You can verify persistent faults by running checkverify anytime after running checkstorage, even after an extended period of hours or days. However, when deciding your schedule, keep in mind that the database state changes over time, and the changes can mask both soft faults and hard faults.
For example, a page that is linked to a table but not allocated is a hard fault. If the table is dropped, the fault is resolved and masked. If the page is allocated to another table, the fault persists but its signature changes. The page now appears to be linked to two different tables. If the page is reallocated to the same table, the fault appears as a corrupt page chain.
Persistent faults that are corrected by a subsequent database change usually do not pose an operational problem. However, detecting and quickly verifying these faults may locate a source of corruption before more serious problems are encountered or before the signature of the original fault changes. For this reason, Sybase recommends that you run checkverify as soon as possible after running dbcc checkstorage.
When checkstorage is executed with
the target database in single-user mode, there will be no soft faults
and no need to execute checkverify.
checkverify runs only one time for each execution of checkstorage. However, if checkverify is interrupted and does not complete, you can run it again. The operation resumes from where it was interrupted.
checkverify may take a long time to complete when processing very large databases. During this time, checkverify does not provide you with any indication of when it will finish. To get a status for the checkverify, use the command_status_reporting command. When the check_status_reporting is on, the checkverify gives progress status reports. The default value is off.
To turn the option on, issue the following command:
set command_status_reporting on
The results appear as:
1> dbcc checkverify (pubs2) 2> go
Verifying faults for’pubs2’. Verifying faults for table 't1'. The total number of tables to verify is 5. This is table number 1. Verifying faults for table 't2'. The total number of tables to verify is 5. This is table number 2. Verifying faults for table 't3'. The total number of tables to verify is 5. This is table number 3. Verifying faults for table 't4'. The total number of tables to verify is 5. This is table number 4. Verifying faults for table 't5'. The total number of tables to verify is 5. This is table number 5. DBCC CHECKVERIFY for database ’pubs2’ sequence 4 completed at Apr 9 2003 2:40PM. 72 suspect conditions were resolved as faults, and 11 suspect conditions were resolved as harmless. 0 objects could not be checked.
