How WebLogic Server (WLS) JMS Unit Of Order Mapping Works with Uniform Distributed Destinations (Doc ID 1310795.1)

Last updated on NOVEMBER 05, 2016

Applies to:

Oracle Weblogic Server - Version: 10.3.3 and later   [Release: and later ]
Information in this document applies to any platform.

Goal

Consider this configuration: there is a cluster of two managed servers with a Uniform Distributed Queue (UDQ) targeted to the cluster: all default configurations. According to Unit of Order (UOO) design, WebLogic Server (WLS) JMS directs messages with the same Unit-of-Order, and having a distributed destination target, to the same distributed destination member. Here are test case results that describe the behavior:

  1. Send messages to the UOO queue: they go on to managed server MS1.
  2. Shut down and remove MS1 from the cluster.
  3. Send in the same messages that were associated to MS1. They are now written to the Unit Of Order queue on managed server MS2.
  4. Assign MS1 back into the cluster, in shutdown mode.
  5. Send the same messages that were associated to MS1. The messages fail because MS1 is still shutdown, but is part of the cluster, so they are still trying to go to the original managed server.
  6. Start MS1.
  7. Send in the same messages that were associated to MS1. The messages are now placed in the UOO queue on MS1.

From the observations taken from this test case, it appears that the UOO messages are permanently pinned to the same managed server. What is recommended to keep up with processing UOO messages when the destination servers are in a cluster and subject to failing over? Is this behavior expected?


Solution

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms