Ken Chou
posted this
03 March 2009
Predictive Data Migration (PDM):
Replace un-healthy disk member in array, and keep array on normal status during the datatransition between healthy HD and replaced HD, Bad Sector Mapping, Media Patrol, SMART, Hard/Soft Reset to recover HD from bad status, HD Powercontrolto recover HD from hung status. Array Error Recovery: Data recovery from bad sector or failed HD for redundant RAID, RAID 5/6 inconsistent data Prevent (Write Hole Table), Data content Error Prevent (Read/Write Check Table) NVRAM event logging
Predictive Data Migration (PDM) is the migration of data from the suspect disk drive to a spare disk drive, similar to Rebuilding a Logical Drive. But unlike Rebuilding, PDM constantly monitors your disk drives and automatically copies your data to a spare disk drive before the disk drive fails and your Logical Drive goes Critical. After the data is copied from the suspect disk drive, the controller marks it with a
Stale configuration and a PFA error.
You can clear the Stale configuration and PFA error and put the disk drive back into service. In some cases, however, you might remove the disk drive for repair or replacement.
PDM Triggers:
The following actions trigger PDM:
• A disk drive with unhealthy status (see below)
• Media Patrol finds a disk critical error*
• You initiate PDM manually
*PDM also counts the number of media errors reported by Media Patrol.
A disk drive becomes unhealthy when:
• A SMART error is reported
• The bad sector remapping table fills to the specified level
Because data would be lost if written to a bad sector, when a bad sector is detected, the disk drive creates a map around it. These maps are saved in the bad sector remapping table, which has a capacity of 512 reassigned blocks and 1024 error blocks. When the table fills to a specified percentage of its capacity, PDM triggers a migration of data from the suspect drive (the disk drive with the bad sectors) to a spare disk drive.
During data migration, you will have access to the Logical Drive but it will respond more slowly to read/write tasks because of the additional operation. The time required for data migration depends on the size of the disk drive.
