[Users] Re: HN & VEs in multi-subnet VLAN

[Suno Ano]

 Tim> Our processes allow for "fail-over" from a failing server onto a
 Tim> good server. We do this by using OpenVZ. So, if "Server5" should
 Tim> fail (let's assume its CPU died), we would fail-over the "Server5"
 Tim> virtual environment (which includes its IP Address) onto "Server1"
 Tim> and clients will still be able to connect to their normal IP
 Tim> Address. Server1 would then have 2 virtual environments, one on
 Tim> 192.168.1.11/27 and the other on 192.168.2.21/27, and it's
 Tim> hardware node configured as 192.168.1.10/27.

sounds to be as if you were up to create a bridge on the HN and add the
VEs interfaces to that bridge


 Tim> A "fail-over" is not an automated process, but a restore from the
 Tim> last backup of the VE onto another hardware node.

with the link below and some modification that should be no problem

 Tim> Will OpenVZ support this type of setup without additional
 Tim> configuration?

yes, I would opt for
http://wiki.openvz.org/HA_cluster_with_DRBD_and_Heartbeat

Whatever you finally come up with, please put your notes etc. onto the
wiki -- maybe write a howto about it ...