2 Replies Latest reply on Nov 28, 2014 3:35 AM by RichardLetts

    Failover Environment - Unique or different IPs

    erik.marcus@openline.nl

      Hello Solarwinds Support,

       

      I’ve got a question:

       

      I’m setting up a new Solarwinds environment. I’ve bought the necessary licenses already

      I want to use the failover engine in an 2 datacenter environment connected by a WAN link and a Dark fiber interconnection. Which looks like in the graph.

       

      Above IP’s are on the principal networkadapter. Which is an identical IP on each side(i.e. 10.239.80.1) + a diferent management IP on each side (10.239.80.11 & 10.239.80.21)

      The private (Solarwinds Channel adresses) will be something like 9.0.0.1 but it did not show that in this graph.

       

      Wtat the FailoverEngineAdminGuide does not make clear is the following:

      1. 1.       It tells me that I can use identical principal networkadapter IP addresses.
      2. 2.       But then it tells me I need different IP addresses because I use a WAN connected DR Site.
      3. 3.       Last it tells me I need a management IP on both sides.

       

      Now I’m confused. Is my graph correct? Or should the primary and secondary server (i.e. VCSW101) have:

      Different Public IPs (10.239.80.1) and different management IPs (80.11. and 80.21)?

      Or

      Unique public IP and only different management IPs ?

       

       

      Thanks in advance,

        • Re: Failover Environment - Unique or different IPs
          aLTeReGo

          It looks like you're trying to use the LAN/HA configuration with a WAN/DR setup. That is technically possible provided you can extend your layer 3 boundary out to your DR location, but most people are not that fortunate. If that's the design you're going with then ignore instructions related to WAN/DR configuration in the Administrators Guide and read only the LAN/HA configuration steps. In a LAN/HA configuration it's expected that the IP address of the primary network interface will be the same for both primary and secondary servers. In a WAN/DR mode, each site has it's own unique layer 3 address space and both primary and secondary servers are in different subnets. As such each primary network interface will have it's own unique IP address. 

          • Re: Failover Environment - Unique or different IPs
            RichardLetts

            The FOE product makes things seem much harder than they need to be. (at its fundamental level all it is doing is replicating files and registry entries between the servers, controlling the state of a network interface, and starting/stopping services as needed. )

             

            there are two configurations that aLTeReGo mentions: LAN or WAN

            In any evironment the servers need a private channel that is always up for them to communicate over; this channel is used to replicate data. These internal/solarwinds channel interfaces need to always be able to talk to each other.

             

            The public/external interfaces follow the active/passive FOE state: the FOE product installs a driver that blocks external communication.

            in a LAN configuration the external/polling interfaces have the same IP address on both the primary and secondary servers; whichever server is the active can access the network using the IP address. You need a common L2 infrastructure to make this work. Routing for the replication channel is simple (they are on the same subnet). you don't get much redundancy in this environment. Honestly, vmware would give you better redundancy/simplicity/manageability for the polliong and application servers.


            In a 'WAN' configuration the external/polling interfaces have different IP addresses (and the FOE implements dynamic DNS updates to point at the currently active server.). Takes a lot more effort to setup, but you can have geographic redundancy. You have to spend more time on the infrastructure setup, and understand windows, DNS, and routing much deeper to manage this.