dbcc checkstorage scales linearly with aggregate I/O throughput for a substantial performance improvement over dbcc checkalloc. The scaling property of dbcc checkstorage means that if the database doubles in size and the hardware doubles in capacity (realizable I/O throughput), the time required for a dbcc check remains unchanged. Doubling the capacity would typically mean doubling the number of disk spindles and providing sufficient additional I/O channel capacity, system bus capacity, and CPU capacity to realize the additional aggregate disk throughput.
Most of the checks performed by using dbcc checkalloc and dbcc checkdb, including text column chain verification, are achieved with a single check when you use dbcc checkstorage, thereby eliminating redundant check operations.
dbcc checkstorage checks the entire database, including unused pages, so execution time is relative to database size. Therefore, when you use dbcc checkstorage, there is not a large difference between checking a database that is nearly empty and checking one that is nearly full, as there is with the other dbcc commands.
Unlike the other dbcc commands, the performance of dbcc checkstorage does not depend heavily on data placement. Therefore, performance is consistent for each session, even if the data placement changes between sessions.
Because dbcc checkstorage does extra work to set up the parallel operation and records large amounts of data in dbccdb, the other dbcc commands are faster when the target database is small.
The location and allocation of the workspaces used by dbcc checkstorage can affect performance and scalability. For more information on how to set up the workspaces to maximize the performance and scalability of your system, see “dbcc workspaces” in Chapter 59, “dbccdb Tables” in the Reference Manual.
To run dbcc checkstorage and one of the system procedures for generating reports with a single command, use sp_dbcc_runcheck. For information on the report generating system procedures, see “Generating reports from dbccdb”.
