We have spatially redundant servers which manage our phone system.
The servers sit in different IP subnets. Both servers have physical IP addresses which are always up if the physical server is powered on, e.g. 192.168.1.2 and 192.168.2.2.
In normal operation only one server will be the “main” server. The server which is “main” makes a virtual IP available (e.g. 192.168.1.1).
A keep alive runs between the 2 servers. If the standby loses sight of the “main” then it kicks into “main” and publishes it’s VIP (e.g. 192.168.2.1)
Only one of the VIPs should exist at once.
In NPM I have a group with the two VIPs as members and have set the group with a roll up status of ‘best status’, which shows we have an active “main” server.
What I need to be able to do is hide the inactive VIP so that it does not show such as a node with packet loss, while it is not active but remains available to monitor as soon as the server it is on becomes main again. Suspending monitoring does not do this for me. This does not serve any funtion as to the availability of the servers themselves as this can be done through the physical IP address.
The logic I need is to ping a single node which could have address 192.168.1.1 OR 192.168.2.1 and avoid showing the address that does not exist as "down".
Any ideas?
Thanks in advance