On 4/26/05, Sean Kennedy <skennedy@tpno-co.org>
wrote:> Hi folks,
>
> I'm curious; What does everyone do for failover? I have two servers,
> same os/compilation. I designate one the master, the other the slave,
> and I rsync the config files once an hour and trigger a restart when
> convenient command on the console. These two servers are setup in the
> dns in a round robin fashion.
>
> What is everyone else doing?
That's kind of a loaded question... Do you plan on expanding? What
is your budget? What are your uptime requirements? Are you serving
customers or is this just for internal use?
Round robin dns is a cheap way of doing load balancing, not failover.
If a server fails, you will still have requests going to the dead
server.
I always prefer to keep volatile data in a real database, and make
sure that if anything is redundant, it's the database. You can
survive downtime, but you might not survive the loss of critical data.
Automate your backups, do them often, and keep a copy both on site
and off site. Keep backups of data beyond what you think you will
need them for at the time, because invariably you will need something
that you deleted 6 months ago.
If you are on a budget and want to just use what you have, then I
would keep your same setup but dump the round robin dns.
Chris