Databases Reference
In-Depth Information
6. Disable the Server Affinity Enabled checkbox and click on the Save button.
7. Click on the Activate Changes button to finish.
How it works...
To illustrate the flow of the process, suppose an application request is being processed in
the first Managed Server,
PROD_Server01
. The application will use the connection factory
JMSAppCF
to publish a JMS message to the
JMSAppQueue
distributed queue. Thanks to the
server affinity option being enabled, the JMS message will be published to the local queue
member of the
JMSAppQueue
distributed queue that is running in the same Managed Server,
PROD_Server01
.
In order to prioritize the delivery of a JMS message to a queue member of the distributed queue,
the balancing algorithm verifies some characteristics such as if the queue member is local to
the Managed Server, if it has a consumer, or if it has a persistent store. The disabled server
affinity removes the influence of being a local queue. So, with the server affinity disabled, the
same request running in
PROD_Server01
will balance the load and publish the JMS messages
to all queue members of the distributed queue,
JMSAppQueue
, running in all Managed Servers,
PROD_Server01
,
PROD_Server02
,
PROD_Server03
, and
PROD_Server04
.
Disabling the server affinity is useful for distributing the JMS messages to
the members of the distributed queue when the application requests are
not yet balanced across the Managed Servers. An example is when the
application requests are not load-balanced at the web tier or when a single
request publishes a large amount of messages.
Analyze each application case and use the option that better suits an even
distribution of the load in production.
Search WWH ::
Custom Search