Re: [Exim] exim in large-scale production environment

Previously, all of the IMAP/POP/SMTP tasks were load balanced between 3
Sun servers. It got to the point where our load was two high for the
servers. Rather than just adding another one to the pool, we decided to
split the mail system up for security and performance reasons.

Rather than put all of our eggs in one basket and buy another enterprise
server; it made more sense to go for more less-expensive servers
(actually, they are Sun Ultra workstations) to deal with POP/IMAP.

Our servers are reliable and we have not had a complete un-expected outage
in well over a year; but, occasionally we do need to stop parts of the
mail-system to deal with things such as OS patches or hard-drive failure. It
is at times like this that our infrastructure proves useful.

-JV

On Thu, Jul 20, 2000 at 08:21:20AM -0700, Tom Samplonius wrote:

[...]

> I have trouble grasping this. What kind of server were you using such
> that it failed more than once a year? I would say that if you
> experiencing more than one failure a year, you need a better server, not
> more of them. In fact you add more servers, the total number of failures
> a year will increase proportionally.

[...]

--
J.R. Van Valen
jrv@???