I've googled and come up with a whole bunch of crap surrounding clustering - this is for a non-clustered environment (although the DB may or may not be a part of a DB cluster)

2 AppServers (really 2 apps that could run on the same server but typically don't) - 1 DB Server (with 1 DB that both applications hit)

why would MSDTC be required? the developers at work say it's required because whenever a DB transaction occurs when more than 1 server is involved it's needed to be able to roll back transactions accurately when failures occur

please provide links/proof if possible

10/7/2009 12:02:14 PM

Quote :

"this is for a non-clustered environment (although the DB may or may not be a part of a DB cluster)"

This is key. I can't speak for the specific technologies you are working with, but typically if in a cluster, you have a "cluster manager" of sorts that helps coordinate everything. In this case, you don't have a cluster manager, so its everybody going solo by itself. As the devs have said, DB transaction rollback itself is the main concern, as you don't want to end up with 2 AppServers independently operating and simultaneously performing the same operation or trying to update the same information in the DB.

No idea why required, but, probably a good idea / alternative if a cluster manager of sorts isn't in the picture otherwise.

10/7/2009 8:16:55 PM