Resolving a problem with dual boot drives

Complete the following steps to resolve most problems with dual boot drives.

Before you begin

When the IBM Storage Virtualize software starts, it reads identity information from the system board and compares it with the information stored on the boot drives in order to identify whether a system has had a fault or has had a part replaced.

Specific node errors are produced under the following conditions:
  • Node error 543: This error indicates that none of the identification information data stored in the drives or the system board match. IBM Storage Virtualize requires that at least two devices contain the same identity information.
  • Node error 545: This error indicates that the identity information on each boot drive match each other but are not the same as the identity information from the system board. If the node canister was replaced, use the command-line interface (CLI) command satask chbootdrive -replacecanister to update the data stored on the boot drives. If the canister was not replaced this error indicates that the node canister and boot drives are from different systems and need to be returned to their original configuration.
  • Node error 743: This error indicates that the identity information cannot be read from one of the two boot drives because that drive failed, is missing, or is out of sync with the other boot drive.
  • Node error 744: This error indicates that the identity information from one of the boot drives belongs to a different node. If boot drives were swapped between drive slots 1 and 2, node error 744 is produced.
  • Node error 745: This error indicates that a single boot drive system has booted with the boot device in slot 2 instead of slot 1.

About this task

An event is displayed in the Monitoring > Events pane of the management GUI if the problem produces node error 743, 744, or 745. Run the fix procedure for that event. Otherwise, connect to the technician port to see the boot drive slot information and use the serial number label on the node canister latch to determine the problem. Alternately, you can use the output of the satask lsbootdrive command.

  • Do not swap boot drives between slots.
  • Each boot drive has a versioned copy of the VPD on the system board.
  • Software upgrading is to one boot drive at a time to prevent failures during CCU.

Procedure

To resolve a problem with a boot drive, complete the following steps in order:

  1. Remove any drive that is in an unsupported slot. Move the drive to the correct slot if you can.
  2. If possible, replace any drive that is shown as missing from a slot. Otherwise, reseat the drive or replace it with a drive from FRU stock.
  3. Move any drive that is in the wrong node back to the correct node.
    Note: If the node serial number does not match the node serial number on the system board, a drive slot has a status of wrong_node. If the serial number on the node canister latch matches the node serial number on the drive, you can ignore this status.
  4. Move any drive that is in the wrong slot back to the correct slot.
  5. Reseat the drive in any slot that has a status of failed. If the status remains failed, replace the drive with one from FRU stock.
  6. If the drive slot has status out of sync and Yes in the can_sync column, then:
    • Use the service assistant GUI to synchronize boot drives, or
    • Use the command-line interface (CLI) command satask chbootdrive -sync.
    • If No is displayed in the can_sync column, you must resolve another boot drive problem first.

Replacing the node canister:

  1. Replace the node canister. Replacing the node canister results in a node error 545.
  2. Use the command-line interface (CLI) command satask chbootdrive -replacecanister to update the data stored on the boot drives to complete the procedure.

What to do next

The status of a drive slot is uninitialized only if the IBM Storage Virtualize software might not automatically initialize the FRU drive. This status can happen if the node serial number on the other boot drive does not match the node serial number on the system board. If the node serial number on the other boot drive matches the MT-M S/N label on the front that is left of the node, you can safely rescue the uninitialized boot drive from the other boot drive. Use the service assistant GUI or the satask rescuenode command to rescue the drive.