Understanding the dbcc checkstorage operation

The dbcc checkstorage operation consists of the following steps:

1. Inspection – dbcc checkstorage uses the device allocation and the segment definition of the database being checked to determine the level of parallel processing that can be used. dbcc checkstorage also uses the configuration parameters max worker processes and dbcc named cache to limit the level of parallel processing that can be used.

2. Planning – dbcc checkstorage generates a plan for executing the operation that takes advantage of the parallelism discovered in step 1.

3. Execution and optimization – dbcc checkstorage uses Adaptive Server worker processes to perform parallel checking and storage analysis of the target database. It attempts to equalize the work performed by each worker process and consolidates the work of under utilized worker processes. As the check operation proceeds, dbcc checkstorage extends and adjusts the plan generated in step 2 to take advantage of the additional information gathered during the check operation.

4. Reporting and control – during the check operation, dbcc checkstorage records in the dbccdb database all the faults it finds in the target database for later reporting and evaluation. It also records the results of its storage analysis in dbccdb. When dbcc checkstorage encounters a fault, it attempts to recover and continue the operation, but ends operations that cannot succeed after the fault. For example, a defective disk does not cause dbcc checkstorage to fail; however, check operations performed on the defective disk cannot succeed, so they are not performed.

   If another session performs drop table concurrently, dbcc checkstorage might fail in the initialization phase. If this happens, run dbcc checkstorage again when the drop table process is finished.