PK60337: WMSG1603E DURING START-UP OF AN APPLICATION SERVER AND POSSIBLY FOLLOWED BY A NULLPOINTEREXCEPTION

Java SDK 1.5 SR8 Cumulative Fix for WebSphere Application Server
Java SDK 1.5 SR8 Cumulative Fix for WebSphere Application Server
Java SDK 1.5 SR10 Cumulative Fix for WebSphere Application Server
6.1.0.31: Java SDK 1.5 SR11 FP1 Cumulative Fix for WebSphere Application Server
6.1.0.33: Java SDK 1.5 SR12 FP1 Cumulative Fix for WebSphere
6.1.0.29: Java SDK 1.5 SR11 Cumulative Fix for WebSphere Application Server
6.1.0.35: Java SDK 1.5 SR12 FP2 Cumulative Fix for WebSphere
6.1.0.37: Java SDK 1.5 SR12 FP3 Cumulative Fix for WebSphere
6.1.0.39: Java SDK 1.5 SR12 FP4 Cumulative Fix for WebSphere Application Server
6.1.0.41: Java SDK 1.5 SR12 FP5 Cumulative Fix for WebSphere Application Server
6.1.0.43: Java SDK 1.5 SR13 Cumulative Fix for WebSphere Application Server
6.1.0.45: Java SDK 1.5 SR14 Cumulative Fix for WebSphere Application Server
6.1.0.47: WebSphere Application Server V6.1 Fix Pack 47
6.1.0.47: Java SDK 1.5 SR16 Cumulative Fix for WebSphere Application Server

APAR status

• Closed as program error.

Error description

• An exception occurred within the JMSRegistration class with
  message WMSG1603E during start-up of an application server,
  this was followed by a NullPointerException from
  MDBCppUtilitiesInterfaceImpl when attempting to access the
  WebSphere MQ messaging provider resources on z/OS.
  
  Attempts to resolve the issue by restarting application servers
  
  on the node failed.
  The node hosted multiple application servers.
  

Local fix

• - Stop ALL application servers on the node
  - Start ONE application server on the node.
  - Check the application server JVM logs for WMSG1603E messages
  - If WMSG1603E occurs, stop the application server again.
  - Start ONE application server a second time.
  - Check the application server JVM logs for WMSG1603E messages
  - If no WMSG1603E is seen, it is now safe to start
    ALL application servers on the node.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:  Users of WebSphere Application Server V6.1  *
  *                  using WebSphere MQ as a messaging provider. *
  ****************************************************************
  * PROBLEM DESCRIPTION: A WMSG1603E message occurs during       *
  *                      application server startup              *
  *                      after making a change to the            *
  *                      MQ_INSTALL_ROOT or updating the         *
  *                      WebSphere MQ installation               *
  *                      pointed to by MQ_INSTALL_ROOT.          *
  *                                                              *
  *                      After this occurs, attempts to          *
  *                      access connection factories via         *
  *                      JNDI and connect to the WebSphere       *
  *                      MQ messaging provider fail.             *
  *                                                              *
  *                      An additional restart of the            *
  *                      application server is required          *
  *                      in order to resolve the issue.          *
  *                      However, when multiple application      *
  *                      servers exist on a node, restarting     *
  *                      the application servers after the       *
  *                      WMSG1603E message occurs may not        *
  *                      resolve the issue - depending on        *
  *                      the order in which the application      *
  *                      servers are restarted.                  *
  ****************************************************************
  * RECOMMENDATION:  Ensure that the MQ_INSTALL_ROOT environment *
  *                  variable is configured ONLY at Node scope.  *
  *                  Configuration of this environment variable  *
  *                  at Server scope can cause problems (which   *
  *                  are not resolved by the code changes in this*
  *                  APAR).                                      *
  ****************************************************************
  When using WebSphere MQ as a messaging provider, a registration
  process occurs within WebSphere Application Server V6.1 to
  build internal knowledge (an OSGi bundle) describing the
  WebSphere MQ installation pointed to by the MQ_INSTALL_ROOT
  environment variable.
  
  If changes are made to the WebSphere MQ installation, such
  as updating the version or changing the MQ_INSTALL_ROOT to point
  at a new location, then the registration process must occur
  again for this this internal knowledge to be rebuilt.
  
  The following WMSG1603E message may occur because the
  application server attempts to perform this re-registration
  during startup of the application server, after the old
  registration information has already been loaded, and hence
  it is already in-use and cannot be re-registered:
  .
  WMSG1603E: An internal error occurred. It was not possible to
  register the WebSphere MQ JMS client with the application server
  due to exception org.osgi.framework.BundleException: The bundle
  could not be resolved Reason: Another singleton version
  selected: com.ibm.ws.runtime.mqjms_6.0.1.1
  .
  Restarting the application server after the WMSG1603E should
  allow the registration process to complete, and connections
  to the WebSphere MQ messaging provider to be successful.
  .
  However, on a node which hosts multiple application servers
  the order in which the application servers are restarted may
  affect the registration process and cause the WMSG1603E
  message to re-occur. This is because each application server
  saves the registration information (the OSGi bundle) to a
  cache when it shuts down, so the shut down of one
  application server cancels out the re-registration performed
  by another application server.
  
  Another possibility for continued occurrence of WMSG1603E
  messages is that the MQ_INSTALL_ROOT environment variable
  has been specified at Server scope (rather than Node scope).
  This causes a problem as all application servers on a node
  share some state (the OSGi cache) and if the MQ_INSTALL_ROOT
  differs between application servers on a node they may
  each overwrite state regarding the MQ_INSTALL_ROOT
  environment variable saved by another application server
  on the node. This results in WMSG1603E messages.
  

Problem conclusion

• This APAR makes two changes, which improve the behaviour of
  WebSphere Application Server V6.1 when changes are made
  to the MQ_INSTALL_ROOT environment variable or the
  WebSphere MQ installation.
  .
  It is still important to ensure the MQ_INSTALL_ROOT
  environment variable is only configured at Node scope after
  applying this APAR. On a Node with MQ_INSTALL_ROOT
  customized at Server level, WMSG1612E messages may occur
  after this APAR is applied (described below) instead of
  WMSG1603E messages, and it will still be necessary to
  perform multiple restarts of application servers in
  order to use the WebSphere MQ JMS resources.
  .
  The changes made in this APAR to improve the behaviour
  of WebSphere Application Server following changes
  to a WebSphere MQ JMS client installation are as follows:
  .
  1) The detection of a change to the WebSphere MQ
  installation now occurs at application server shutdown
  (before the cache information is written) as well as
  application server startup. This has the following benefits:
  * During shutdown of the application server it is possible
  for the re-registration process to complete successfully.
  This means that if changes occur while an application
  server is running, only a single restart of the
  application server is required, rather than two restarts.
  * If multiple application servers on a node exist, then
  they can be restarted in any order without affecting
  the success of the re-registration process.
  .
  2) There are still some circumstances in which an additional
  restart of an application server is required.
  These cases are unavoidable. The most likely circumstance
  is when a change is made to the WebSphere MQ installation
  while the application server is stopped, and then the
  application server is started.
  .
  For this reason, the following messages have been added which
  are outputted when an additional application server restart
  is required, and give the reason the restart is required:
  .
  Error Message "WMSG1612E":
  If the MQ_INSTALL_ROOT environment variable is altered
  while an application server is stopped, the following
  message will be outputted when the application server
  is next started:
  .
  WMSG1612E: The MQ_INSTALL_ROOT has been updated from {0} to
  {1}. The WebSphere MQ Messaging Provider has been disabled.
  A server restart is required to re-enable the function.
  .
  After this message occurs, any attempt to resolve a
  WebSphere MQ connection factory via JNDI will result in
  an exception with the following message:
  .
  WMSG2002E: It was not possible to lookup the specified
  WebSphere MQ administered object because the application
  server needs a restart to pick up the current WebSphere MQ
  installation.
  .
  A restart of the application server is REQUIRED when a
  WMSG1612E message occurs during startup.
  .
  Error Message "WMSG1614E":
  If significant changes to the WebSphere MQ JMS client
  pointed to by MQ_INSTALL_ROOT occur while an application
  server is stopped, then the following message will be
  outputted when the application server next starts:
  .
  WMSG1614E: The WebSphere MQ installed at {0} has been
  updated and a server restart is required for the
  WebSphere MQ Messaging Provider to work with this update.
  The WebSphere MQ Messaging Provider has been disabled.
  .
  An example of a significant change to the WebSphere MQ
  JMS client is a new version that adds/removes Jar files.
  .
  After this message occurs, any attempt to resolve a
  WebSphere MQ connection factory via JNDI will result in
  an exception with the following message:
  .
  WMSG2003E: It was not possible to lookup the specified
  WebSphere MQ administered object because the application
  server needs a restart to register an update to the
  WebSphere MQ installation.
  .
  A restart of the application server is REQUIRED when a
  WMSG1614E message occurs during startup.
  .
  Warning Message "WMSG1613W":
  If maintenance is applied to the WebSphere MQ installation
  while an application server is stopped, but the changes
  are not significant enough to mean a re-registration is
  a necessity, the following message is outputted during
  application server startup:
  .
  WMSG1613W: The WebSphere MQ installed at {0} has been
  updated from version {1} to version {2}. A server restart
  may be required to properly register this update.
  .
  Applications will remain able to access WebSphere MQ
  messaging provider connection factories, and create
  connections to WebSphere MQ queue managers.
  A restart of the application server is suggested when a
  WMSG1613W message occurs during startup.
  

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:

  PK60497

Fix information

• Fixed component name

  PLAT MSG COM

• Fixed component ID

  620600101

Applicable component levels

• R200 PSY

     UP