Start and Stop Accelerator

To start the Accelerator vovserver, login as user rtdamgr (or the user chosen as Accelerator administrator) on the machine that is designated as the vovserver, and the use the following command:

% ncmgr start

To stop the Accelerator vovserver, use the following command:

% ncmgr stop

Note: To stop the vovserver while there is an active workload (running jobs), use the option -freeze. When -freeze is used, all currently running jobs are preserved. If -freeze is not used, all vovtaskers are stopped along with vovserver, and any running job is also terminated. The -freeze option feature can be important when upgrading Accelerator.

% ncmgr stop -freeze
% ncmgr start -force

Reset Accelerator and Restart Taskers

Changes made to Accelerator's configuration files can be read in by performing a reset of the queue. The default behavior is to re-read all configuration files. There are also two optional reset types: soft and hard.

Soft Reset (-soft): A soft reset instructs all running taskers to re-read minor configuration changes and starts configured taskers that are not running at the time. Running jobs are not disturbed.
Hard Reset (-hard): A full reset stops and restarts all taskers. This is a forceful command and will kill all running jobs.


vncmgr: Usage Message
  
          Utility to reread configuration files and start/restart slaves.
  
          USAGE:
          % ncmgr reset [OPTIONS]
  
          OPTIONS:
          -help               -- Print this message
          -queue <name>       -- Specify the queue name
  
          -soft               -- Starts any stopped, but configured slaves
          -slaves             -- Same as -soft
  
          -hard           *   -- Stops and starts all slaves (jobs are lost)
          -full           *   -- Same as -hard
  
          * Warning: These options will forcefully stop all running jobs
  
          EXAMPLES:
  
          % ncmgr reset
          % ncmgr reset -slaves
          % ncmgr reset -full
  
          % ncmgr reset -soft
          % ncmgr reset -hard

% ncmgr reset                  - Reread configuration files
% ncmgr reset -queue vnc2      - Reread configuration files for vnc2 queue
% ncmgr reset -soft            - Reread configuration files and start taskers that are not running
% ncmgr reset -hard            - Reread configuration files and restart all taskers

For an example of using the full reset option, refer to Directories and Files.

Start Accelerator at System Boot Time

The instructions in this section are valid for Linux only. This part of the installation requires root permission.

Note: This step is optional.

The Accelerator vovserver can be restarted at reboot by installing the proper script in both the /etc/rc3.d and /etc/rc5.d directories.

Run the following commands on the host that was selected as the Accelerator vovserver.

% /bin/su
% cp $VOVDIR/etc/boot/S99nc /etc/rc2.d/S99nc
% chmod 755 /etc/rc2.d/S99nc
% vi /etc/rc2.d/S99nc
....
Edit configurable items as needed.

Note: sudo should be used where configured. To avoid Trojan Horse programs, su should always be called by full path /bin/su.

% ./S99nc start

Messages about the vovserver starting should be displayed. An example is shown below:

% ./S99nc status

The output of nc info should be displayed. An example is shown below:

% ./S99nc stop

Note: Re-start the Accelerator server with the command S99nc start after testing the stop capability.