Robert DiFalco wrote:

> Actually, I'm not concerned about down time do to FB having a problem.
> I'm worried about down time do to the server FB runs on having a
> problem. Right now, it is difficult to do a hot fail-over with FB. This
> is an issue at large enterprises. Even if FB could run without a
> software glitch for years yet could not be easily setup to do a hot
> failover, that would be enough to keep it out of some enterprises. This
> is the primary reason many databases support clustering; so that a
> mission critical application can continue uninterrupted even when "a"
> server goes down -- it's not usually to compensate for the server
> "software" being unreliable. JMTC.

On some of my PHP setup's, if the main Firebird Server does not respond,
the connection is switched to the backup machine. The users just see a
slightly longer delay in a page loading. 'Load Sharing' is done in a
similar way, since 90% of users are read only. Even where it is not
automatic, just changing a single entry switches to a backup machine.

This is just a matter of system design, clustering is not necessarily
the only solution and I haven't needed to worry about it (yet). Just a
fast network connections and redundant routing.

--
Lester Caine
-----------------------------
L.S.Caine Electronic Services