Installing and Configuring Windows Compute Cluster Server 2003 for Batch Monitoring

CATIA must already be installed on the computer running the batch launcher.

The operating system can be:

• Microsoft Vista
• Microsoft Windows XP Professional Edition Service Pack 2 (SP2)
• Windows XP Professional x64 Edition
• Windows Server 2003 32-bit
• Windows Server 2003 x64 Edition.

1. Install the second CD which contains the Microsoft Compute Cluster Pack.

   When the Microsoft Compute Cluster Pack Installation Wizard is displayed, click Next several times until you reach the Select Installation Type dialog box:

   
   
   

   In our example, note that the only choice available is Install only the Microsoft Compute Cluster Pack Utilities for cluster users and administrators. The other choices are only available when installing the cluster pack on head or compute nodes running a Windows Server 2003 system.

   Click Next several times and complete the installation.

2. Install the Microsoft Compute Cluster Pack SDK.

   The Microsoft Compute Cluster Pack SDK can be obtained from the following site:

   http://www.microsoft.com/downloads/details.aspx?FamilyID=d8462378-2f68-409d-9cb3-02312bc23bfd&DisplayLang=en

CATIA must already be installed on the head and compute nodes.

Windows Compute Cluster Server 2003 is based on Windows Server 2003 Standard x64 Edition and can only be installed on 64-bit hardware.

There are two cases.

• on a node without an operating system, you must install both the first CD containing Windows Server 2003 Compute Cluster Edition, and then the second CD containing the Microsoft Compute Cluster Pack
• on a node running one of the following operating systems:
  
  • Windows Server 2003 Compute Cluster Edition
  • Windows Server 2003 Standard x64 Edition
  • Windows Server 2003 Enterprise x64 Edition
  • x64 Editions of Windows Server 2003 R2

  you only need to install the Microsoft Compute Cluster Pack.

1. Install the Microsoft Compute Cluster Pack on the head node.

2. Install the Microsoft Compute Cluster Pack on each compute node.

When Compute Cluster Pack installation is complete, it will display a To Do List page that shows you the steps necessary to complete configuration of your compute cluster. These steps include defining the network topology, configuring RIS using the Configure RIS wizard, adding compute nodes to the cluster, and configuring cluster users and administrators.

A certain number of configurations tasks must be performed on the batch launcher and/or nodes before submitting batches.

Configuration data are the input and output data for the batches (XML files). All the configuration data used by the grid must be accessible in the same way from any node. For example, if the data are stored in a database, all the nodes must have access to this database. If the configuration data are files-based, all the paths referenced in the batch monitor GUI must be UNC paths(\\machine\shared) and shared by all the nodes. Therefore, for file-based data, you must create a shared directory location to store these data.

On the batch launcher and the head node:

1. Check that your local disk is accessible for everybody through a share.

   For example,

   E:\

   must be accessible as:

   \\machine\HOME

2. Create a directory called GRID_HOME under each share.

   At this point, you should have:

   \\batchlauncher\HOME\GRID_HOME

   and:

   \\head\HOME\GRID_HOME

   visible and accessible from any node of the grid.

Certain PLM infrastructure and Windows Compute Cluster Server 2003 files need to be accessed by any node, therefore it is logical to store them on the head node.

These files are:

• the

  catstart.exe

  executable program (copied from a CATIA installation) and the license configuration file

  LicenseSetup.txt

• Windows Compute Cluster Server 2003 configuration files in XML format.

Perform the following operations on the head node:

1. Create the following directories:

   \\head\HOME\GRID_HOME\Env

   and:

   \\head\HOME\licences

2. Create the following directory:

   \\head\HOME\GRID_HOME\Config

   which will be used to store the Windows Compute Cluster Server 2003 XML configuration files.

3. Copy a version of the catstart.exe taken from the following directory of any CATIA 64-bit installation:

   C:\Program Files\Dassault Systemes\B211\win_b64\code\bin

   to the directory:

   \\head\HOME\GRID_HOME\Env

4. Create a text file named LicenseSetup.txt in the directory:

   \\head\HOME\licences

   The file must contain one line per batch.

   Each line contains:

   
   
   • the name of the batch as it appears in the Batch Monitor
   • the pipe symbol:|
   • the name of the license to be reserved for this batch (three-letter prefix for license).

   For example:

   ...
   CATDMUUtility|AL2.slt
   MigrateV5ToV4|AL1.slt
   ...

   If you need several products/configurations for a given batch, use the following syntax :

   CATDMUUtility|AL2.slt|AL1.slt

   In grid mode, this license configuration file will be used instead of the license information usually specified using the License setup button in the batch monitor when not using the grid mode.

This tasks explains how to set variables for the WCC server.

1. On the batch launcher, start CATIA and select the Tools > Options... command.

2. Select the PLM Batches tab and configure the parameters.

   For full details, refer to the description of the PLM Batches tab in the Infrastructure Users Guide.

This tasks explains how to submit a batch in grid computing mode.

Once you have configured the PLM Batches tab and started the batch monitor, batches are submitted in the same way as in normal mode.

However, there is one slight variation in the batch monitor user interface:

1. Start the batch monitor, select a batch and access the GUI of the batch.

2. Specify the batch parameters.

   You have two choices:
   

   • click the Run button to submit the batch to the grid, using the default Windows Compute Cluster Server 2003 configuration parameters
   • in grid computing mode, there is a Grid button at the bottom of the batch parameters dialog box; clicking this button opens a file explorer which allows you to select an XML configuration file which contains Windows Compute Cluster Server 2003 parameters.

   This is an example of what a file may contain:

   <?xml version="1.0" encoding="utf-8"?>
   <Job xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
   xmlns:xsd="http://www.w3.org/2001/XMLSchema" SoftwareLicense="" 
   MaximumNumberOfProcessors="1" MinimumNumberOfProcessors="1" 
   Runtime="Infinite" IsExclusive="false" Priority="Normal" 
   Name="MigrateV5ToV4" Project="" RunUntilCanceled="false">
   </Job>

   When you submit the batch, the configuration file will be saved on the head node in the directory:

   \\head\HOME\GRID_HOME\Config

   with the name of the current batch, for example:

   MigrateV4ToV5.xml

   If the same batch is submitted later, the same file will be used. Therefore, be aware that using the Grid button sets the Grid parameters for a specific batch, not only for the current batch execution.

   If the Grid button is used for an already configured batch, the new XML file will overwrite the old one.

   Note: when submitting a batch in grid computing mode, all input and output paths references in the dialog box for submitting batches must be in UNC format. They must also be accessible by all nodes with, as a minimum, read authorization for inputs and write authorization for output files.

   For an explanation of compute cluster submission parameters, please refer to the Microsoft Compute Cluster User's Guide.

This tasks explains how to submit a batch in command line mode.

1. To start a batch on the grid through the command line interface, run the command:

   CATBatchStarter -input xmlparameterfile

   where xmlparameterfile is any parameter file generated using the batch monitor.

2. Check the batch has been executed successfully.