OpenVPN Support Forum

Hi,

We've been using the OpenVPN for a while. We have three remote offices and several hosted servers. They are all connecting to one OpenVPN server. The IP network is 172.22.100.0/24. The OpenVPN server assigns the static IP addresses to the hosted servers and we've given a host name to each hosted server so that the team can access these servers from their computers in the remote offices. For Window's client, we added ip address/hostname pair in the hosts file.

Now, we want to add a secondary OpenVPN server for failover. We've configured the secondary network as 172.22.200.0/24. We've tested the failover. It works beautifully.

But here raised an issue on accessibility. In a failover scenario, the same hosted server (says, serverA) could have two ip address: 172.22.100.53 (from primary VPN server) or 172.22.200.53 (from secondary VPN server). Now if the client switch from primary server to secondary server, any program uses the host name (e.g. bookmarks in the web browser) won't work any more.

My Question:
Is there a way that server can push a list static ip address/host name pair over to the client when the switch occurs.

Thanks for help!

MF

just to make sure I got it right: all hosted servers have TWO IP addresses, depending on which VPN server they are connected to? or are the hosted servers connected via a LAN interface to both servers?

in the latter case it would be just a matter of setting up routing correctly.
If the hosted servers are also connected to the VPN then I'd set up a DNS server on both VPN servers and push it out to all clients - that way serverA.example.com would always resolve to the right IP.