PM80336: In environments with co-existing WXS JVMs, upon termination of one of these JVMs, the clusterstore may stop retrieving

APAR status

• Closed as program error.

Error description

• When two servers share a common JVM, and one of the servers is
  terminated, either manually or do to an external error, the
  ClusterStore will stop updating its view of the route table.
  Calls directly to the ClusterStore would result in
  empty route tables, regardless of the current state of the
  environment. Remote transactions may also fail due to
  TargetNotAvailable exceptions caused by these empty route
  tables.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:  Users that have client Java virtual         *
  *                  machines (JVM) that are                     *
  *                  attempting                                  *
  *                  to                                          *
  *                  read                                        *
  *                  routing                                     *
  *                  information wih the ClusterStore class.     *
  ****************************************************************
  * PROBLEM DESCRIPTION: In environments with co-existing        *
  *                      WebSphere eXtreme Scale                 *
  *                      JVMs,  the clusterstore might stop      *
  *                      retrieving routes upon termination of   *
  *                      one of                                  *
  *                      these JVMs.                             *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  When two servers share a common JVM, and one of the servers is
  stopped, either manually or due to an external error, the
  ClusterStore stops updating its view of the route table.
  Calls directly to the ClusterStore result in empty route
  tables, regardless of the current state of the environment.
  Remote transactions might also fail due to TargetNotAvailable
  exceptions that are caused by these empty route tables.
  

Problem conclusion

• With the logic corrected, the ClusterStore continue to
  update its view of the route table after attempts to retrieve
  the information have been made.
  

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

  WS EXTREME SCAL

• Fixed component ID

  5724X6702

Applicable component levels

• R850 PSY

     UP

• R860 PSY

     UP