RabbitMQ modules

The RabbitMQ module bundle provides an alternative way of publishing content and synchronizing instances in Magnolia.

RabbitMQ is open source message broker software (a.k.a message-oriented middleware) that implements the Advanced Message Queuing Protocol (AMQP). 

The Magnolia RabbitMQ modules are suitable for: 

• Environments consisting of more than five public instances.  
• Environments that need to synchronize with other environments.

The RabbitMQ alternative:

• Gives you full control over activation and synchronization.
• Reduces the load on your author environment.
• Persists activations in queues, allowing for easier setup of continuous deployment environments.

This document guides you through the process of setting up RabbitMQ in your environment and installing and configuring the modules.

RabbitMQ bundle

The RabbitMQ bundle contains three modules:

• RabbitMQ Connector module: Integrates RabbitMQ with Magnolia.
• RabbitMQ Activation module: Allows you to synchronize and activate over RabbitMQ.
• RabbitMQ Monitoring module. Allows you to monitor behavior on update for each instance that receives activation messages.

Installing Magnolia RabbitMQ modules

The modules are interdependent. Please install all three modules. 

Maven is the easiest way to install the modules. Add the following dependencies to your bundle:

Installing RabbitMQ service

Follow the instructions at RabbitMQ to download and install the service. Instructions and packages are provided for all operating systems.

Note:

• Magnolia modules are needed to set up RabbitMQ so install the three jars that are shipped with the parent POM first on your Magnolia instances. 
• Connector module allows you to configure RabbitMQ directly from Magnolia. The module comes with a set of example consumers and producers that you can extend as necessary.
• Activation module allows you to sync and activate content to other instances. Those instances can be public or author instances.
• Monitoring module provides tools and a consumer for visualizing and monitoring the consuming process on all instances that receive activation messages.

Configuring the Rabbit MQ connector

RabbitMQ exchanges and consumers are configured in the connector module. The exchange is configured on the author or producing instance and the consumer on the public or receiving instance.

Configuring the RabbitMQ client

The module ships with an example client configuration in /modules/rabbitmq-connector/rabbitmq-client/sampleClient.

Properties:

Exchange configuration

The exchange configuration is configured on the author or producing instance.

The exchange is the kind of router that the queues connect to. The producer instance pushes messages to the exchange and the exchange then decides what to do with them.

The example exchange configuration is in /modules/rabbitmq-connector/rabbitmq-client/sampleClient/exchangeConfig.

Properties:

Testing the connection

You can test the connection in the RabbitMQ console.

When the client is enabled, the exchanges and queues should be visible in the RabbitMQ console. You can test this at http://localhost:5672, for example.  Make sure the consumer is disabled on the author instance or it will re-consume what it had just tried to publish.

You should be able to see the queue in Queues. 

You should be able to see the exchange in Exchanges. 

RabbitMQ activation commands

The RabbitMQ Activation module bootstraps preconfigured activation commands in the standard Activation module.

You can select commands/catalogs to match your purpose for using RabbitMQ and use them in your apps. You can also customize them to suit your needs. 

The commands are configured in /modules/rabbitmq-activation(activation)/commands/<catalog name>.

Catalogs:

The activate and deactivate commands defined in the above catalogs depend on your client configuration and exchange configuration. The activate and deactivate commands each contain a sub-command that contain the client name and and exchange name as properties. 

For example:

Adding commands to apps

You can use the RabbitMQ commands in your apps, for example, to sync your content with another environment after an activation.

Example: Content syncing in the Pages app in /modules/pages/commands/website.

Testing activation

To test the new commands activate content and then open the RabbitMQ console. You should see messages in the queue (fan1) . 

Configuring the activation consumer

When the messages have successfully reached the queue:

1. Start the receiving instance.
2. Configure the activation consumer on the receiving instance to consume to the correct queue (queueName property).
3. Enable the consumer.

 Make sure the sampleClient configuration is enabled on both instances.

The example consumer definition is in /modules/rabbitmq-connector/rabbitmq-client/sampleClient/consumerDefinitions.

Properties:

Testing message consumption

When the activation consumer is configured, the receiving instance should start consuming messages.

Open the RabbitMQ console to verify that the specific queue (fan1) is being emptied and that the nodes appear on the receiving instance. 

Content Syncing

While transactions are possible with RabbitMQ, they are complicated and slow down publishing. The ultimate would be to have the same activation message received by all instances at the same time, but this is technically very difficult to achieve and does not add much advantage in any event.

What is important however is that a Magnolia instance knows when a node is out of sync with another node, and that the load balancer knows which node is late. This is achieved by recording the time stamp and latest message state. 

Whenever a message has been consumed and the content activated on the instance, the instance stores the message id (which is a long that is incremented) and the date the message was activated.

Example:

This information is either sent to the load balancer or made available over a servlet. The goal is that the load balancer knows which instance is behind and by how much. In the example above the load balancer controlling tool knows that Public 1 and Public 2 are synced. It can decide to wait for Public 3 to sync, and if this does not happen, to alert the systems administrator or trigger the creation of a new instance.

SyncState and SyncStore mechanism

The Magnolia RabbitMQ implementation introduces the concepts of SyncState and SyncStore to achieve content synchronicity between public instances.

The mechanism is basically a counter that is incremented when each activation message is sent. The counter is sent in the message header. When the receiving instance gets the message, the instance increments its own activation counter. This means that two public instances which have different SyncStates are out of sync until the counters once again have the same number.

The load balancer calls the REST service of the public nodes and in this way knows which public node is most up to date and which one is out of sync.

Examples: 

The loadbalancer also knows when the public instances are getting activated. SyncStates are persisted to the JCR.

Guaranteeing synchronicity 

The Dynamically Weighted Least Connections Algorithm on your load balancer guarantees synchronicity and ensures that public instances do not get too far out of sync. See Least-Connection Algorithm based on variable weight for multimedia transmission for more.

The load balancer plays a big role in this. During a very big and long activation process, the load balancer can decide to redirect the traffic to the instances which have the highest and most equal ids. This increases traffic on these instances, which in turn slows down the activation process. Instances with initial low and unequal ids now have a higher chance of reaching the same state as the instances under load. 

The algorithm creates a kind of "auto-damping" system that slows down instances by giving them more traffic, while speeding up the activation process on the others . Having the load balancer actively take the activation mechanism into consideration stabilizes the whole public node group. 

Public Monitoring app

The RabbitMQ Monitoring module includes the Public Monitoring app and the Public monitoring REST service

In the Public Monitoring app can can follow the state of any instance that is consuming on a queue. The app will return information messages containing the status of activations, provided the ACK exchange is configured correctly, 

Configuring monitoring

Monitoring is configured on the instance responsible for monitoring, typically the author or producing instance.

Configuration involves three steps:

1. Configuring the ACK client .
2. Adding the ACK exchange name to the activation consumer definition. 
3. Configuring REST endpoints.

Configuring the ACK client

The ACK client is configured in the connector module.

The example ACK client configuration is in /modules/rabbitmq-connector/rabbitmq-client/ackClient/consumerDefinitions/ackConsumer.

Properties: 

Adding the ACK exchange name

The next step to configuring monitoring is to link the ACT client in the the activation consumer definition . 

To do this add the actExchangeName to the activation consumer definition in /modules/rabbitmq-connector/rabbitmq-client/sampleClient/consumerDefinitions/activationConsumer/ackExchangeName, for example.

Configuring REST endpoints

The final step to configuring monitoring is to configure REST endpoints and add the access rights to them. Permissions to issue REST requests are controlled by Magnolia's standard role-based security mechanism.

REST endpoints are used for getting information from, monitoring and controlling the activation status

SyncState REST service

The SyncState REST service allows you to get the current state of activation immediately (without checking the ACK return queue). 

The service is configured in the activation module in /modules/rabbitmq-activation/rest-endpoints/syncstate.

Properties:

To enable the service:

1. In the Security app add the following permissions to the rest role: 
   

   Access control lists:

   Workspace	Permission	Scope	Path
   Config	Read only	Selected and sub nodes	/rabbitmq-activation

   Web access:

   Permission	Path
   Get & Post	.rest/syncstate/state

2. Assign the rest role to a user or to anonymous if you need access without credentials. 

You can use the REST endpoint as follows:

curl http://publicaddress:publicport/.rest/syncstate/state

You should get the following (example):

{"seqNbr":2011,"stamp":1467280192928,"topoTag":""}

Public monitoring REST service

The Public monitoring REST service allows you to get the results of all consuming instances returning messages on the ACK queue. This service is used by the Public Monitoring app.

The service is configured in the monitoring module in /modules/magnolia-rabbitmq-monitoring/rest-endpoints/pubstate.

Properties:

To enable the service:

1. In the Security app add the following permissions to the rest role: 
   

   Access control lists:

   Workspace	Permission	Scope	Path
   PblcMntrng	Read only	Selected and sub nodes	/

   Web access:

   Permission	Path
   Get & Post	.rest/public-monitoring*

2. Assign the rest role to a user or to anonymous if you need access without credentials.

You can use the REST endpoint as follows:

curl http://monitoringinstance:port/.rest/public-monitoring/all

This should return something like:

[{"workspace":"dam","address":"192.168.10.85","name":"administrators-MacBook-Pro.local","syncStamp":"1466422838236","diff":"0","uuid":"2acd7055-7168-4926-b66a-47400ab78d2e","seqNbr":"1977","exceptions":"{}"}]

Properties:

Client control REST service

The client control REST service allows you to restart the consumer on the remote instance via a cURL command.

The service is configured in the connector module in /modules/rabbitmq-connector/rest-endpoints/consumer.

Properties:

To enable the service:

1. In the Security app add the following permission to the rest role: 

   Web access:

   Permission	Path
   Get & Post	.rest/rbmqClients*

2. Assign the rest role to a user or to anonymous if you need access without credentials.

You can use the REST endpoint as follows:

curl http://publicaddress:publicport/.rest/rbmqClients/restartAll

The command will restart all consumers and clients on that instance.

Continuous deployment strategy

How to synchronize new public instances

Public instance backups should be done when the SyncState message id is identical for all public nodes. You can then choose one public node and to do a dump. Once the dump is done, create a new historical queue and empty the old queue by consuming all messages. The instance will then be ready to receive new activation messages which post-date the last dump. 

Whenever you need to create a new instance, always use the latest dump. The new instance will be registered to the new historical queue. This will mean that the message id is in SyncState with the latest activation message. New activation messages will be sent to the new historical queue before they register in the old historical queue. The new historical queue will start getting all messages after the last message in old historical queue.

Adding new instances

This series of diagrams shows the best setup for scaling up when adding new public instances to the load balancer.

1. The basic idea is to always keep a spare queue (which does not have a consumer) connected to the exchange. 
   
2. The spare queue stores all activation messages until a new consumer connects. The newly created instance is created with the same initial data as Public 1.
   
3. Public 2 is created with the same data set as Public 1. Before Public 2 starts consuming a new queue (without a consumer) is created.
   

4. Once the remaining messages are consumed a backup is created. This serves as the new state for new instances. This is also the time to add the new public instance to the load balancer.

   When to do a backup?

   Each message has a version number or tag in its header. When a new public is added it can tell the author instance or the exchange about its presence by sending its identity before consuming starts. This enables the exchange or author to create the new queue for the public instance.

   From this point, all activation messages contain a new version or tag and the consumer starts consuming messages in the queue. When the consumer comes across a message with the newest version number, it does the backup.

5. Activation continues and the spare queue (queue_pub_3) is filled with new messages that represent the difference between Public 2 and newest content.
   

Adding a queue after backup

This example shows you how to add a new queue to the example testfanoutAck exchange:

1. Add the queue:

   curl -i -u guest:guest -H "content-type:application/json" \
     -XPUT -d'{"auto_delete":false,"durable":true,"arguments":{},"node":"rabbit@localhost"}' \
     http://localhost:55672/api/queues/%2f/fan2

2. Declare it to be bound to correct exchange:

   curl -i -u guest:guest -H "content-type:application/json" \
      -d'{"routing_key":"","arguments":[]}' \
     http://localhost:55672/api/bindings/%2f/fan2/testfanoutAck

See RabbitMQ Management HTTP API for more.

Troubleshooting and special uses cases

Node Activation Policy

The node order is respected when a node is created, but it is not respected on update.

To ensure that the node order is respected on update you can add the PUB_POLICY property to the parent of the nodes that need a specific order.  

The PUB_POLICY property has two valid values that result in the following behaviors:

Value	Behavior
respect_order	Nodes are always deleted on update. You need to use the Publish incl.subpages/nodes action to activate or the subnodes will not be created.
remove_orphan_nodes	Nodes whose parent was not found in the correct place are ignored. (Not implemented yet.)

ACK queue blocked by unacked node

When a consumer encounters an issue because of an activation exception, the consumer stops consuming. The message is sent back to the queue with a non-ack, and an exception is sent over the ACK exchange. 

To illustrate this we deleted a node on the public instance and then tried to deactivate it on the author instance. 

The deletion and attempted deactivation resulted in the following error:

ERROR lia.rabbitmq.activation.jobs.ActivationConsumerJob: /bla

The exception is visible in the Public Monitoring app. You can see from the PathNotFoundException that the consumer did not find the path that was deleted directly of the public instance.

In the RabbitMQ console the:

• Message is not consumed in the queue.
• Message is marked unacked.
• Consumer is gone,

When more nodes are published into the queue they become blocked behind the unacked message.

There are two ways to solve this:

1. Recreate the node on the public instance and restart the consumer.
   •  The sequence number count remains correct.
   •  You have to connect to the public instance and do "manual manipulation".
2. Remove the faulty activation message from the queue and restart the consumer.
   •  This can be done remotely and there is no need to connect "manually" to the public instance. 
   •  Because there is one "missing" message the sequence number is lost.

Here's a step-by-step solution using option 2 (because it demonstrates how to delete an unacked message):

1. Close the channel in the RabbitMQ console. This acks all unacked messages and is needed because messages can no longer be consumed.
    
2. Force close the connection
    
3. All messages are now acked in the queue.
    
4. Consume only the "first" message.
   

   

   

5. Restart the consumer on the public instance.

   curl http://localhost:8081/.rest/rbmqClients/restartAll

6. The instance is getting activated again.