We've been using the FOE product since it replaced the Hot Standby product several years ago and have been very unsatisfied with the way it works. One of the complications with it is that if you have an NPM implementation with multiple additional polling servers there is no way we have found to make every server with an FOE pair up fail over at the same time. This means that all the servers and their paired FOE server have to be in the same location. In other words if we wanted to have a redundant arrangement of all the main poller and all the addtiional pollers at a physically diverse location we would have to manually fail all the additional servers if the main server failed. The reason I say this is because Solarwinds doesn't support the main poller and the addtional pollers seperated across a WAN. In the scenario above if the main poller failed to a secondary server at a location different from where the addtional pollers are still up, the addtional latency could cause communications and performance problems.
In addition , we're having problems with the FOE installation process wiping out the license key on the secondary server. This started with Windows Server 2008 due to a change by Microsoft in the licensing registration process.
So having said all this what is my feature request? I'm suggesting that Solarwinds come up with a different fail over process that takes into account a customer having a completly duplicate set of servers at a location connected by a MAN or WAN link, an arrangement that doesn't use the Neverfail FOE product. I know there is a white paper on doing it this way, but it requires the customer to purchase a complete duplicate set of licenses for the backup set of servers which doesn't take into account the fact that those licenses are not used until a failure occurs. Also, fail over in the white paper is a manual process. The issue is that FOE does not scale to large scale operations that require physicaly seperate backup facilities.