Kevin,

> -----Original Message-----
> From: kgdayau [mailto:kdp@...]
> Sent: Tuesday, December 07, 2004 9:25 PM
>
> I have to assure the customer that the end result will be
> 100% reliable with fail safe mechanisms in place. I am
> considering that I recomend a backup database server (with a
> replicator addon product
> operating) that can become the primary with only small
> amounts of configuration and a maximum of about ten minutes
> from the I.T. staff (no dedicated DBA on staff). The I.T.
> mentioned server clustering which I know nothing about.

Someone mentioned a few months ago that they were either testing or
were using a cluster to provide automatic switching between servers
upon server failure. The key was a shared external raid enclosure or
san. You'll have to check the archives for details.

I've been looking to take the approach you mention above - replication
to a backup server, which can replace the primary server with a little
manual intervention.

> The server will be a dual
> processer Xeon, running Win Server 2003 with separate raid
> for DB and disk for OS - all SCSSI, RAM >= 1 GB.

I would suggest that you use the Classic version of FB with these
hardware specs. I moved from NT4 with FB Superserver to FB Classic on
Win2003 and dual Xeons with hyperthreading enabled. The performance
has been excellent, and fixed one of the problems with the older
configuration - a user running a large report would slow down, or even
block, all of the other users while the report ran. Now I can see
that one CPU maxed out at 100% utilization, while the other three are
available to service other users.

Bob M..