> Anmol Sheth wrote:
>
> Hi,
> In the LVS HOWTO there was a short paragraph in the Failover Section (
> Section 12 ) about the realserver taking over the director in case of
> a director failure. Effectively the same code running on all the nodes
> of the LVS cluster. Is this already implemented by someone ?? If it
> is, I would be really glad if someone can direct me to some more
> documentation and code.
I did an installation where I had two nodes running LVS/redirector in
conjunction with heartbeat and the real services on it (apache). In
normal state one node would run the director and directs the traffic to
itself and the other node. In case the directing node fails the second
node takes over the cluster IP and directs the traffic to itself. When
the failed node comes back the second node directs the traffic to itself
and the first node.
So you have high available load balancing with just two machines. You
can add additional real servers anytime you need them.
Maybe this is what you are looking for?
juri
--
juri.haberland@xxxxxxxxxxxxx
system engineer innominate AG
clustering & security networking people
phone: +49-30-308806-45 fax: -77 http://innominate.de
|