Update to my question above: we had to restore the ntuser.dat file to resolve that issue but we have still been unsuccessful.
Eugene - in order to test out a "real life" disaster recovery, what would you recommend we delete .zdb files, registry, etc. because deleting someone from the BB Manager wouldn't prove to be a "real life" example. What is a good test case scenario that we could emulate?
Please try to use instructions posted by RIM that i referenced in the beginning of the thread:
From the testing prospective, i would try to execute as follows:
1. configure BES1(production server), add, provision and activate users.
2. stop BB services on BES1 and backup DB, registry and zdbs
3. shutdown BES1
4. configure BES2(disastery recovery) with the same BES name, SRP ID and licenses as BES1
5. stop BB services on BES2 and restore DB, registry and zdbs from BES1
6. restart BES2
7. Verify that service on user devices has been restored
8. Shutdown BES2 and bring BES1 up
Starting from this moment on, in case if production server(BES1) goes down you will just need to bring disaster recovery(BES2) up and email service on devices should be automatically restored.
Please also note that i have certain amount of confidence that above should work however i haven't tried it myself.
As a FYI, in case anyone was interested, this did not work. Due to the fact we have to stop the controller each time a backup is performed, which causes issues on its own, we have decided not to pursue this matter further.
We are going to a virtualize our BES environment in the next few months, in the mean time we're just going to do disk mirroring.