next up previous
Next: Shutting down PRM Up: Starting up and Previous: Startup using the

Manual startup

  1. On each host that will be part of the PRM environment start up a node manager in the background:

            nodemngr -s sysmngr_host_name  other-options &

    Write a node-configuration file for the system manager's use. This file should contain the list of nodes that the system manager will manage. The format of this file is:

            hostname-1   architecture-type
            hostname-2   architecture-type
              ..               ..
              ..               ..

    A sample node-config file can be found in $PRM_TOP/demo/ncfile.example. If sysmngr is started through the PRM_startup script, a node-config file is automatically created by the script.

  2. Start up the system manager on host sysmngr_host_name.

    	sysmngr ncfile &

Note: If a node or system manager is aready running, attempts to start another node or system manager on the same host will result in an error message of the form "can not bind socket". If so, use ps(1) to check whether a node manager, system manager, PRM server (such as term_iotask or file_iotask) or PRM user tasks are already running and kill them. Rerun the startup procedure.

Santosh Rao
Tue Feb 13 13:59:23 PST 1996