Two server fail safe setup possible?

[embeeweezer]

Hello all,

here's my setup: two softether servers in the cloud, several softether clients. I'd like to create an automatic fail safe system where clients automatically connect through 'the other' server in case one fails.
I've been checking about the cluster member servers and cluster controller. But since there's only one cluster controller:
*what if the cluster controller goes down* ?
Or is there some other solution to this?

Looking forward to your input!

Best,

Mike.

[thisjun]

How much time of downtime is allowed?
If allowed downtime is not so short, copy the configure from main server to backup-server periodically.
When the main server is broken, in backup-server, set global IP address same as the main server, and turn on VPN server on the backup-server.

[embeeweezer]

Hello thisjun, thanks for your reply. Well such a solution would be possible to script but because of cached dns / ip's on the clients, it will take (i don't know how) long before everything will reconnect to the new dynamic ip. This will be trouble.
So I wonder how could Softether function in a business critical environment? Who wants to run their business on software where the core networking has a *single* point of failure: the cluster controller...

[qupfer]

embeeweezer wrote:
> Hello thisjun, thanks for your reply. Well such a solution would be
> possible to script but because of cached dns / ip's on the clients, it will
> take (i don't know how) long before everything will reconnect to the new
> dynamic ip. This will be trouble.

A non elegant solution could be achieved with the hosts file.
I mean you could set the config to use "myvpnserver" as host and add to your hosts-file
111.222.222.111 myvpnserver
With the help of a script, you can change the file to
222.111.111.222 myvpnserver
and SE will use the new IP at the next reconnect.