PM63898: OpenJPA's implementation of the allocationSize in a @SequenceGenerator is incorrectly implemented.

APAR status

• Closed as program error.

Error description

• IBM customers requested to backport JIRA OPENJPA-1376 defect to
  WebSphere Application Server V6.1 (OpenJPA 1.0.x).
  https://issues.apache.org/jira/browse/OPENJPA-1376
  

Local fix

• NA
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All users of IBM WebSphere Application      *
  *                  Server V6.1.0 Feature Pack for Enterprise   *
  *                  JavaBeans 3.0 who make use of the           *
  *                  'allocationSize' when using the             *
  *                  @SequenceGenerator annotation.              *
  ****************************************************************
  * PROBLEM DESCRIPTION: OpenJPA's implementation of the         *
  *                      allocationSize in a @SequenceGenerator  *
  *                      is incorrectly implemented.             *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  The Javadoc for the 'allocationSize' property of the
  @SequenceGenerator annotation defines the 'allocationSize' as
  follows: "The amount to increment by when allocating sequence
  numbers from the sequence."  OpenJPA is using this value as a
  caching value.  That is, when generating a native sequence, an
  SQL like the following my be generated:
  CREATE SEQUENCE entityE2_seq_gen START WITH 1 INCREMENT BY 50
  CACHE 50
  In this SQL statement, OpenJPA is using a 'CACHE 50'
  statement, where the value 50 is the allocationSize.  By using
  caching, OpenJPA is requesting a sequence value from the
  database each time it needs a new sequence.  The database is
  in turning caching sequence values.  This results in extra
  calls to the database.  A more efficient approach would be for
  OpenJPA to cache the sequence value in memory, rather than
  relying on the database, and thus honoring the true meaning of
  'allocationSize'.  Allocating values in memory allows the
  OpenJPA runtime to avoid accessing the database for every
  sequence request.  That is, when 'allocationSize' is set to
  50, OpenJPA will set aside the next 50 numbers each time it
  accesses the sequence, which in turn means it only has to make
  a database trip to get new sequence numbers once every 50
  sequence number requests.  In this case, a sequence similar to
  the following should be generated:
  CREATE SEQUENCE entityE2_seq_gen START WITH 1 INCREMENT BY 50
  

Problem conclusion

• To enable the code of this fix, please use the
  following JVM system property:
  
  -Dopenjpa.jdbc.DBDictionary=useNativeSequenceCache=false
  
  Or add the following to your persistence.xml file:
  
  <property name="openjpa.jdbc.DBDictionary"
  value="useNativeSequenceCache=false"/>
  
  With this fix, and the above system property, OpenJPA will use
  the 'allocationSize' as intended (i.e. as the increment when
  allocation sequence numbers).
  
  The fix for this APAR is currently targeted for inclusion
  in Enterprise JavaBeans 3.0 Feature Pack Fix Pack 45
  (6.1.0.45) of WebSphere Application Server version 6.1.
  
  Please refer to the recommended updates page for delivery
  information:
  http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980
  

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

  WAS EJB3 FEATUR

• Fixed component ID

  5724J0851

Applicable component levels

• R610 PSY

     UP