AMCC 720-0138-00 ユーザーズマニュアル

Chapter 6. Maintaining Units

72

Verification makes use of the same error checking and error repair techniques 
used during ordinary use of drives configured through 3ware RAID 
controllers. 

When verification encounters an error, the controller typically retries the 
command. If there are cable CRC errors, there may be multiple retries 
including downgrade of the UDMA mode. If the error persists and is 
unrepairable (e.g., ECC errors), an error notification is issued to indicate the 
problem. (See AEN “0026 Drive ECC error reported” on page 122.) 

If the disk drive is part of a redundant unit that is in a redundant state (not 
degraded or rebuilding), then Dynamic Sector Repair automatically rewrites 
the redundant data to the error location to force the drive to reallocate the 
error location. A notification of repair is posted to the alarms list. The result is 
a restoration of drive and data integrity; the primary and redundant data are 
again both valid. 

If the unit is not redundant, it is recommended that you perform a file-system 
check to correct the issue. On Mac OS X, you can do this using the First Aid 
tab in the Disk Utility—select the disk on the left and then click 

. 

If verification encounters problems, you can then use the 

 option 

on the same screen.If the errors persist and cannot be overwritten from a 
backup copy, perform a final incremental backup. You will need to replace the 
defective drive, recreate the array, and reinstall the data. 

Starting a Verify Manually

Verification of units can be done automatically, on a schedule, or can be 
started manually, as described below. (See “Setting Auto Verify for a Unit” on 
page 46 and “Scheduling Background Tasks” on page 76.)

                                      

 

If the unit has not previously been initialized and you manually select

 

Unit the initialization process starts

.