MegaRAID Storage Manager Online Help System

Maintaining and Managing Storage Configurations > Rebuilding a Drive

Rebuilding a Drive

If a drive in a mirrored virtual drive (RAID 1 or RAID 1E) fails, MegaRAID Storage Manager software automatically rebuilds the data on a hot spare drive to prevent data loss. The rebuild is a fully automatic process, so it is not necessary to issue a Rebuild command. You can monitor the progress of drive rebuilds in the Group Show Progress window. To open this window, select Group Operations->Show Progress.

A failed drive must be replaced, and the data on the drive must be rebuilt on a new drive to restore the system to fault tolerance. (You can choose to rebuild the data on the failed drive if the drive is still operational.) If hot spare drives are available, the failed drive is rebuilt automatically without any user intervention.

A red circle to the right of the drive icon indicates that a drive has failed. A yellow circle appears next to the icon of the virtual drive that uses this drive. This indicates that the virtual drive is in a degraded state; the data is still safe, but data could be lost if another drive fails.

When the new drive spins up, the drive icon changes to green as the rebuild begins, and an up-arrow appears above the drive icon.

Follow these steps to rebuild a drive:

  1. Right-click the icon of the failed drive, and select Rebuild.
  2. Click Yes when the warning message appears. If the drive is still good, a rebuild will start.
  3. You can monitor the progress of the rebuild in the Group Show Progress window by selecting Manage->Show Progress. If the drive cannot be rebuilt, an error message appears. Continue with the next step.
  4. Shut down the system, disconnect the power cord, and open the computer case.
  5. Replace the failed drive with a new drive of equal capacity.
  6. Close the computer case, reconnect the power cord, and restart the computer.
  7. Restart the MegaRAID Storage Manager software.
  8. When the new drive spins up, the drive icon changes back to normal status, and the rebuild process begins automatically. You can monitor the progress of the rebuild in the Group Show Progress window by selecting Manage->Show Progress.

If you want to force a drive into Fail status in order to trigger a rebuild, right-click the drive icon, and select Make Drive Offline. A red circle appears next to the drive icon. Right-click the icon and select Rebuild from the pop-up menu. A drive rebuild cannot be aborted.

A drive rebuild is also started if you select Make Drive Online from the pop-up menu.

New Drives Attached to a MegaRAID Controller

When you insert a new drive on a MegaRAID system, if the inserted drive does not contain valid DDF metadata, the drive displays as JBOD for MegaRAID entry-level controllers, such as the SAS 9240-4i/8i. If the drive does contain valid DDF metadata, its drive state is Unconfigured Good.

A new drive in JBOD drive state is exposed to the host operating system as a stand-alone drive. Drives in JBOD drive state are not part of the RAID configuration because they do not have valid DDF records. The operating system can install and run anything on JBOD drives.

Automatic rebuilds always occur when the drive slot status changes, for example, when you insert a drive or remove a drive, so that a hot spare can be used. However, a new drive in JBOD drive state (without a valid DDF record), does not perform an automatic rebuild.

To start an automatic rebuild on the new JBOD drive, you have to change the drive state from JBOD to Unconfigured Good. (Rebuilds start on Unconfigured Good drives only.) After you set the drive state to Unconfigured Good, the drive state information always remains on the drive, and you can use the drive for configuration.


DB09-000202-04  37857-02  Rev. E  October 2010
Copyright© 2010 by LSI Corporation. All rights reserved.
You can find a list of the LSI U.S. distributors, international distributors, sales offices, and design resource centers on the LSI web site at:  http://www.lsi.com/cm/ContactSearch.do