PM49734: WHEN USING MVS LOGSTREAMS FOR WEBSPHERE APP SERVER ERROR LOGS SERVANT MAY CONNECT TO LOGSTREAM MULTIPLE TIMES AND USE HIGH SQA

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• When WebSphere Application Server is configured to use the z/OS
  Logstream for recording of it's 'error log' information, it
  uses the z/OS Logger utility.
  The WebSphere Application Servant region Address Space will
  connect to the Logstream (IXGCONN) and then Write (IXGWRITE) to
  the Logstream.
  Logger limits the number of writes that any one connection can
  have in it's 'write queue'.  Each of the 'queued writes' is
  backed by SQA Storage in the z/OS System.
  It has been found that the WebSphere Servant Regions may
  'connect' multiple times.  If Logger is not able to process the
  'Writes' that are queued to it, SQA Storage will be used to hold
  the SRB';s associated with the queued messages.
  Depending on system configuration and availability of common
  storage on the system, this queue of SRB's may cause an SQA
  Storage Shortage on the System.
  
  This apar will also change the handling of some error/warning
  codes that are returned from Logger. For example when
  RC=00000004, RSN=00000407 is returned, WebSphere will log a
  message that there has been a possible loss of data, but it will
  continue writing to the logstream instead of current design of
  disconnecting from the logstream and sending the messages to
  CERR instead.
  

Local fix

• If it is determined that an SQA Storage shortage is caused by
  Logger queue of SRB's from the WebSphere Servant Region Address
  Space, then the WebSphere configuration can be changed to not
  use Logstream for it's 'error log'
  The WebSphere variable that is set to point to a Logstream is:
      ras_log_logstreamName
  If this is set to point to a Logstream dataset name, you can
  simply remove this variable, and re-start the WebSphere Server
  to pick up the change.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All users of IBM WebSphere Application      *
  *                  Server V7.0                                 *
  ****************************************************************
  * PROBLEM DESCRIPTION: WebSphere Application Server for z/OS   *
  *                      makes multiple concurrent connections   *
  *                      to the errorlog logstream.              *
  *                      Also, WebSphere stops writing to the    *
  *                      errorlog logstream when it receives a   *
  *                      IXGCONN RC=00000004 and RSN=00000407.   *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  The WebSphere errorlog logstream (ras_log_logstreamName)
  implementation does not interpret various IXGWRITE and IXGCONN
  return and reason codes correctly.
  This can lead to a WebSphere address with multiple connections
  to the logstream.   This can negate the restriction of
  outstanding IXGWRITE requests.  Exceeding the limit for
  outstanding IXGWRITE requests can exhaust system resources,
  such as storage.  ABENDS878 RC=4 was observed for a case were
  the previous condition ran the system out of SQA storage.
  Another observed scenario resulted from an incorrect handling
  of the IXGCONN RC=4, RSN=407.  In this case, WebSphere stopped
  directing messages to the logstream.  Instead the messages
  were written to the default message location (typically CERR).
  

Problem conclusion

• Code has been modified to properly re-act to the IXGWRITE and
  IXGCONN return and reason codes.  WebSphere will properly
  recognize when our logstream connection has been lost and only
  then re-connect based on the IXG reason codes.
  
  APAR PM49734 is currently targeted for inclusion in Service
  Level (Fix Pack) 7.0.0.21 of WebSphere Application Server V7.0.
  
  Please refer to URL:
  //www.ibm.com/support/docview.wss?rs=404&uid=swg27006970
  for Fix Pack availability.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

  PM46270

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  WEBSPHERE FOR Z

• Fixed component ID

  5655I3500

Applicable component levels

• R700 PSY UK74996

     UP12/01/18 P 1201

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.