In my opinion (not saying im right), the answers for the new questions are:
4. a. Use dynamic management views
12. b. Grant the ALTER TRACE permission to the developers.
16. SELECT TWO:
b. Register a server principal name (SPN) for SQLSERVER1 and SQLSERVER2
and also
d. Create a map for each SQL login from SQLSERVER1 to SQLSERVER2 and use the impersonate option.
21. a. Drop and recreate the table using the FILESTREAM data type.
24. I have no idea about this one, page level but which one i dont know.
25. Select TWO:
c. Set the database recovery model to FULL
and
e. Set the PAGE_VERIFY database option to TORN_PAGE_DETECTION
26. a. Disable partition-level lock escalation
28. c. 64-KB
29. Not sure but i suspect b. Execute the following command on the server: exec sp_configure 'maximum degree of parrallelism",8
35. d. create three different non-administrative local user accounts and assign them to each of the services.
28. b. Kerberos authentication.
40. c. Enable the nested triggers server configuration option
41. a. Enable the nested triggers server configuration option. (i think this is the answer, not sure that they would have it twice in the exam though).
42. b. Use a logon trigger.
43. a. Set only Execute on SetCustomerStatus
47. d. Define a SQL server policy and import it on each database server
51. d. Transactional replication that uses updating subscriptions (im just guessing this one).
54. b. Configure manual failover role switching.
64. c. Backup the database master key
65. Two part answer (not 100% sure if this is correct):
b. Login accounts
and
c. Master database
66. No idea about this one, i'm guessing b. Perform a partial restore of the PRIMARY filegroup and filegroupB from a partial backup.
****************************************************************************************************************************************************
the following questions i got different answers /thoughts about, so please talk about
12. c. policyAdministratorRole grants the required permissions
21. d. stram_blob_columns = true in my opinion suits best together with merge replication (+ Newid())
24. b. row-level data compression has less performance impact as requested.
25. c. + d. identify I/O path errors seems to me more checksum feature
29. a. I understand that there is just 0 or 1 values and 1 is the value to stop the parallism feature
51. c. each site has to work their own data (update/insert etc.) so it should be peer-to-peer replication
54. d. why manual failover? wont automatic not reduce the time of reaction?
65. a + c login accoutns are in the master database included
best regards
andreas