X-Git-Url: https://review.fuel-infra.org/gitweb?a=blobdiff_plain;f=website%2Freference%2Fplugins%2Fconnector_activemq.md;fp=website%2Freference%2Fplugins%2Fconnector_activemq.md;h=f38bade522e33ccb1561f455d44465ffd7817673;hb=b87d2f4e68281062df1913440ca5753ae63314a9;hp=0000000000000000000000000000000000000000;hpb=ab0ea530b8ac956091f17b104ab2311336cfc250;p=packages%2Fprecise%2Fmcollective.git

diff --git a/website/reference/plugins/connector_activemq.md b/website/reference/plugins/connector_activemq.md
new file mode 100644
index 0000000..f38bade
--- /dev/null
+++ b/website/reference/plugins/connector_activemq.md
@@ -0,0 +1,135 @@
+---
+layout: default
+title: ActiveMQ Connector
+toc: false
+---
+[STOMP]: http://stomp.codehaus.org/
+
+The ActiveMQ connector uses the [STOMP] rubygem to connect to ActiveMQ servers.  It is specifically optimiszed for ActiveMQ
+and uses features in ActiveMQ 5.5.0 and later.
+
+This code will only work with version _1.1.8_ or newer of the Stomp gem.
+
+## Differences between ActiveMQ connector and Stomp Connector
+
+The ActiveMQ connector requires MCollective 2.0.0 or newer and introduce a new structure to the middleware messsages.
+
+ * Replies goes direct to clients using short lived queues
+ * Agent topics are called */topic/<collective>.<agent_name>.agent*
+ * Support for point to point messages are added by using _/queue/<collective>.nodes_ and using JMS selectors.
+
+The use of short lived queues mean that replies are now going to go back only to the person who sent the request.
+This has big impact on overall CPU usage by clients on busy networks but also optimize the traffic flow on
+networks with many brokers.
+
+Point to Point messages means each node has a unique subscription, the approach using JMS Selectors means
+internally to ActiveMQ only a single thread will be dedicated to this rather than 1 per connected node.
+
+Before using this plugin you will need to make appropriate adjustments to your ActiveMQ Access Control Lists.
+
+## Configuring ActiveMQ
+For best behavior there are a few settings you need in your _activemq.xml_
+
+### Remove unused queues
+We use uniquely named queues for replies.  As queues will live forever we need to get ActiveMQ to remove
+queues we are done with else they will just add up and grow forever.
+
+{% highlight xml %}
+<destinationPolicy>
+  <policyMap>
+    <policyEntries>
+      <policyEntry queue="*.reply.>" gcInactiveDestinations="true" inactiveTimoutBeforeGC="300000" />
+    </policyEntries>
+  </policyMap>
+</destinationPolicy>
+{% endhighlight %}
+
+The above policy will instruct ActiveMQ to remove dead queues after 5 minutes.
+
+### Optimize network usage for direct requests in a network of brokers
+If you are using a network of brokers you will need to make a big change to how that works.
+At present we tend to have 1 bi-directional connection for everything, with direct requests
+we dedicate a bi-directional connection for these queues leaving the other just for topics.
+
+{% highlight xml %}
+<networkConnectors>
+  <networkConnector
+        name="stomp1-stomp2-topics"
+        uri="static:(tcp://stomp2.xx.net:6166)"
+        userName="amq"
+        password="secret"
+        duplex="true"
+        decreaseNetworkConsumerPriority="true"
+        networkTTL="2"
+        dynamicOnly="true">
+        <excludedDestinations>
+                <queue physicalName=">" />
+        </excludedDestinations>
+  </networkConnector>
+  <networkConnector
+        name="stomp1-stomp2-queues"
+        uri="static:(tcp://stomp2.xx.net:6166)"
+        userName="amq"
+        password="secret"
+        duplex="true"
+        decreaseNetworkConsumerPriority="true"
+        networkTTL="2"
+        dynamicOnly="true"
+        conduitSubscriptions="false">
+        <excludedDestinations>
+                <topic physicalName=">" />
+        </excludedDestinations>
+  </networkConnector>
+</networkConnectors>
+{% endhighlight %}
+
+You will need to adjust the TTL for your network.  Note the queue connection has a different
+_conduitSubscriptions_ policy than the topic one, you have to create these different connections
+and set this policy for everything to work correctly.
+
+## Configuring MCollective
+
+### Common Options
+The most basic configuration method is supported in all versions of the gem:
+
+### Failover Pools
+A sample configuration can be seen below.  Note this plugin does not support the old style config of the Stomp connector.
+
+{% highlight ini %}
+connector = activemq
+plugin.activemq.pool.size = 2
+plugin.activemq.pool.1.host = stomp1
+plugin.activemq.pool.1.port = 6163
+plugin.activemq.pool.1.user = me
+plugin.activemq.pool.1.password = secret
+
+plugin.activemq.pool.2.host = stomp2
+plugin.activemq.pool.2.port = 6163
+plugin.activemq.pool.2.user = me
+plugin.activemq.pool.2.password = secret
+{% endhighlight %}
+
+This gives it 2 servers to attempt to connect to, if the first one fails it will use the second.  Usernames and passwords can be set
+with the environment variables STOMP_USER, STOMP_PASSWORD.
+
+If you do not specify a port it will default to _61613_
+
+You can also specify the following options for the Stomp gem, these are the defaults in the Stomp 1.1.6 gem:
+
+{% highlight ini %}
+plugin.activemq.initial_reconnect_delay = 0.01
+plugin.activemq.max_reconnect_delay = 30.0
+plugin.activemq.use_exponential_back_off = true
+plugin.activemq.back_off_multiplier = 2
+plugin.activemq.max_reconnect_attempts = 0
+plugin.activemq.randomize = false
+plugin.activemq.timeout = -1
+{% endhighlight %}
+
+### Message Priority
+
+ActiveMQ messages support priorities, you can pass in the needed priority header by setting:
+
+{% highlight ini %}
+plugin.activemq.priority = 4
+{% endhighlight %}