Hazelcast Idempotent Repository Tutorial
Apache Camel enables you to integrate the Idempotent Consumer Pattern in a really simple way. In this tutorial we will setup a cluster consisting of two OSGi containers (Apache Karaf). Inside this both nodes we will deploy two equal Camel bundles. This two bundles will receive a message from a REST interface, send them via a Hazelcast SEDA queue and finally store it into a distributed Hazelcast map. To avoid that a (equal) message will be consumed twice we're using the idempotent repository implementation based on Hazelcast. This implementation guarantees cluster wide that no message will be processed twice.
When to Use
If you use camel in a clustered environment (2 nodes or more) and you have to be sure that two equal messages never will be processed twice.
Why to use
Because you can create with very few effort a Idempotent Repository over n nodes inside your Camel application.
How to Use
We will create a camel route with a restlet consumer and a hazelcast producer. These routes will be deployed on the servers and form our cluster.
The restlet consumer will get XML strings that represent an invoice:
The Invoice id will be our Unique message ID.
You will need two servers with an OSGI runtime environment (we will use Karaf in this tutorial).
In the Karaf console add the camel features and install the needed packages:
Our routes (similar on both nodes):
Testing the routes :
Switch to your project directory and build the jar (with maven):
Copy the jar to your servers and install them in Karaf:
In the java project create a class with the following code:
While executing the project the console will output the send and received messages:
If the message hasn't been sent before, you will get a