IT41342: ORACLE RESTORE JOB UNMOUNTS INCORRECT VOLUME DURING CLEANUP

APAR status

• Closed as program error.

Error description

• When IBM Spectrum Copy Data Management performs an Oracle
  restore job, it mounts temporary clone volumes to the
  Oracle server. When the restore job ends, the temporary
  volumes are detached as part of the cleanup process.
  
  The devices are detached based on their device names
  (e.g. /dev/sdb) as observed when the restore was first
  performed. If the Oracle server is restarted while the
  restore job was active, then device names in the Linux
  kernel can change. When the restore job is ended later,
  it can mistakenly detach another legitimate device which
  now has the same name previously used by the temporary
  clone volume.
  
  Detaching the incorrect volume can cause other failure of
  other services which rely on that device. In the worst
  case, this can cause a crash of the operating system.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * IBM Spectrum Copy Data Management levels 2.2.10, 2.2.11,     *
  * 2.2.12, 2.2.13, 2.2.14, 2.2.15, and 2.2.16                   *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * See Error Description                                        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Apply fixing level when available. This problem is currently *
  * projected to be fixed in IBM Spectrum Copy Data Management   *
  * level 2.2.16.1 and 2.2.17. Note that this is subject to      *
  * change at the discretion of IBM.                             *
  ****************************************************************
  

Problem conclusion

• Code fixes have been implemented to change the way volumes are
  unmounted during restore cleanup. Instead of saving device names
  during restores, the device UUIDs are stored. During cleanup,
  the UUIDs are used to find the device name and unmount the
  volume based on that. This ensures the correct device is
  unmounted during cleanup even if its name has changed at some
  point between restore and cleanup.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Modules/Macros

• Apps     Oracle
  

Fix information

• Fixed component name

  SPECTRUM COPY D

• Fixed component ID

  5737B3400

Applicable component levels