IT23793: FDC generated when calling "com.ibm.mq.MQDestination.get(MQMessage,MQGetMessageOptions)"

APAR status

• Closed as program error.

Error description

• A message containing an MQHRF2 header and a message payload that
  has a format of MQSTR is put onto a queue.  The coded character
  set identifier (CCSID) field in the MQHRF2 header is set to
  1208, which indicates that the bytes contained within the
  message payload represent a string encoded using CCSID 1208.
  The bytes contained within the message payload contained the
  byte sequence:
  
    0xE2 0x80 0x99
  
  which is the valid UTF-8 character "RIGHT SINGLE QUOTATION MARK"
  (U+2019).
  
  The MQMD has defined within it a CCSID value of 819, meaning
  that the following header (the RFH2) used character data encoded
  in CCSID 819, meaning the following structure:
  
    MQMD.ccsid = 819
    RFH2.ccsid = 1208
    body = encoded in 1208
  
  
  When an IBM MQ classes for Java application calls the method:
  
    com.ibm.mq.MQDestination.get(MQMessage, MQGetMessageOptions)
  
  to get this message from a queue,  the method call completes
  without throwing an exception, however an FFDC containing the
  following information was also generated:
  
  *************************************************************
  Source Class     :- com.ibm.mq.MQMessage
  Source Method    :- readPropertiesRfh2()
  ProbeID          :- 1
  
                                        Data
                                        ----
  
  |   Description  :-  An unexpected error occurred while parsing
  the message properties RFH2 header
  |   Exception    :-  java.io.UnsupportedEncodingException:
  java.nio.charset.UnmappableCharacterException: Input length = 1
  *************************************************************
  
  When the method call completed, the "com.ibm.mq.MQMessage"
  object passed into the get(...) call was populated with message
  data, but the body of the message was corrupt.
  

Local fix

• Alter the message putting application such that the MQMD.ccsid
  value used in the message matches the CCSID which the body of
  the message was encoded with.
  

Problem summary

• ****************************************************************
  USERS AFFECTED:
  This issue affects users of the IBM MQ classes for Java who have
  applications that call one of the following methods:
  
  - MQDestination.get( MQMessage )
  - MQDestination.get( MQMessage, MQGetMessageOptions )
  - MQDestination.get( MQMessage, MQGetMessageOptions, int )
  
  to get messages that contain:
  
  - An MQHRF2 header.
  - And a message payload that has the format MQSTR.
  
  from either a queue or a topic.
  
  The issue affects messages of this form which have a specified
  MQMD.ccsid value which is different to the encoding CCSID of the
  message body.
  
  
  Platforms affected:
  MultiPlatform
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  When an IBM MQ classes for Java application calls one of the
  following methods
  
  - MQDestination.get(MQMessage)
  - MQDestination.getMQMessage, MQGetMessageOptions)
  - MQDestination.get(MQMessage, MQGetMessageOptions, int)
  
  to get a message containing:
  
  - An MQMD.
  - An MQHRF2 header.
  - And a message payload that has the format MQSTR.
  
  from either a queue or a topic, the IBM MQ classes for Java will
  issue an MQGET API call to get a message from the specified
  destination.
  
  If a message is returned, the IBM MQ classes for Java will:
  
  - Store the MQHRF2 header as a byte array.
  - Decode the message payload using the CCSID field within the
  MQHRF2 header and store it as a "java.lang.String" object.
  
  The IBM MQ classes for Java will then perform some processing of
  the MQHRF2 header, removing any message properties if required.
  
  After this processing has been performed, the IBM MQ classes for
  Java re-encoded the MQHRF2 header using the CCSID specified in
  the CCSID field within the MQMD of the message.  However, the
  IBM MQ classes for Java also attempted to re-encode the message
  payload as a byte sequence using this same CCSID, rather than
  the CCSID originally contained with the MQHRF2 header of the
  message.
  
  If the message body contained a byte sequence that:
  
  - Could be successfully decoded using the CCSID specified in the
  MQHRF2 header of the message, using a
  "java.nio.charset.CharsetDecoder"
  - Could not be re-encoded into a byte sequence using the CCSID
  as specified within the MQMD, using a
  "java.nio.charset.CharsetEncoder" without an exception being
  thrown
  
  then the IBM MQ classes for Java would generate an FFDC
  containing the following information:
  
  *************************************************************
  Source Class     :- com.ibm.mq.MQMessage
  Source Method    :- readPropertiesRfh2()
  ProbeID          :- 1
  
                                        Data
                                        ----
  
  |   Description  :-  An unexpected error occurred while parsing
  the message properties RFH2 header
  |   Exception    :-  java.io.UnsupportedEncodingException:
  java.nio.charset.UnmappableCharacterException: Input length = 1
  *************************************************************
  
  However, the method that the application was using to get the
  message completed without error and returned an
  "com.ibm.mq.MQMessage" object.  As the attempt to rewrite the
  message payload into the MQMessage object had failed, the
  MQMessage object contained corrupted data and could not be
  processed correctly by the application.
  

Problem conclusion

• Two changes have been made as a result of this APAR:
  
  Change 1:
  -------------
  The method:
  
  - MQHeaderList.write(DataOutput, int, int, boolean)
  
  used by the IBM MQ classes for Java to re-encode a message's
  headers and payload, has been updated so that, if the last
  header on the message has the Format field set to the value
  "MQSTR", then the CCSID specified in the CCSID field for that
  header is used to encode the payload into a byte sequence. This
  ensures that the same CCSID is used to:
  
  - Decode the byte sequence contained within the message payload
  into a String.
  - Re-encode the String back into a byte sequence.
  
  which will prevent any exceptions being thrown when the String
  is re-encoded.
  
  
  Change 2:
  -------------
  In addition to this, the IBM MQ classes for Java methods:
  
  - MQDestination.get(MQMessage).
  - MQDestination.getMQMessage, MQGetMessageOptions).
  - MQDestination.get(MQMessage, MQGetMessageOptions, int).
  
  have been updated to:
  
  - Throw an MQException containing reason code 2111
  ('MQRC_SOURCE_CCSID_ERROR') if the attempt to re-encode the
  message body using the CCSID specified by the CCSID field in the
  last header before the message body failed.
  
  - Throw an MQException containing reason code 2195
  ('MQRC_UNEXPECTED_ERROR') if the attempt to update the MQHRF2
  header on the message and rewrite the message body fails for any
  other reason.
  
  This will alert applications that have called one of these
  get(...) methods to the fact that there was an issue processing
  the message returned from the queue manager, and that the
  application should not attempt to process the message as it will
  be in an inconsistent state.  If the message was being consumed
  inside of a syncpoint, the application can then choose to
  rollback the syncpoint and put the message back onto the queue
  that it came from.
  
  Note that if an MQException containing reason code 2195 is
  thrown, an FFDC containing the following information:
  
  *************************************************************
  Source Class     :- com.ibm.mq.MQMessage
  Source Method    :- readPropertiesRfh2()
  ProbeID          :- 1
  
                                        Data
                                        ----
  
  |   Description  :-  An unexpected error occurred while parsing
  the message properties RFH2 header
  |   Exception    :-  <Exception details>
  *************************************************************
  
  will be generated.  The information in the FFDC will help IBM
  Support identify the root cause of the MQException.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v8.0       8.0.0.10
  v9.0 LTS   9.0.0.4
  
  The latest available maintenance can be obtained from
  'WebSphere MQ Recommended Fixes'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006037
  
  If the maintenance level is not yet available information on
  its planned availability can be found in 'WebSphere MQ
  Planned Maintenance Release Dates'
  http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006309
  ---------------------------------------------------------------
  

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

  IBM MQ BASE MP

• Fixed component ID

  5724H7251

Applicable component levels