I’m setting up my company’s instance of Icinga to use two master nodes in an HA setup. It seems that most examples of Icinga HA setups have the DB IDO on a separate server or even cluster of servers and I’m wondering what the advantages of this are. In particular I have two questions:
If the database becomes unavailable, is it the case that Icinga checks and notifications will continue running but just the web front-end and history logging will stop?
If so, are there any advantages to having a separate DB server or clustered DB setup besides improving web front-end reliability? Is it purely for performance and availability reasons?
Check this thead and see if it answers some of your questions:
There are also various other threads concerning the database in a HA setup (just search on the forums).
AS for this I’m not sure.
I think I read something about some kind of queuing if the database is not available. But I couldn’t find it just now. And even if there is, I think it will only be for a limited time.
Maybe other after more realiable information on this