The just released E2k7 MP has two important bugs in it. I communicated this to the product group, and they are working on releasing a fix. Meanwhile workarounds are (thanks to Ake for quick response to this):
1. Discovery discovers E2k3 clustered backend servers as Ex.Common and All Exchange 2007 roles.
To resolve this follow:
1. Create a group in OpsMgr consisting of all the E2003 cluster instances (the target is Windows Server so you need those instances)
2. Disable the Microsoft.Exchange.2007.Microsoft_Exchange_2007_All_Servers_Installation.Discovery for the group
3. Disable the Ex.Common discovery for the group
4. Run remove-disabledmonitoringobject to remove the discovered instances (note that this will remove all instances with discovery disabled, not just the Exchange specific ones). This should remove the all server installation instances and remove the monitoring from the servers.
2. The disk space monitoring rules do not have alert suppression and have a 60seconds frequency. This will result in alert storms when an E2k7 server runs out of disk space.
I worked around it by disabling the disk space related rules and monitors bound to Ex.Common/Exchange 2007 All servers object types.
Since Base OS MP ships with good disk monitoring logic, I decided this is sufficient.