PK93428: 6.1.0.5 Cumulative fix 3 for WebSphere eXtreme Scale

WebSphere® eXtreme Scale cumulative fix 3 resolves the following problems:

ERROR DESCRIPTION:
This APAR is a cumulative fix of several problems found by customers and the WebSphere® eXtreme Scale development team. Refer to the Problem Summary section for a detailed list of all of the fixes that are included in this APAR.

LOCAL FIX:

PROBLEM SUMMARY

USERS AFFECTED:
All users of WebSphere eXtreme Scale (formerly Data Grid) Version 6.1 who are using the ObjectGrid component.

PROBLEM DESCRIPTION:
This APAR is a cumulative fix of several problems. The Problem Summary section lists all of the included fixes.

RECOMMENDATION:
None

The following items are the field APARs addressed in this cumulative fix.

*PK86110: DISTRIBUTED OBJECTGRID TRANSACTIONS ARE EXTREMELY SLOW TO COMMIT WHEN USING THE ENTITYMANAGER API.
*PK86546: CLIENT THREADS APPEAR TO HANG WHEN TRYING TO COMMUNICATE TO THE PRIMARY CATALOG SERVER PROCESS.
*PK91001: MAXINACTIVEINTERVAL VALUE FOR HTTP SESSION MANAGEMENT NOT SETTING CORRECTLY ON GRID SERVERS HOSTING REPLICA SHARDS

The following items are the additional fixes included in this cumulative fix:

1. When the availability state of an ObjectGrid is set to OFFLINE, the write-behind loader is now flushed.
2. When routing to a primary shard, client request timeout is now based on transaction timeout, instead of failing fast.
3. The PlacementServiceMBean.retrieveAllServersJMXAddress method now returns addresses for ObjectGrid servers that are running in a WebSphere Application Server process.
4. Timing windows were closed that caused LockTimeoutExceptions in catalog service processes.
5. Iterations of Query results were limited to Short.MAX_VALUE. This limitation is now fixed by returning
   up to Integer.MAX_VALUE results.
6. The getForUpdate method returned the read committed value instead of getting the value again and storing the new value into the DiffMap.
7. During brownout conditions, the DuplicateKeyException exception that can occur for a newly promoted primary has been fixed.
8. The xsadmin tool now supports the "-teardown" parameter with filter options. For example, you can stop an entire zone using a zone filter.
9. Fixed the NoClassDefFoundError error that can occur with Java security enabled
   when committing EntityManager transactions.
10. Fixed the ClassCastException exception that can occur when transactions time out.
11. The StopServer script takes a list of servers to prevent shard movement, when only stopping one server.
12. Fixed the DuplicateKeyException exception that can occur with asynchronous replication when an insert transaction follows an invalidate transaction for the same key.
13. Improved stopping of servers when transitioning shards to other servers.
14. A problem with unbalanced shards was fixed in multiple zone environment.
15. Improvements to start hundreds of containers simultaneously.
    
    PROBLEM CONCLUSION:
    See the Recommended Updates page for delivery information:
    http://www.ibm.com/support/docview.wss?rs=3023&uid=swg27005709

Download the UpdateInstaller below to install this fix.

Review the readme.txt for detailed installation instructions.