LibraryToggle FramesPrintFeedback

Available as of Camel 2.7

The hazelcast: component allows you to work with the Hazelcast distributed data grid / cache. Hazelcast is a in memory data grid, entirely written in Java (single jar). It offers a great palette of different data stores like map, multi map (same key, n values), queue, list and atomic number. The main reason to use Hazelcast is its simple cluster support. If you have enabled multicast on your network you can run a cluster with hundred nodes with no extra configuration. Hazelcast can simply configured to add additional features like n copies between nodes (default is 1), cache persistence, network configuration (if needed), near cache, enviction and so on. For more information consult the Hazelcast documentation on http://www.hazelcast.com/documentation.jsp .

Maven users will need to add the following dependency to their pom.xml for this component:

<dependency>
    <groupId>org.apache.camel</groupId>
    <artifactId>camel-hazelcast</artifactId>
    <version>x.x.x</version>
    <!-- use the same version as your Camel core version -->
</dependency>

hazelcast:[ map | multimap | queue | seda | set | atomicvalue | instance]:cachename[?options]

[Warning]Warning

You have to use the second prefix to define which type of data store you want to use.

Hazelcast provides event listeners on their data grid. If you want to be notified if a cache will be manipulated, you can use the map consumer. There're 4 events: put, update, delete and envict. The event type will be stored in the "hazelcast.listener.action" header variable. The map consumer provides some additional information inside these variables:

Header Variables inside the response message:

Name Type Description
hazelcast.listener.time Long time of the event in millis
hazelcast.listener.type String the map consumer sets here "cachelistener"
hazelcast.listener.action String type of event - here added, updated, envicted and removed
hazelcast.objectId String the oid of the object
hazelcast.cache.name String the name of the cache - e.g. "foo"
hazelcast.cache.type String the type of the cache - here map

The object value will be stored within put and update actions inside the message body.

Here's a sample:

fromF("hazelcast:%sfoo", HazelcastConstants.MAP_PREFIX)
.log("object...")
.choice()
    .when(header(HazelcastConstants.LISTENER_ACTION).isEqualTo(HazelcastConstants.ADDED))
         .log("...added")
         .to("mock:added")
    .when(header(HazelcastConstants.LISTENER_ACTION).isEqualTo(HazelcastConstants.ENVICTED))
         .log("...envicted")
         .to("mock:envicted")
    .when(header(HazelcastConstants.LISTENER_ACTION).isEqualTo(HazelcastConstants.UPDATED))
         .log("...updated")
         .to("mock:updated")
    .when(header(HazelcastConstants.LISTENER_ACTION).isEqualTo(HazelcastConstants.REMOVED))
         .log("...removed")
         .to("mock:removed")
    .otherwise()
         .log("fail!");

A multimap is a cache where you can store n values to one key. The multimap producer provides 4 operations (put, get, removevalue, delete).

Header Variables for the request message:

Name Type Description
hazelcast.operation.type String valid values are: put, get, removevalue, delete
hazelcast.objectId String the object id to store / find your object inside the cache
Comments powered by Disqus