IZ16437: MSG2 ABEND ON OCCASION WHEN LENGTH OF INPUT TO LARGE

APAR status

• Closed as program error.

Error description

• * Environment:
  *   Windows for this PMR but should not be platform specific
  *
  * Problem Description:
  *   If the system command is a message, and the message is too
  large,
  *   the TEMS may abend.  It is difficult to determine the
  message size
  *   due to substitution parameters.
  *
  *
  * Detailed Recreation Procedure:
  *   Here is a sample situation which caused the TEMS to abend:
  
  <TABLE>
  <ROW>
  <SITNAME>
  H3C_UNIX_NETWORK_STATUS_G
  </SITNAME>
  <TEXT>
  <![CDATA[H3C_UNIX_NETWORK_STATUS_G, Interface Status != UP)]]>
  </TEXT>
  <AFFINITIES>
  00f2000000000000000000000000000000########F
  </AFFINITIES>
  <PDT>
  <![CDATA[*IF *SCAN Network.Interface_IP_Address *EQ . *AND
  *VALUE
  Network.Interface_Status *NE UP]]>
  </PDT>
  <REEV_DAYS>
  0
  </REEV_DAYS>
  <REEV_TIME>
  000500
  </REEV_TIME>
  <AUTOSTART>
  *YES
  </AUTOSTART>
  <ADVISE>
  <![CDATA[H3C_UNIX_NETWORK_STATUS_G, Interface Status != UP)]]>
  </ADVISE>
  <CMD>
  <![CDATA[blat d:\blat\filename.txt -to public-monitor@h3c.com
  -subject
  "!!! Tivoli Critical: H3C_UNIX_NETWORK_STATUS_G now
  &{Network.Timestamp}  is  true at  &{Network.System_Name} "
  -body "!!!
  Tivoli Critical: Situation H3C_UNIX_NETWORK_STATUS_G now
  &{Network.Timestamp}  is true at  &{Network.System_Name} . In
  detail,
  &{Network.Interface_IP_Address}  is  &{Network.Interface_Status}
  , now
  is  critical status(value!=UP). For detail, please login to TEPS
  and
  check the Network information.]]>
  </CMD>
  <AUTOSOPT>
  NNN
  </AUTOSOPT>
  <DISTRIBUTION>
  *ALL_UNIX
  </DISTRIBUTION>
  <ALERTLIST>
  </ALERTLIST>
  <HUB>
  </HUB>
  <QIBSCOPE>
  E
  </QIBSCOPE>
  <SENDMSGQ>
  *NONE
  </SENDMSGQ>
  <DESTNODE>
  </DESTNODE>
  <LOCFLAG>
  </LOCFLAG>
  <LSTCCSID>
  zh_CN
  </LSTCCSID>
  <LSTDATE>
  1080225152901000
  </LSTDATE>
  <LSTRELEASE>
  V100
  </LSTRELEASE>
  <LSTUSRPRF>
  SYSADMIN
  </LSTUSRPRF>
  <NOTIFYARGS>
  </NOTIFYARGS>
  <NOTIFYOPTS>
  </NOTIFYOPTS>
  <OBJECTLOCK>
  </OBJECTLOCK>
  <PRNAMES>
  </PRNAMES>
  <REFLEXOK>
  </REFLEXOK>
  <SITINFO>
  <![CDATA[ATOM=UNIXNET.FIPADDR;~]]>
  </SITINFO>
  <SOURCE>
  </SOURCE>
  </ROW>
  </TABLE>
  
  Import this situation and start it.  Or create own situation
  with long
  message length for the system command.
  
  KDSMAIN.RAS contains:
  
  Exception 0xC0000005 (ACCESS_VIOLATION)
  
   Computer: H3CMON03-NT                Windows: 5.2-SP2
   Current Time: 2008/02/25, 17:48:05    Start Time: 2008/02/25,
  12:47:57
   Current Clock: 0x47C28ED5             Start Clock: 0x47C28ECD
   Process ID: 0x00000D28               Thread ID: 0x00000B78
   Program: D:\IBM\ITM\CMS\kdsmain.exe
   Origin: KGL01P1.dll<>kgltmst+0x225E
  

Local fix

• Limit message size (no larger than 510 bytes)
  

Problem summary

• If a situation with a large CMD field is imported, TEMS may stop
  and produce a core dump.  Typically the failure will occur when
  SITMON tries to start the situation.  You will see a situation
  started message in the kdsmain.msg file.  TEMS will continue to
  fail when it is started back up again until this situation is
  deleted.  These symptoms were found to be caused by an internal
  memory buffer overflow.
  

Problem conclusion

• The code was modified such that if the CMD field is too large,
  an error is logged in the trace files and the situation will not
  be started.  The following messages will be seen in the log
  files when this problem id dected:
  
  
   (48DCE2DE.1928-5FC:kpxcloc.cpp,1530,"KPX_CreateProxyRequest")
  Reflex command length <531> is too large, the maximum length is
  <512>
  (48DCE2DE.1929-5FC:kpxcloc.cpp,1532,"KPX_CreateProxyRequest")
  Try shortening the command field in situation <TEST_UNIX_STATUS>
  
  kdsmain.ras log messages will be seen when the CMD field is too
  large:
  09/26/08 08:52:58 KO46256    Situation definition
  TEST_UNIX_STATUS created by *ENTERPRISE.
  09/26/08 08:52:58 KO41047    Situation TEST_UNIX_STATUS
  distribution l337.tivlab.austin.ibm.com:KUX added.
  09/26/08 08:52:58 KO41039    Error in request sqlRequest.
  Status= 1102. Reason= 58.
  09/26/08 08:52:58 KO41039    Error in request TEST_UNIX_STATUS.
  Status= 1131. Reason= 1131.
  09/26/08 08:52:58 KO48156    Not able to start monitoring for
  situation TEST_UNIX_STATUS.
  
  
  The fix for this APAR is contained in the following maintenance
  packages:
  
     | interim fix | 6.1.0.7-TIV-ITM-IF0003
  

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:

  OA27238 OA27766 OA28022

Fix information

• Fixed component name

  TEMS

• Fixed component ID

  5724C04MS

Applicable component levels

• R610 PSY

     UP