If an Opsview host shuts down before all its Opsview components stop, old
.lock files might prevent the components starting properly. Follow these steps to cleanly restart all Opsview components on an Opsview host, including tidying up left-over
Monitoring outage likelyRestarting the Opsview components on an Opsview host will likely cause a monitoring outage. The severity and extent of such an outage will vary depending on the host and its role in your Opsview environment.
- Log in to the Opsview host as
- Stop the Opsview components using
/opt/opsview/watchdog/bin/opsview-monit stop all
Wait for all the components to stop. All components should be Not monitored when you check the
/opt/opsview/watchdog/bin/opsview-monit summary -B
Depending on the speed of your system, stopping all the components may take a few minutes.
- Stop the
systemctl stop opsview-agent opsview-watchdog
- Kill any remaining
opsviewuser processes using
pkill -u opsview
- Check that there are no remaining processes running under the
ps -fu opsview
If there are any such processes, kill them with
- a common process not to stop, which needs to be killed is the
kill -9 <pid>
- Delete any Opsview
find /opt/opsview -name *.pid -delete find /opt/opsview -name *.lock -delete
- Start the
systemctl start opsview-agent opsview-watchdog
- Start the Opsview components using
opsview-monit- the initial opsview-monit command is necessary to initialise opsview-monit first
/opt/opsview/watchdog/bin/opsview-monit start all
Depending on the speed of your system, starting all the components may take a few minutes.
watch /opt/opsview/watchdog/bin/opsview-monit summary -B
- exit the commandwith CTRL+C
Please sign in to leave a comment.