[lug] High Availability and Failover options

[Alan Robertson]

"Riggs, Rob" wrote:
> 
> I agree that it is done. I do not agree that it common.

Hard to argue with that when  High-Availability (failover) systems are
uncommon to start with ;-)

> There are frequently
> better ways of investing one's money to achieve reliability.
> 
> -Rob
> 
> -----Original Message-----
> From: Alan Robertson [mailto:alanr at unix.sh]
> 
> I believe that people really are successful at keeping session state (for
> HTTP) in databases and fail them over.

It's a design decision which should (preferably) be made earlier rather than
later -- according to what you believe your real availability and
performance needs are.  Failovers are normally rare - even though 2/3 of
them are done for administrative reasons (and not for real failures).

The more high-availability you want to achieve, the more work you have to be
willing to put out - and Rob rightly points out that this is only a small
piece of a rather large puzzle. TANSTAAFL.

	-- Alan Robertson
	   alanr at unix.sh