IV93320: CRASH OR DATA LOSS AFTER REMOVING LARGE J2 FILES ON AIX 7.2.1 APPLIES TO AIX 7200-01 17/09/29 PTF PECHANGE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• ****************************************************************
  * USERS AFFECTED:
  * Systems running the AIX 7200-01 Technology Level
  * with bos.mp64 between and including the 7.2.1.0 and 7.2.1.3
  * levels.
  ****************************************************************
  * ERROR DESCRIPTION:
  * After removing large files in jfs2, on AIX 7200-01, the
  * system may crash when an inode is re-used before garbage
  * collection is completed.
  *
  * The stack trace will be similar to:
  * Ý003E4728¨iUnlinkZombie+0000A8 (...)
  * Ý003E3B68¨zombieGarbageXT+000048 (..)
  * Ý003E38D0¨zombieGarbageTrunc+0000D0 (...)
  * Ý002EA670¨procentry+000010 (...)
  *
  * There is also potential data loss if this process does
  * not assert and continues garbage collection of the
  * re-used inode.
  *
  * With the default system settings, 'large files' may be any
  * files larger than around 15.8 MB.  Removing files larger
  * than this utilizes the background garbage collection and
  * opens up the possibility of encountering this bug.
    **************************************************************
  * RECOMMENDATION:
  * Install APAR IV93320.
  * Prior to fix availability, an interim fix is available from
  * either
  * ftp://aix.software.ibm.com/aix/ifixes/iv93320/
  * https://aix.software.ibm.com/aix/ifixes/iv93320/
  * The ifixes can be installed using AIX Live Update.
  * The ifixes can also be installed using emgr's Concurrent
  * Update.
  * If neither of these options are used, installation of these
  * fixes requires a reboot.
  ****************************************************************
  

Local fix

Problem summary

• A system crash with the stack:
   003E4728 iUnlinkZombie+0000A8 (...)
   003E3B68 zombieGarbageXT+000048 (..)
   003E38D0 zombieGarbageTrunc+0000D0 (...)
   002EA670 procentry+000010 (...)
  
  Note that if data corruption occurs, this may manifest itself
  in unpredictable ways.
  

Problem conclusion

• Disable the new background unlink processing for large files.
  

Temporary fix

•   *********
    * HIPER *
    *********
  

Comments

• 7200-01 - use AIX APAR IV93320
  

APAR Information

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

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

  IV94792

Fix information

• Fixed component name

  AIX V7.2

• Fixed component ID

  5765CD200

Applicable component levels

• R720 PSY U873438

     UP17/09/21 I 1000 Ž

PTF to Fileset Mapping

• U873438 bos.mp64 7.2.1.4