JR35363: JOBS WITH SEQUENTIAL FILE STAGES THAT INPUT LARGE COLUMNS OF MULTI-BYTE CHARACTERS MAY FAIL INTERMITTENTLY

Download Fix Pack 2 for InfoSphere Information Server Version 8.1

APAR status

• Closed as program error.

Error description

• A job which contains a sequential file stage may intermittently
  fail, and cause a core dump, if one of that stage's input
  columns contains multi-byte characters and has a length greater
  than 128 bytes.
  

Local fix

• Enable 'Allow per-column mapping' in the sequential file stage.
  

Problem summary

• ****************************************************************
  USERS AFFECTED:
  DataStage 7.5.3 users that have NLS enabled and import
  multi-byte sequential data.
  ****************************************************************
  PROBLEM DESCRIPTION:
  A job reading delimited data in a sequential file stage may fail
  and cause a core dump if:
  (a) NLS is enabled, and
  (b) a column's length is greater than 128 bytes, and
  (c) the input data for that column contains multi-byte
  characters.
  ****************************************************************
  RECOMMENDATION:
  Apply the patch for JR35363.
  ****************************************************************
  

Problem conclusion

• The sequential file stage implementation was writing beyond a
  column's allocated buffer during multi-byte processing.  This
  was corrected by adding additional, multi-byte specific,
  checks that can trigger buffer reallocation.
  

Temporary fix

• Enable 'Allow per-column mapping' in the sequential file stage.
  

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

  WIS DATASTAGE

• Fixed component ID

  5724Q36DS

Applicable component levels

• R753 PSY

     UP