MaintainingYourDatabase•Chapter7 297 Performing Ongoing Maintenance SQL Server Agent is a service that can create multi step jobs, notification alerts, and notify operators by e-mail, pager or NET SEND command. SQL Server Agent service must be running for jobs and alerts to execute. It is also key to system functionality like replication, Policy-Based Management and data collector. The Database Maintenance Plan wizard allows sysadmins to create, modify, schedule and run customized maintenance plans. A Maintenance Plan is a sequence of ongoing maintenance tasks like rebuilding indexes, validate databases, and perform backup. Policy-Based Management is a new feature of SQL Server 2008 that allows you to create and apply policies to multiple target servers or database objects. A policy consists of a set of conditions that must be met by the object for it to be compliant with the policy. Objects that do not comply with the policy are reported to you when the policy is evaluated. Policies can be evaluated manually or automatically. This is known as “on demand”, “on change: prevent”, “on change: log only” and “on schedule” policy evaluation modes. These modes are used to check whether the target complies with the policy, and possibly prevent a change that will cause the object to become noncompliant. The “on change: prevent” and “on change: log only” policies are enforced using recursive DDL triggers. The “on schedule” evaluation is performed using SQL Server Agent. Performance Data Collection The Data Collector is a feature of SQL Server 2008 that allows you to systematically collect data from various sources and store it in a data warehouse. The Data Collector can be used to collect performance data, but it is not limited to this use. The Data Collector uses SQL Agent jobs to collect data, and SQL Agent must be running for the Data Collector to work. SQL Server Reporting Services reports are created on collected data. Built-in reports are available, or you can create custom reports. 298 Chapter7•MaintainingYourDatabase Exam Objectives Frequently Asked Questions Q: What happens to the database collation when a database is restored from backup to a server that has a different data collation? A: The database collation remains as it was on the source server. Q: How do I change collation settings at column and database level once they have been assigned? A: Use ALTER TABLE and ALTER DATABASE statements with the COLLATE clause. Unfortunately, there are many restrictions preventing this operation. Q: What happens when I join two tables, but the join keys have different collations? A: You will receive an error message similar to ‘Cannot resolve collation conflict between ‘Latin1_General_CS_AS’ and ‘Greek_CI_AS’ in equal to operation.’ Q: Which is more efficient: row-level compression or page-level compression? A: Page-level compression results in more efficient disk space usage as it enables row-level compression, prefix compression, and dictionary compression. However, page-level compression also results in higher performance overhead. Q: Can I create a column set on the entire table and update each row as an XML fragment? A: No, because column sets are defined on sparse columns and the primary key column cannot be marked as sparse. You must always have a primary key column in a table. Q: Is the Fill Factor of an index automatically maintained by SQL Server? A: No, the Fill Factor is only an initial setting. This is why you must rebuild your indexes as an ongoing task. Q: What is the overhead of Policy-Based Management on my SQL Server? A: Unless the policy is currently being evaluated, there is no overhead on SQL Server. Two exceptions are the “On Change: Prevent” and “On Change: Log MaintainingYourDatabase•Chapter7 299 Only” evaluation modes. These modes use DDL Triggers, which execute to perform a check every time an object in the policy’s target set is altered. This results in some overhead. Q: Can you manage earlier versions of SQL Server using Policy-Based Management? A: Yes, except not all settings are applicable to all versions of SQL Server. Q: When a SQL Server Agent job executes, what credentials are used? A: The credentials used are those of the SQL Server Agent service account, unless a proxy is explicitly configured or the statement is performing custom imper- sonation. Q: Is it better to create Maintenance Plans or individual SQL Server Agent jobs to perform my routine management and maintenance tasks? A: It is recommended that you use Maintenance Plans because they are easier to view and edit. Additionally, the Maintenance Plan Wizard will not let you misconfigure a task, which cannot be said for complicated homegrown Transact-SQL scripts. Q: Differential and log backups both record changes from the last full backup, so what is the difference between them? A: The difference between these backup types is in the way changes are backed up. The differential backup records all database extents that have changed since the last backup. When a differential backup is restored, all changes must be restored as there is no order or sequence of data changes. This makes the differential backup faster to restore than a log backup. Log backups record transactions in sequence. Replaying the log means redoing all transactions recorded in it. This is a time consuming operation, but it does allow you to roll forward to a point-in- time, provided all sequential log file backups are available. Q: I am missing a log file backup from the middle of a log sequence. Is it possible to restore all the other logs I have and roll the database forward, losing only the transactions recorded in the lost log? A: No, logs can only be restored in sequence. In this case, you can only roll the database forward to the point of the lost log. 300 Chapter7•MaintainingYourDatabase Q: How often should I back up the Master database? A: The Master database stores much of the server configuration information including security settings, database configuration, and logins. You should back it up regularly in the live environment. The exact frequency depends on your organization’s tolerance for loss of configuration data and how often this config- uration changes. Q: I have configured the performance Data Collector, but it has collected no data. Why? A: The most likely reason that the Data Collector is not collecting data is that the SQL Agent is not running. The Data Collector uses SQL Agent jobs to perform data collection. MaintainingYourDatabase•Chapter7 301 Self Test 1. You have a large centralized line of business database used by hundreds of users all around the world. One of the key tables is the Orders table. This table is frequently updated and frequently searched. A nonclustered index named IX_Orders_CustomerID exists on the Orders table, CustomerID column. Recently your company has experienced record sales, and many new customers have been added. You have received several user reports of degraded performance when trying to search for orders by a specific customer. After running some diagnostics, you are convinced that the reason for the performance degradation is index fragmentation. You wish to resolve the problem for your users, but, due to the 24-hour nature of your business, you are not allowed to take the index offline. Additionally, you wish to be able to interrupt reindexing operations without losing progress, in case the maintenance operation affects performance severely while it is running. What is the best way to resolve the problem while meeting your objectives? A. Run ALTER INDEX _Orders_CustomerID ON Orders.CustomerID REBUILD B. Run ALTER INDEX IX_Orders_CustomerID ON Orders.CustomerID REORGANIZE C. Run CREATE INDEX IX_Orders_CustomerID ON Orders. CustomerID WITH DROP EXISTING D. Run DBCC INDEXDEFRAG (LOBDatabase, “Orders”, IX_Orders_ CustomerID) 2. You have recently removed a lot of historical data from your SQL Server 2008 database named HR by archiving it to a different server. However, you notice that the disk space consumed by the database file remains unchanged and the .MDF file is taking up the majority of space on the drive. How can you decrease the size of the database file? A. Run DBCC CHECKDB (‘HR’, REPAIR_ALLOW_DATA_LOSS) B. Run DBCC SHRINKFILE (‘HR_Data’, TRUNCATEONLY) C. Run DBCC SHRINKFILE (‘HR_Data’, NOTRUNCATE) D. Use the Copy Database Wizard to copy the database, then delete the original database. Finally, rename the new database you have just copied to the same name as the original database. . task. Q: What is the overhead of Policy-Based Management on my SQL Server? A: Unless the policy is currently being evaluated, there is no overhead on SQL Server. Two exceptions are the “On Change:. recursive DDL triggers. The “on schedule” evaluation is performed using SQL Server Agent. Performance Data Collection The Data Collector is a feature of SQL Server 2008 that allows you to systematically. Transact -SQL scripts. Q: Differential and log backups both record changes from the last full backup, so what is the difference between them? A: The difference between these backup types is in the