Thanks Elliott.

Site1 has 90% of the users so i what i was thinking is sql1 would take all the requests from site1 and site2.

If sql1 is down, all the requests will go to sql2.

Peer-to-Peer, as i understood it, can replicate changes both ways, so if sql1 was taking all the requests, all the changes will be replicated to sql2. if sql1 went down, sql2 will be taking all the requests and once sql1 comes back up, sql1 will start taking all the requests and all the changes that happened on sql2 will be replicated to sql1.

now, would that work with a windows cluster where i can specify that sql1 server is the primary and sql2 server is the backup?

thanks again.