Configuring a remote shell script for the parallel engine (Windows)

To run parallel jobs on remote Microsoft Windows hosts, define a remote shell script on the conductor node.

About this task

A sample script named remsh.example is provided in the PXEngine\etc directory. Copy this script file and modify it for your configuration.

Also, modify the remcp.example file. The remcp script controls which remote copy utility (rcp or scp) is used by the IBM® InfoSphere® DataStage® distribute-component script. The distribute-component script is also located in the PXEngine\etc directory. It is used when the transformer libraries in jobs that contain compiled PX transforms must be copied to the compute nodes.

Procedure

  1. Copy the remsh.example script file, and rename it remsh. Save the file in the PXEngine\etc directory.
  2. Edit the script file and locate the line that begins with rsh:
    #!/bin/sh 
    # Example $APT_ORCHHOME/etc/remsh 
    rsh "$@"
  3. Change the line as shown in the following example:
    #!/bin/sh 
    # Example $APT_ORCHHOME/etc/remsh 
    $(ROOTDIR)/mksnt/rsh.exe "$@"
  4. Copy the remcp.example script file, and rename it remcp. Store the file in the PXEngine\etc directory.
  5. Edit the script file and locate the line that begins with rcp:
    #!/bin/sh 
    # Example $APT_ORCHHOME/etc/rercp 
    rcp "$@"
  6. Change the line as shown:
    #!/bin/sh 
    # Example $APT_ORCHHOME/etc/rercp 
    $(ROOTDIR)/mksnt/rcp.exe "$@"
  7. Save your changes.
    You are now set up to run parallel jobs that contained transform operators in a multi-node clustered SMP environment.

What to do next

Set up a method for distributing compiled transform libraries to all physical compute nodes in the cluster. See Distributing transform libraries to compute nodes.