LVS on FreeBSD vs LVS on Linux

[Robert Bernier]

Hi,

I'd appreciate your input.

I've been doing research on high availability clustering for database usage and I'm currently suffering indecision as to which OS I should be using.

What's the current Feature difference between LVS Linux and LVS FreeBSD?

What are your thoughts between: a 100% pf solution on FreeBSD, using LVS on Linux, a hybrid combination (if that's reasonable) using pf/psync/carp and LVS on FreeBSD?

What appeals to me about LVS are the sheer number of available load balancing algorithms and it dynamic feedback system. Contrary to pf where I don't, as yet, have the experience emulating the same kind of behaviour nor am I keen on reinventing the wheel. But on the other hand, I love the FreeBSD failover mechanism of psync, and carp (the retention of state seems so simple and elegant).

Thank you for your time.


best regards

Robert Bernier