A rebuild does not start after replacing a defunct drive - IBM ServeRAID



Source

RETAIN tip: H192840

Issue

There are several symptoms related to the same issue:

  1. After replacing a defunct drive in a redundant RAID configuration with a previously used or new hard disk drive, the disk does not start rebuilding automatically.
  2. After replacing a defunct drive with a previously used or new hard disk drive, the physical drive immediately transitions to an online state, but the logical drive remains critical.
  3. After replacing a missing drive in a mirrored pair with a drive previously used in the same mirrored pair, the physical drive immediately transitions to an online state and data corruption occurs.
  4. After adding a previously used or new hard disk drive to expand a logical volume, the disk does not start rebuilding automatically.

Affected configurations

This tip is not hardware specific.

The system is configured with one or more of the following IBM Options:

  • ServeRAID-8i Controller, Option 13N2227
  • ServeRAID-8k SAS Controller, Option 25R8064
  • ServeRAID-8k-l SAS Controller, Option 39R8729
  • ServeRAID-8s SAS PCIe Controller, Option 39R8765

This tip is not software specific.

Solution

The proper procedure when adding any previously used good or absolutely new drive to an array, either as a replacement drive or to expand an array, is to first initialize the drive.

Reusing good hard disk drives within a RAID configuration requires the drive to be initialized to remove existing configuration information. Existing configuration information may cause the ServeRAID controller to behave differently than expected and in some rare cases, can result in data loss.

Hard disk drives can be initialized by using the Array Configuration Utility (ACU), the IBM Support CD, the ServeRAID Manager installable application, or by using the ARCCONF TASK sub-command within the CLI.

Offline Initialization Method:

  1. Power off the server.
  2. Insert the "previously used" or "absolutely new" good hard disk drive.
  3. Power on the server. When prompted, use CTRL-A to launch the ACU.
  4. From the main menu, choose "Array Configuration Utility". Then, choose the "Initialize Drives" submenu.
  5. Select the replaced drive by the slot location and initialize the replacement drive by Enclosure ID and Slot number location. The drives locations are presented as follows:
  Dev:Box:Slot
00:00:00
01:00:01
02:00:02
03:00:03
04:00:04
05:00:05

The first and second digits is the Device ID. The third and fourth digits is the Enclosure ID. The fifth and sixth digits is the slot location which should match the label on the front bezel of the server or external chassis.

Important Note: Initializing the wrong disk, in an existing array, will cause data loss if the array is already critical, or will cause an optimal array to become critical. Ensure to properly correlate the physical slot location to the software presentation.

The referenced file will be available from the Servers - ServeRAID Software Matrix.

Additional information

The ServeRAID-8 series controllers will always try to use the values stored in the configuration data on newly inserted hard disk drives to determine how the hard drive will be used within the current working configuration. Depending on the values and how the drive was previously used, the controller may ignore all values and behave as expected by automatically rebuilding, or in other situations it will try to use those values resulting in the unexpected behavior as described in the symptoms above.

In most cases, users expect the default behavior to be what occurs when a new unused drive is inserted. The process of initializing drives will cause the ServeRAID controller to write meta-data correctly to the previously used or absolutely new hard disk drive.

Applicable countries and regions

 


Document id:  MIGR-5074510
Last modified:  2008-12-01
Copyright © 2014 IBM Corporation