IT28470: DELETION OF EXTENTS IS NOT REDUCING

8.1.8.000-IBM-SPOC-Linuxs390x
8.1.8.000-IBM-SPSRV-Linuxppc64le
8.1.8.000-IBM-SPCMS-WindowsX64
8.1.8.000-IBM-SPCMS-WindowsI32
8.1.8.000-IBM-SPCMS-Linuxx86_64
8.1.8.000-IBM-SPOC-WindowsX64
8.1.8.000-IBM-SPOC-Linuxx86_64
8.1.8.000-IBM-SPOC-LinuxPPC64le
8.1.8.000-IBM-SPOC-AIX
8.1.8.000-IBM-SPSRV-WindowsX64
8.1.8.000-IBM-SPSRV-Linuxx86_64
8.1.8.000-IBM-SPSRV-Linuxs390x
8.1.8.000-IBM-SPSRV-AIX
IBM Spectrum Protect Server V8.1.6.X interim fix downloads
IBM Spectrum Protect Server V8.1 Fix Pack 8 (V8.1.8) Downloads

APAR status

• Closed as program error.

Error description

• The deletion of extents from a directory container pool can be
  inefficient and at times may appear not to remove extents at
  all.
  
  The QUERY EXTENTUPDATES command provides the number of extents
  which are queued for deletion by the deletion engine thread.
  Example output  :
  
  
  Number of Extents Not Referenced
  
  ---------------------------------
  323,656,906
  
  
  Number of Extents Eligible for
  Deletion
  ------------------------------
  320,149,903
  
  
  The number of extents eligible for deletion should reduce over
  time and is the responsibility of the background deletion worker
  threads.  This then frees the space back to the container pool.
  If the number of extents eligible for removal is not reducing or
  even seems to grow, this issue is being encountered.
  
  The cause of the issue, is that the same container can be added
  to the queue several times to spread the work of extents to the
  deletion worker threads. The worker threads can spend a lot
  of time examining the same container for extents which may now
  no longer exist if other threads have removed extents and the
  container itself.
  
  
  Customer/L2 Diagnostics (If Applicable)
  
  
  IBM Spectrum Protect Versions Affected:
  Spectrum Protect Server versions 7.1.x and 8.1.x on all
  platforms.
  
  Initial Impact:
  Medium
  
  Additional Keywords:
  TSM, TS001890294, ANR0522W, space, freed, directory
  

Local fix

• Recycle the server instance, which rebuilds the list of extents
  to be removed and set the following option :
  
  setopt ChunkDeletionDelay 300
  
  The above option controls how often containers are examined for
  eligible chunks and placed on the queue. The default is 30
  seconds, so changing this to 300 seconds reduces the chance of
  the same container being placed on the queue
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * All IBM Spectrum Protect server users.                       *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * All IBM Spectrum Protect server users.                       *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Apply fixing level when available. This problem is currently *
  * projected to be fixed in levels 8.1.6.200, 8.1.7.100 and     *
  * 8.1.8. Note that this is subject to change at the discretion *
  * of IBM.                                                      *
  ****************************************************************
  

Problem conclusion

• This problem was fixed.
  Affected platforms for reported release:  AIX, Linux, and
  Windows.
  Platforms fixed:  AIX, Linux, and Windows.
  

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:

Fix information

• Fixed component name

  TSM SERVER

• Fixed component ID

  5698ISMSV

Applicable component levels

• R81A PSY

     UP

• R81L PSY

     UP

• R81W PSY

     UP