So here's the thing. We have several HA pairs of pollers including the primary poller. Our server team utilise BMC Patrol Agent to monitor the health of the server but we also utilise an edited version of it for generating what we call heartbeats. Essentially an alert is created by the man active poller every 5m which is in a set format. The BMC Patrol SOLW' agent picks this up and drops it into BMC. If it misses 2 in a row an incident is raised for the server team to initially look at..
All well and good until you have to restart the server and forget (because you were never told this) to re-start this service.
So, I'm into thinking about one of two options ...
1. Create an alert that identifies which is the active HA partner and check, and if necessary, start this service.
2. Alternatively, create a script/batch file/start up process that identifies which is the HA master and kickstarts the service.
The initial issue I have with either of these ideas is how to automate the identifying of the HA master. The secondary issue is the how to do this in an alert (which seems to me to be the most obvious/easiest way).
Thoughts?