IBM ServeRAID, User Reference

Page: 174 / 226

Note: (For a configuration that contains a hot-spare drive) If you replace a
failed physical drive, it is not necessary to position the new physical drive
on the same SCSI channel and SCSI ID as the original hot-spare drive.
The replacement physical drive is automatically incorporated into the
configuration as a hot-spare drive. Here is an example of how this works:
a. The original configuration consists of a RAID level-5 logical drive
composed of four physical drives. The physical drives are connected
to SCSI channel 1; they are assigned SCSI IDs 0, 1, 2, and 3. SCSI
ID 3 is a hot-spare drive.
b. The physical drive at Channel 1, SCSI ID 2, fails; the logical drive
enters the critical state.
c. The hot-spare drive at Channel 1, SCSI ID 3, is rebuilt into the array.
d. You remove the failed physical drive at Channel 1, SCSI ID 2, and
replace it with a new physical drive. The new physical drive at
Channel 1, SCSI ID 2, is automatically assigned to be a hot-spare
drive.
2. If a rebuild operation is in progress, wait until the rebuild is complete. Otherwise,
go to step 3.
Note: If you are replacing multiple defunct drives, you must wait for each
rebuild operation to complete before starting subsequent rebuild
operations.
3. Verify the cables, physical drives, and controllers are installed properly.
4. Attempt to rebuild the defunct physical drive by performing a hot-swap rebuild.
See “Rebuilding a hot-swap drive” for instructions.
5. If the hot-swap rebuild fails, contact your IBM service representative.
Rebuilding a hot-swap drive
A hot-swap rebuild refers to a rebuild operation that is started by the ServeRAID
controller when it detects that a drive that is part of an array and in the defunct
state has been removed and reinserted on the SCSI cable or backplane. The
reinsertion of the physical drive, whether it is the same drive or a new drive, will
trigger the ServeRAID controller to start the rebuild operation. During the rebuild
operation, the drive being rebuilt is in the rebuild state, and the logical drive remains
critical until the rebuild operation has been successfully completed.
On IBM servers, when a hot-spare drive is available, the rebuild operation begins
automatically without the need to replace the failed drive. If more than one drive
fails within the same array, no rebuild takes place. If multiple drives fail in separate
arrays (one physical drive per array), the controller initiates a rebuild operation for
the logical drives within the array containing the first failed physical drive. This
rebuild operation is performed on the
first hot-spare drive of sufficient size to
become a valid member of the array.
Complete the following steps to start a hot-swap rebuild:
1. Without removing the drive completely, gently remove the physical drive from
the server, using the handle of the hot-swap tray. If necessary, see the
documentation that comes with your server for information about removing a
physical drive.
158
IBM ServeRAID: User’s Reference, Version 5.10