Table of ContentsView in Frames

How the maintenance center helps prevent drive errors

Data ONTAP provides a mechanism to test drives called the maintenance center. Sometimes Data ONTAP puts drives into the maintenance center automatically; you can also put a suspect drive into the maintenance center manually. Knowing how the maintenance center works helps you manage your storage effectively.

When a disk is in the maintenance center, it is subjected to a number of tests. If the disk passes all of the tests, it is redesignated as a spare. Otherwise, Data ONTAP fails the disk.

The maintenance center is controlled by the disk.maint_center.enable option. It is on by default.

Data ONTAP puts disks into the maintenance center only if there are two or more spares available for that disk.

You can control the number of times a disk is allowed to go to the maintenance center by using the disk.maint_center.allowed_entries option. The default value for this option is 1, which means that if the disk is ever sent back to the maintenance center, it is automatically failed.

You can also put a disk into the maintenance center manually by using the disk maint start command. If the target disk is in use, it does not enter the maintenance center until its contents have been copied to another disk (unless you include the -i option).

Data ONTAP informs you of these activities by sending messages to the following destinations:

When Data ONTAP puts a disk into the maintenance center and that disk is housed in a storage shelf that supports automatic power cycling, power to that disk might be turned off for a short period of time. If the disk returns to a ready state after the power cycle, the maintenance center tests the disk. Otherwise, the maintenance center fails the disk immediately.

You can see the power-cycle status for ESH4 storage shelves by using the environment shelf_power_status command.

You can access the options and commands to control the maintenance center by using the nodeshell. For more information about the nodeshell, see the man page for the system node run command.

For information about best practices for working with the maintenance center, see Technical Report 3437: Storage Best Practices and Resiliency Guide.