Controlling Unix Platform Agents

Starting and Stopping Agents on UNIX using SYSV init

You use the scheduler script to start and stop network processors. You start all network processors or start a specific instance and can optionally provide an instance directory, which is the installation directory of the instance.

${InstallDir}/latest/etc/scheduler {start|start-instance|stop|stop-instance|list|list-instances|restart|status}

start                                                         start all network processors
start-instance <instance_name> [<instance_directory>]         start a specific instance
stop                                                          stop all network processors
stop-instance <instance_name> [<instance_directory>]          stop a specific instance
list | list-instances                                         list all instances
restart                                                       restart all network processors
status                                                        list all instances with the current status

Starting and Stopping Agents on Linux using SystemD

You use the systemctl command to perform actions on the platform agent(s):

systemctl {start|status|stop} scheduler[@<instance>]

systemctl start scheduler          start all instances
systemctl start "scheduler@<instance>"                        start instance <instance>
systemctl stop scheduler          stop all instances
systemctl stop "scheduler@<instance>"         stop instance <instance>
systemctl status scheduler                                    print status of scheduler unit
systemctl status "scheduler@<instance>"                       print status of <instance> unit

Configuring Start at System Startup

When you enable a service for an instance, the platform agent instance will be started at system start up, if you disable a service, it will not start at system startup.

Per instance:

systemctl enable "scheduler@<instance>"
systemctl disable "scheduler@<instance>"

Starting (`enable`) and stopping an instance using

All instances:

systemctl enable scheduler
systemctl disable scheduler

When you make changes to the systemd configuration, you need to reload the configuration:

systemctl daemon-reload
systemctl reset-failed

Starting and Stopping Agents on Solaris

You use the following tools to control services in SMF:

• svcs - inspect service status
• svcadm - enable, disable, clear maintenance flag on services
• svccfg - configure SMF

Starting and Stopping Services

if you do not specify a specific instance to start (enable) or stop (disable), then all platform instances will be affected.

svcadm enable redwood/scheduler[:<instance_>]
svcadm disable redwood/scheduler[:<instance_>]

Troubleshooting Services

When a service instance fails, its maintenance flag is set and the service cannot be enabled again unless the flag is cleared. In general, you inspect the SMF log using svcs -x redwood/scheduler:<instance> and the platform agent logs under <install_dir>/var/<server_sid>/<instance>/trc/

Once you have resolved the issue, you check if any processes associated with the instance are running, terminate any existing processes, and clear the maintenance flag for the service:

#Check if any processes are still running for the service:
svcs -p redwood/scheduler:<instance>
#Terminate any processes, note that these might be associated with RunMyJobs processes
pkill -9 <process_name>
#Clear the maintenance flag
svcadm clear redwood/scheduler:<instance>