2.3.6 Handling Disk Replacement

If a disk has a problem, the physical disk status changes.

When a physical disk is removed, its status becomes not present. Oracle ASM may take a grid disk offline when getting I/O errors while trying to access a grid disk on the physical disk. When the physical disk is replaced, Oracle Exadata System Software automatically puts the grid disks on the physical disk online in their respective Oracle ASM disk groups. If a grid disk remains offline longer than the time specified by the disk_repair_time attribute, then Oracle ASM force drops that grid disk and starts a rebalance to restore data redundancy. Oracle ASM monitors the rebalance operation, and Oracle Exadata System Software sends an e-mail message when the operation is complete.

The following table summarizes the physical disk statuses, and how Oracle ASM handles grid disks when the physical disk has a problem.

Table 2-1 Physical Disk Status

Physical Disk Status	Oracle Exadata System Software Action
normal Disk is functioning normally	No action.
not present Disk has been removed	Oracle Exadata System Software offlines disk, then uses the DROP ... FORCE command after disk_repair_time limit exceeded. The rebalance operation begins.
predictive failure Disk is having problems, and may fail	Oracle Exadata System Software drops the grid disks on the affected physical disk without the FORCE option from Oracle ASM, and the rebalance operation copies the data on the affected physical disk to other disks. After all grid disks have been successfully removed from their respective Oracle ASM disk groups, administrators can proceed with disk replacement.
critical Disk has failed	Oracle Exadata System Software drops the grid disks using the DROP ... FORCE command on the affected physical disk from Oracle ASM, and the rebalance operation restores data redundancy. Administrators can proceed with disk replacement immediately. This status is only available for releases 11.2.3.1.1 and earlier.
poor performance Disk is performing poorly	Oracle Exadata System Software attempts to drop the grid disks using the FORCE option on the affected physical disk from Oracle ASM. If the DROP ... FORCE command is successful, then the rebalance operation begins to restore data redundancy and administrators can proceed with disk replacement immediately. If the DROP ... FORCE command fails due to offline partners, Oracle Exadata System Software drops the grid disks on the affected physical disk without the FORCE option from Oracle ASM, and the rebalance operation copies the data on the affected physical disk to other disks. After all grid disks have been successfully removed from their respective Oracle ASM disk groups, administrators can proceed with disk replacement.

After a physical disk is replaced, Oracle Exadata System Software automatically creates the grid disks on the replacement disk, and adds them to the respective Oracle ASM disk groups. An Oracle ASM rebalance operation relocates data to the newly-added grid disks. Oracle ASM monitors the rebalance operation, and Oracle Exadata System Software sends an e-mail message when the operation is complete.