ITEM: AM6670L

Customer has ghost disks on an HACMP cluster.

Question:

  Customer is having problems with phantom drives. He removes them
  but when he power cycles the system, they come back.

Response:

Explained to the customer that the ghost drives are a normal
occurance for an HACMP cluster and that they should not be removed
except when an importvg is going to be performed.  

The ghost disks apear when the system is booted and another system has 
the shared volume group varied on.  Here is why this happens,  the 
config manager tries to perform an lqueryvg on the drives that are 
defined and were previously available.  Of course, the lqueryvg fails
since the drives are already varied on to another system.  This causes 
the config manager to mark the drives as defined instead of available. 
Then config manager sees drives out on the scsi bus so he creates new 
hdisks for these drives.  Of course, these new drives are the same as 
the defined drives but config manager is unable to read the VGDA to 
determine this.  Therefore,  the newly added drives are made available 
but do not have the PVID and are not associated with any volume group 
(since the config manager thinks these are new drives).  These ghost
disks are handled correctly by HACMP when a failover occurs.

The customer had removed the defined drives thinking that they were
ghost disks.  Now HACMP is unable to failover to the secondary system.

You can reference chapter 7 in the "Administration Guide" for hacmp 3.1,
or you can follow the steps bellow.

To correct this problem varyoff the volume group on the primary,  
export the volume group from both systems,  remove the hdisks associated
with the shared volume group, run config manager, and reimport the 
volume group.

 The following commands need to be executed on both systems.

 1. smit clstop              
_______________________________________________________________________
                             Stop Cluster Services

 Type or select values in entry fields.
 Press Enter AFTER making all desired changes.

                                                        [Entry Fields]
 * Stop now, on system restart or both                 now                     +

   BROADCAST cluster shutdown?                         true                    +
 * Shutdown mode                                       graceful                +
          (graceful, graceful with takeover, forced)
_______________________________________________________________________

 2. varyoffvg \

 3. exportvg \

 4. rmdev -d -l \

 5. cfgmgr

 6. importvg -y \ -V \ \

 7. Set auto varyon to "no".
    smit chvg
 ______________________________________________________________________
                          Change a Volume Group

 Type or select values in entry fields.
 Press Enter AFTER making all desired changes.

                                                        [Entry Fields]
 * VOLUME GROUP name                                   testvg
 * Activate volume group AUTOMATICALLY                 no                      +
     at system restart?
 * A QUORUM of disks required to keep the volume       yes                     +
    group on-line ?

 _____________________________________________________________________




Support Line: Customer has ghost disks on an HACMP cluster. ITEM: AM6670L

Dated: August 1995 Category: N/A

This HTML file was generated 99/06/24~13:30:26

Comments or suggestions?
Contact us