> I like it. Can't see anything wrong with it off the top of my head.
I've been using it for about a year -- so I'm sure it works. :)
> > The key difference in this scenario is that the director does NOT have
the
> > IP address installed; rather it is a router that knows how to get to
the
> > IP address.
The thing is the LVS DOES have to have the IP address installed in order for
LVS to get it.
LVS is *LIKE* a router -- but it is isn't exactly a router. In order for LVS
to grab each VIP, look at them, change how they are routed dynamically, and
then route them on, it will need the VIPs.
Just assign the VIPs to a hidden loopback device. Doing so will end up
getting the LVS to listen to *EVERY* IP on the entire network you assign to
the loopback device, which is VERY handy.
> > Rather than having to write our own failover and load balancing code for
> > the directors themselves, perfect it, etc, we can just use a proven
> > routing protocol (like BGP) that already has all that built in.
> If there are 2 directors (one running and one on standbye), how will BGP
> know that the running one has failed and switch over to the other?
Joe -- remember how we were talking about my LVS which elected a master
rather than depending on a pre-defined master and slave?
I did this by way of LVS-DR using routing. You've seen all of this before,
you just don't realize it. :)
All the best --
Ted
|