HPlogo Mirrored Disk/iX User's Guide: HP 3000 MPE/iX Computer Systems > Chapter 4 Troubleshooting

Disk failures

» 

Technical documentation

Complete book in PDF
» Feedback

 » Table of Contents

 » Glossary

 » Index

The following disk failures will now be described along with their recovery procedures:

  • Drives that do not mount at system start.

  • Drives that are disabled after they have mounted.

Disk mounting failure

The system automatically mounts a mirrored volume set after volumes have been added to a mirrored volume set, upon power on of the disk drive or the boot of the system.

When a mirrored volume set is mounted, it is possible that one or more mirrored partners may be missing or not responding.

Example: If a disk does not mount

This example shows how to recover from a disk that did not mount.

  1. If LDEV 32's partner did not mount, LDEV 32 is placed in the PENDING state and a console message is displayed to alert you of this condition.

    If the partner of LDEV 32 comes online, it is recognized; otherwise, the message displays every thirty seconds.

  2. Reply to the (Y/N)? question to stop it from repeating.

       ?09:09/12/MIRRORED PARTNER MISSING FOR LDEV# 32 
    
    
    
       ?09:09/22/ACKNOWLEDGE MIRRORED PARTNER MISSING FOR LDEV# 32(Y/N)?
    
    
    
       :REPLY 22,Y
    
  3. You will not be able to access MEMBER2 due to the PENDING state of one of its disks until you do one of the following:

    1. Power on the missing partner (if the drive had previously been powered off).

    2. Issue the SUSPENDMIRRVOL command to place the PENDING disk in the SUSPEND-MIRR state and make it accessible without mirroring.

      NOTE: Issuing the SUSPENDMIRRVOL command does not stop the repeating message. Your reply stops the message.
  4. Use the DSTAT command to verify that LDEV 32's partner did not mount.

       :DSTAT
    
    
    
       LDEV-TYPE   STATUS       VOLUME  (VOLUME SET - GEN) 
    
       _ _ _ _ _ _ _ _ _ _ _    _ _ _ _ _ _ _ _ _ _ _ _ _ 
    
       30- 079370   MASTER-MD     MEMBER1     (PROD_SET-0)
    
       31- 079370   MASTER-MD     MEMBER1     (PROD_SET-0)
    
       32- 079370  PENDING-MD     MEMBER2     (PROD_SET-0)
    

    The previous screen shows that MEMBER2 is waiting (PENDING) for the mount of its partner (which is not even listed). The MEMBER2 volume remains in the PENDING state and remains unavailable until you issue the SUSPENDMIRRVOL command to override and tell the system to proceed without mirroring on that volume.

    NOTE: The SUSPENDMIRRVOL command can only be issued on a disk in the PENDING state.
  5. Use the VOLUTIL SUSPENDMIRRVOL command to access the MEMBER2 volume without mirroring.

       :VOLUTIL
    
    
    
       Mirvutil A.00.00, (C) Hewlett-Packard Co., 1990.  All Rights Reserved.
    
    
    
       volutil:SUSPENDMIRRVOL PROD_SET:MEMBER2 32
    
    
    
       *Verify:SUSPEND THE MIRROR PENDING VOLUME ON LDEV 32 [Y/N]?Y
    
    CAUTION: Care must be taken when using SUSPENDMIRRVOL to ensure that the PENDING disk is good. This command forces the system to mount and use this drive. Because of drive errors, it may not have been possible to mark the drive as bad. The drive could contain data that has not been updated. This could lead to application errors and force a reload of the volume set.
  6. After you have suspended a mirrored volume, use the SHOWSET command with the MIRROR parameter to verify that the volume can be accessed and is in the SUSPEND-MIRR state.

       volutil:SHOWSET PROD_SET MIRROR
    
    
    
       Volume Name  Vol Status   Mirr Status  Ldev  Mirr ldev
    
       - - - - - -  - - - - - -  - - - - - - - - -  - - - - - 
    
       MEMBER1      MASTER       NORMAL       30    31   
    
       MEMBER1      MASTER       NORMAL       31    30   
    
       MEMBER2      MEMBER       SUSPEND-MIRR 32    
    

    The previous screen shows that the MEMBER2 volume is available and does not have a mirrored partner.

  7. Check the disk that did not mount to see if it was powered on. Once the disk has been repaired by either physically replacing the drive or fixing the disk problem, power on the disk.

Example: Replacing a disk that did not mount

This example shows how to replace a disk that did not mount at system start.

  1. Use the DSTAT command to verify that the new volume can be initialized (SCRATCH or UNKNOWN status). If you need to scratch the volume and it does not contain any data that you want to save, use the SCRATCHVOL command.

    NOTE: The new volume must be mounted in the SCRATCH or UNKNOWN state. It does not need to have the same LDEV or I/O path as the disk that did not mount.
       volutil: :DSTAT
    
    
    
       LDEV-TYPE   STATUS       VOLUME  (VOLUME SET - GEN) 
    
       _ _ _ _ _ _ _ _ _ _ _    _ _ _ _ _ _ _ _ _ _ _ _ _ 
    
       30- 079370   MASTER-MD     MEMBER1 (PROD_SET-0)
    
       31- 079370   MASTER-MD     MEMBER1 (PROD_SET-0)
    
       32- 079370   MEMBER-MD     MEMBER2 (PROD_SET-0)
    
       33- 079370   SCRATCH
    
    NOTE: If the new volume mounts in the PENDING state, do not issue the SUSPENDMIRRVOL command on the new volume. Disk mirroring cannot work properly if both partners of a mirrored pair are placed in the SUSPEND-MIRR state.
  2. Use the REPLACEMIRRVOL command to initialize LDEV 33 as the new mirrored disk partner of LDEV 32.

       volutil: REPLACEMIRRVOL PROD_SET:MEMBER2 33
    

    The system now recognizes (mounts) the replaced volume, resumes disk mirroring, and starts the repair process.

  3. Use the SHOWSET command to verify that MEMBER2 is under repair.

       volutil: SHOWSET PROD_SET MIRROR
    
    
    
       Volume Name  Vol Status  Mirr Status  Ldev  Mirr ldev
    
       - - - - - -  - - - - - - - - - - - - - - -  - - - - - 
    
       MEMBER1      MASTER      NORMAL       30    31   
    
       MEMBER1      MASTER      NORMAL       31    30   
    
       MEMBER2      MEMBER      REPAIR-SRCE  32    33   
    
       MEMBER2      MEMBER      REPAIR-DEST  33    32   
    

    The above screen shows that LDEV 33 (REPAIR-DEST) is being repaired by LDEV 32 (REPAIR-SRCE). Programs and data residing on MEMBER2 are available while repairs are taking place.

Disk failure after mounting

The system automatically recovers from a failure of a single disk that is a partner of a mirrored pair during normal mirrored operation. Normal mirrored operation means that both partners are fully mounted and no repair operation is taking place.

A drive can fail and be marked DISABLED in the following ways:

Errors being returnedThe drive is marked as having failed (DISABLED) immediately, and the application continues to use the remaining drive in the NON-MIRROR state.
Drive not respondingThere is a slight delay (less than two minutes) while the system waits for the drive to respond. During this waiting period, processes performing I/O will be suspended. If the drive responds before the timeout, normal mirroring resumes. If the drive does not respond, the drive is marked as having failed (DISABLED), and the application continues to use the remaining drive in the NON-MIRROR state.

Example: If a disk fails after mounting

This example shows how to recover from a disk that failed after mounting.

  1. If LDEV 32 fails, a console message alerts you of this condition.

       ?09:09/12/MIRRORED VOLUME DISABLED ON LDEV# 32 
    

    The system automatically continues the application without mirroring and places the good disk in the NON-MIRROR state. The following message displays every thirty seconds, asking you to acknowledge this condition.

  2. Your reply stops the repeating message on the console. It causes no other action to take place.

       ?09:09/22/ACKNOWLEDGE MIRRORED VOLUME DISABLED ON LDEV# 32 [Y/N]?
    
       :REPLY 22,Y
    
    NOTE: The repeating message continues until a reply is given - even if the drive is replaced.
  3. Use the DSTAT command to show that LDEV 32 has failed and is no longer available.

       :DSTAT
    
    
    
       LDEV-TYPE   STATUS       VOLUME  (VOLUME SET - GEN) 
    
       _ _ _ _ _ _ _ _ _ _ _    _ _ _ _ _ _ _ _ _ _ _ _ _ 
    
       30- 079370   MASTER-MD     MEMBER1 (PROD_SET-0)
    
       31- 079370   MASTER-MD     MEMBER1 (PROD_SET-0)
    
       32- 079370   DISABLED-MD   MEMBER2 (PROD_SET-0)
    
       33- 079370   MEMBER-MD     MEMBER2 (PROD_SET-0)
    
  4. Use the SHOWSET command to confirm that LDEV 32 is disabled and that LDEV 33 is functioning in a NON-MIRROR state.

       :VOLUTIL
    
    
    
       Mirvutil A.00.00, (C) Hewlett-Packard Co., 1990.
    
       All Rights Reserved.
    
    
    
       volutil:SHOWSET PROD_SET MIRROR
    
    
    
       Volume Name  Vol Status  Mirr Status  Ldev  Mirr ldev
    
       - - - - - -  - - - - - - - - - - - - - - -  - - - - - 
    
       MEMBER1      MASTER      NORMAL       30    31   
    
       MEMBER1      MASTER      NORMAL       31    30   
    
       MEMBER2      MEMBER      DISABLED     32    33   
    
       MEMBER2      MEMBER      NON-MIRROR   33    32   
    

Example: Replacing a disabled disk

This example shows how to replace a disk in the DISABLED state.

  1. Use the DSTAT command to verify that the new volume mounts in the DISABLED state and has the same LDEV as the previous volume that was disabled.

       :VOLUTIL
    
    
    
       Mirvutil A.00.00, (C) Hewlett-Packard Co., 1990.
    
       All Rights Reserved.
    
    
    
       volutil:SHOWSET PROD_SET MIRROR
    
    
    
       Volume Name  Vol Status  Mirr Status  Ldev  Mirr ldev
    
       - - - - - -  - - - - - - - - - - - - - - -  - - - - - 
    
       MEMBER1      MASTER      NORMAL       30    31   
    
       MEMBER1      MASTER      NORMAL       31    30   
    
       MEMBER2      MEMBER      DISABLED     32    33   
    
       MEMBER2      MEMBER      NON-MIRROR   33    32   
    
    NOTE: When replacing a volume in the DISABLED state, you must use the same LDEV number and I/O path as the failed disk. Any volume mounted on that LDEV mounts in the DISABLED state and is available as the target of the REPLACEMIRRVOL command.
  2. Use the REPLACEMIRRVOL command to replace LDEV 32 and resume mirroring (after the repair).

       volutil: REPLACEMIRRVOL PROD_SET:MEMBER2 32
    

    The system now recognizes (mounts) the replaced volume, resumes disk mirroring, and starts the repair process. The replacement volume has the same characteristics specified when the disabled volume was first initialized using the NEWMIRRVOL or NEWMIRRSET commands.

    NOTE: REPLACEMIRRVOL always initiates an immediate repair (no staging), even if the maximum number of repairs is already taking place.

    If the destination drive fails or is not responding during a repair operation, it is marked DISABLED, and the source drive returns to the NON-MIRROR state.

  3. You can use the SHOWSET command to verify that MEMBER2 is under repair.

       volutil: SHOWSET PROD_SET MIRROR
    
    
    
       Volume Name  Vol Status   Mirr Status  Ldev  Mirr ldev
    
       - - - - - -  - - - - - -  - - - - - - - - -  - - - - - 
    
       MEMBER1      MASTER       NORMAL       30    31   
    
       MEMBER1      MASTER       NORMAL       31    30   
    
       MEMBER2      MEMBER       REPAIR-SRCE  32    33   
    
       MEMBER2      MEMBER       REPAIR-DEST  33    32   
    
Feedback to webmaster