of your job:
  
  Get-HpcJob <jobID>|select id, Progress, ProgressMessage
  
  

Related Resources

• Understanding the Progress and Progress Message Job Properties
  
  
• Setting Custom Job Progress Information in Windows HPC Server 2008 R2 Step-by-Step Guide
  
  
• HPC R2 Demo: New job progress, job troubleshooting, and service-balanced policy – video (5 min.)
  
  

Goal

Allow tasks that are canceled time to perform cleanup or completion steps before exiting.

Steps

1. Submit a job that runs an application that includes code to handle a CTRL_BREAK event.
   
   
2. Cancel the job while it is running.
   
   
3. Verify that the actions in the application’s CTRL_BREAK event handler were performed.
   
   

You can use one of the following methods to change the Task Cancellation Grace Period to 10 seconds:

• Run HPC PowerShell as an administrator, and then type:
  
  Set-HpcClusterProperty –TaskCancelGracePeriod 10
  
  
• Open a Command Prompt window as an administrator, and then type:
  
  cluscfg setparams TaskCancelGracePeriod=10
  
  

Related Resources

None.