load balanced configuration

[Alex Dean]

On May 19, 2010, at 4:07 PM, keith smith wrote:

>
> In a nutshell I am trying to create an automated backup that is a  
> automated fail over solution.

Keep in mind that failover/HA is not the same as backup.  It makes  
sense that both servers can do both jobs (you could keep your backups  
on both your production servers), but the failover/replication process  
should not be considered a backup process.

I understand 'high availability' to mean that both servers will have  
access to the same (or nearly the same) data, so either one could be  
primary.  In that case, anything which happens on your primary should  
automatically be replicated/synced/whatever to your secondary.  If you  
get hacked and someone does 'DROP DATABASE yourcart;', it's gone from  
both servers.

Making an hourly or daily backup of your data & critical files, and  
storing them elsewhere, is crucial.  Then... practice the steps needed  
to actually use those backups.  Discovering that your backups require  
18 hours of restore time before you can get them into production is  
awful.  Doubly awful if you discover it only when you actually need to  
use them.  Yes, it happened to me. :)

alex