I have found an undocumented CON of splitting the Orion NPM junctions to seperate PCs or VMs. The problem occurs when the SQL Server PC or VM looses connection to the network or gets rebooted while the website and pollers are still running. It causes all of the configured, active alerts to get triggered. I have documented this problem in SQL server on a different Server than the Orion install.. I have created two .bat files that will stop and start the services and scheduled them to run on the Poller/Web Server in Scheduled Tasks. By stopping the Orion NPM services, it prevents the alerts for being triggered and stops the alerting engine from sending out e-mail (if you have an alert setup to e-mail) for each node, interface, and volume saying that it went down then another e-mail for each when SQL comes back up.
orstop.bat (241 bytes)