com.atlassian.confluence.content.render.xhtml.migration.exceptions.UnknownMacroMigrationException: The macro 'next_previous_link3' is unknown.

Viewing Cluster Information

Owned by Former user (Deleted)
Mar 14, 2014
3 min read

When WSO2 Message Broker is running in clustered mode, queues and topic subscriptions are distributed. Although this is transparent to the user, he may still need to know what nodes are there in the cluster and how queue ownership is dynamically decided. This usability aspect is managed by the Cluster management feature, which allows viewing cluster information using the management console of each Message Broker node. 

Following capabilities are provided by Cluster Management:

  • View Zookeeper coordination details.
  • View Cassandra connection details.
  • View how queue ownership has been decided.
  • View individual, node-wise queue information.

To access each capability,

1. log in to the Message Broker management console and select "Cluster information - Node List" in the "Main" menu.

2. The "Cluster Management" window opens. For example,

The following information is provided by the window.

Cassandra Connection Details

WSO2 MB uses Apache Cassandra as the Message Storage. Therefore, during cluster configuration, the MB instance should be pointed to a running Apache Cassandra server. These connection details are shown here.

Zookeeper Connection Details

WSO2 MB should be pointed to a running Apache Zookeeper server during cluster configuration. User can view that configuration through the "Cluster Management" window.

Node ID

Node ID is assigned to each node joined in the cluster, offering a unique way to identify each node. This is useful when troubleshooting. Each node is considered as a peer, and there's no master-slave concept in WSO2 Message Broker. This section shows the node id assigned to the node and if the node is configured to be clustered or running in standalone mode.

For example,

node ID display

View how queue ownership is decided

When clustered, each queue has an owner (handling the actual processing of the queue), which is dynamically decided according to the load distributed among the cluster nodes. User can view what nodes process which queues, through a table in the "Cluster Management" window as follows.

 

  • Node ID: This table lists all the nodes that are visible to the server accessed by the management console. After setting up the cluster, users can verify using this table that each server can see one another in the cluster.
  • IP Address of Node: Displays the IP address of each node listed.
  • Total Node Subscriptions: Shows the total number of queue subscriptions each node is having. You can view queue-wise subscriptions for the node clicking on the Number of Queue Workers in front of each node. 
  • Number of Queue Workers: Shows how many queues are owned by that server node. 

3. Click the number in the "Number of Queue Workers" column to view individual queue information.

View individual, node-wise queue information

4. This shows the names of the queues owned by the node clicked in step 3. The number of messages yet to be copied from Global Queue (queue space common to all nodes in cluster) to Node Queues (queue space private to each node) is also shown. For example,

 

In the same page analysis of 

  • How many messages to be delivered are accumilated on the selected node for each user definded queue in the cluster
  • How many subscriptions selected node is having for each queue defined in the cluster

is shown. Below is an example of that view. 

com.atlassian.confluence.content.render.xhtml.migration.exceptions.UnknownMacroMigrationException: The macro 'next_previous_links2' is unknown.