Setting up a Cluster

This section describes how to set up a WSO2 product worker/manager separated cluster and how to configure this cluster with different third-party load balancers. The following sections give you information and instructions on how to set up your cluster.

Worker/manager separated clustering deployment pattern

In this pattern there are three WSO2 Application Server nodes; 1 node acts as the manager node and 2 nodes act as worker nodes for high availability and serving service requests. Although you can use any standard WSO2 product, WSO2 AS is used here for the purposes of this example. In this pattern, we allow access to the admin console through an external load balancer. Additionally, service requests are directed to worker nodes through this load balancer. The following image depicts the sample pattern this clustering deployment scenario will follow.

Here, we use two nodes as well-known members, one is the manager node and the other is one of the worker nodes. It is always recommended to use at least two well-known members to prevent restarting all the nodes in the cluster in case a well known member is shut down.

See Worker/Manager separated clustering patterns for a wider variety of options if you prefer to use a different clustering deployment pattern.

Configuring the load balancer

The load balancer automatically distributes incoming traffic across multiple WSO2 product instances. It enables you to achieve greater levels of fault tolerance in your cluster, and provides the required balancing of load needed to distribute traffic.

The following are some of the load balancers that you can configure and their respective configurations.

• Configuring NGINX Plus
• Configuring Apache Httpd
• Configuring AWS ELB
• Configuring HAProxy

Setting up the databases

See Setting up the Database for information on how to set up the databases for a cluster. The datasource configurations must be done in the <PRODUCT_HOME>/repository/conf/datasources/master-datasources.xml file for both the manager and worker nodes. You would also have to configure the shared registry database and mounting details in the <PRODUCT_HOME>/ repository/conf/registry.xml file.

Configuring the manager node

1. Download and unzip the WSO2 AS binary distribution. Consider the extracted directory as <PRODUCT_HOME>.
2. Set up the cluster configurations. Edit the <PRODUCT_HOME>/repository/conf/axis2/axis2.xml file as follows.
   1. Enable clustering for this node: 
      <clustering class="org.wso2.carbon.core.clustering.hazelcast.HazelcastClusteringAgent" enable="true">
   2. Set the membership scheme to wka to enable the well-known address registration method (this node sends cluster initiation messages to the WKA members that we define later). 
      <parameter name="membershipScheme">wka</parameter>
   3. Specify the name of the cluster this node will join.
      <parameter name="domain">wso2.as.domain</parameter>
   4. Specify the host used to communicate cluster messages.
      <parameter name="localMemberHost">xxx.xxx.xxx.xx2</parameter>

   5. Specify the port used to communicate cluster messages. This port number is not affected by the port offset value specified in the <PRODUCT_HOME>/repository/conf/ carbon.xml. If this port number is already assigned to another server, the clustering framework automatically increments this port number. However, if two servers are running on the same machine, you must ensure that a unique port is set for each server.
      <parameter name="localMemberPort">4100</parameter>

   6. Specify the well known members. In this example, the well known member is a worker node. The port value for the WKA worker node must be the same value as it's localMemberPort (in this case 4200).

      <members>
      	<member>
      		<hostName>xxx.xxx.xxx.xx3</hostName>
      		<port>4200</port> 
      	</member>
      </members>

      Although this example only indicates one well-known member, it is recommended to add at least two well-known members here. This is done to ensure that there is high availability for the cluster.

      You can also use IP address ranges for the hostName. For example,  192.168.1.2-10. This should ensure that the cluster eventually recovers after failures. One shortcoming of doing this is that you can define a range only for the last portion of the IP address. You should also keep in mind that the smaller the range, the faster the time it takes to discover members, since each node has to scan a lesser number of potential members.

   7. Change the following clustering properties. 

      <parameter name="properties">
                  <property name="backendServerURL" value="https://${hostName}:${httpsPort}/services/"/>
                  <property name="mgtConsoleURL" value="https://${hostName}:${httpsPort}/"/>
      </parameter>

3. Configure the HostName. To do this, edit the <PRODUCT_HOME>/repository/conf/carbon.xml file as follows.

   <HostName>as.wso2.com</HostName>
   <MgtHostName>mgt.as.wso2.com</MgtHostName>

4. Enable SVN-based deployment synchronization with the AutoCommit property marked as true. To do this, edit the <PRODUCT_HOME>/repository/conf/carbon.xml file as follows. See Configuring Deployment Synchronizer for more information on this.

   <DeploymentSynchronizer>
   	<Enabled>true</Enabled>
   	<AutoCommit>true</AutoCommit>
   	<AutoCheckout>true</AutoCheckout>
   	<RepositoryType>svn</RepositoryType>
   	<SvnUrl>https://svn.wso2.org/repos/as</SvnUrl>
   	<SvnUser>svnuser</SvnUser>
   	<SvnPassword>xxxxxx</SvnPassword>
   	<SvnUrlAppendTenantId>true</SvnUrlAppendTenantId>
   </DeploymentSynchronizer>

5. In the <PRODUCT_HOME>/repository/conf/carbon.xml file, you can also specify the port offset value. This is ONLY applicable if you have multiple WSO2 products hosted on the same server.

    Click here for more information on configuring the port offset.

   When you run multiple products/clusters or multiple instances of the same product on the same server or virtual machines (VMs), you must change their default ports with an offset value to avoid port conflicts. An offset defines the number by which all ports in the runtime (e.g., HTTP/S ports) are increased. For example, if the default HTTP port is 9763 and the offset is 1, the effective HTTP port will change to 9764. For each additional product instance, you set the port offset to a unique value. The offset of the default ports is considered to be 0.

   The port value will automatically be increased as shown in the Port Value column in the following table, allowing all five WSO2 product instances or servers to run on the same machine.

   WSO2 product instance	Port Offset	Port Value
   WSO2 server 1	0	9443
   WSO2 server 2	1	9444
   WSO2 server 3	2	9445
   WSO2 server 4	3	9446
   WSO2 server 5	4	9447

   Important: This step is optional and only required if all server instances are running in the same machine. This is not a recommended approach for production environments. Note that you need to change all ports used in your configurations based on the offset value if you are to do this.

   <Ports>
   	...
   	<Offset>0</Offset>
   	...
   </Ports>

6. Map the host names to the IP. Add the below host entries to your DNS, or “/etc/hosts” file (in Linux) in all the nodes of the cluster. You can map the IP address of the database server. In this example, MySQL is used as the database server, so <MYSQL-DB-SERVER-IP> is the actual IP address of the database server.

   <IP-of-MYSQL-DB-SERVER>	carbondb.mysql-wso2.com

7. Allow access the management console only through the load balancer. Configure the HTTP/HTTPS proxy ports to communicate through the load balancer by editing the <PRODUCT_HOME>/repository/conf/tomcat/catalina-server.xml file as follows.

   <Connector protocol="org.apache.coyote.http11.Http11NioProtocol"
   	port="9763"
   	proxyPort="80"
   	...
   	/>
   <Connector protocol="org.apache.coyote.http11.Http11NioProtocol"
   	port="9443"
   	proxyPort="443"
   	...
   	/>

    Click here for more information on this configuration.

   The Connector protocol tag sets the protocol to handle incoming traffic. The default value is HTTP/1.1, which uses an auto-switching mechanism to select either a blocking Java-based connector or an APR/native connector. If the PATH (Windows) or LD_LIBRARY_PATH (on most UNIX systems) environment variables contain the Tomcat native library, the APR/native connector will be used. If the native library cannot be found, the blocking Java-based connector will be used. Note that the APR/native connector has different settings from the Java connectors for HTTPS.

   The non-blocking Java connector used is an explicit protocol that does not rely on the auto-switching mechanism described above. The following is the value used:
   org.apache.coyote.http11.Http11NioProtocol

   The TCP port number is the value that this Connector will use to create a server socket and await incoming connections. Your operating system will allow only one server application to listen to a particular port number on a particular IP address. If the special value of 0 (zero) is used, Tomcat will select a free port at random to use for this connector. This is typically only useful in embedded and testing applications.

Configuring the worker node

1. Download and unzip the WSO2 AS binary distribution. Consider the extracted directory as <PRODUCT_HOME>.
2. Set up the cluster configurations. Edit the <PRODUCT_HOME>/repository/conf/axis2/axis2.xml file as follows.
   1. Enable clustering for this node.
      <clustering class="org.wso2.carbon.core.clustering.hazelcast.HazelcastClusteringAgent" enable="true">
   2. Set the membership scheme to wka to enable the well-known address registration method (this node will send cluster initiation messages to WKA members that we will define later).
      <parameter name="membershipScheme">wka</parameter>
   3. Specify the name of the cluster this node will join.
      <parameter name="domain">wso2.as.domain</parameter>
   4. Specify the host used to communicate cluster messages.
      <parameter name="localMemberHost">xxx.xxx.xxx.xx3</parameter>

   5. Specify the port used to communicate cluster messages. If this node is on the same server as the manager node, or another worker node, set this to a unique value, such as 4000 and 4001 for worker nodes 1 and 2. This port number will not be affected by the port offset in carbon.xml. If this port number is already assigned to another server, the clustering framework will automatically increment this port number.
      <parameter name="localMemberPort">4200</parameter>

   6. Specify the well known member by providing the host name and localMemberPort values. Here, the well known member is the manager node. Defining the manager node is useful since it is required for the Deployment Synchronizer to function in an efficient manner. The deployment synchronizer uses this configuration to identify the manager and synchronize deployment artifacts across the nodes of a cluster.

      <members>
      	<member>
      		<hostName>xxx.xxx.xxx.xx2</hostName>
      		<port>4100</port> 
      	</member>
      </members>

      Although this example only indicates one well-known member, it is recommended to add at least two well-known members here. This is done to ensure that there is high availability for the cluster.

      You can also use IP address ranges for the hostName. For example,  192.168.1.2-10. This should ensure that the cluster eventually recovers after failures. One shortcoming of doing this is that you can define a range only for the last portion of the IP address. You should also keep in mind that the smaller the range, the faster the time it takes to discover members, since each node has to scan a lesser number of potential members.

3. Configure the HostName. To do this, edit the <PRODUCT_HOME>/repository/conf/carbon.xml file as follows.

   <HostName>as.wso2.com</HostName>

4. Enable SVN-based deployment synchronization with the AutoCommit property marked as false. To do this, edit the <PRODUCT_HOME>/repository/conf/carbon.xml file as follows.

   <DeploymentSynchronizer>
   	<Enabled>true</Enabled>
   	<AutoCommit>false</AutoCommit>
   	<AutoCheckout>true</AutoCheckout>
   	<RepositoryType>svn</RepositoryType>
   	<SvnUrl>https://svn.wso2.org/repos/as</SvnUrl>
   	<SvnUser>svnuser</SvnUser>
   	<SvnPassword>xxxxxx</SvnPassword>
   	<SvnUrlAppendTenantId>true</SvnUrlAppendTenantId>
   </DeploymentSynchronizer>

5. In the <PRODUCT_HOME>/repository/conf/carbon.xml file, you can also specify the port offset value. This is ONLY applicable if you have multiple WSO2 products hosted on the same server.

    Click here for more information on configuring the port offset.

   When you run multiple products/clusters or multiple instances of the same product on the same server or virtual machines (VMs), you must change their default ports with an offset value to avoid port conflicts. An offset defines the number by which all ports in the runtime (e.g., HTTP/S ports) are increased. For example, if the default HTTP port is 9763 and the offset is 1, the effective HTTP port will change to 9764. For each additional product instance, you set the port offset to a unique value. The offset of the default ports is considered to be 0.

   The port value will automatically be increased as shown in the Port Value column in the following table, allowing all five WSO2 product instances or servers to run on the same machine.

   WSO2 product instance	Port Offset	Port Value
   WSO2 server 1	0	9443
   WSO2 server 2	1	9444
   WSO2 server 3	2	9445
   WSO2 server 4	3	9446
   WSO2 server 5	4	9447

   Important: This step is optional and only required if all server instances are running in the same machine. This is not a recommended approach for production environments. Note that you need to change all ports used in your configurations based on the offset value if you are to do this.

   <Ports>
   	...
   	<Offset>0</Offset>
   	...
   </Ports>

6. Map the host names to the IP. Add the below host entries to your DNS, or “/etc/hosts” file (in Linux) in all the nodes of the cluster. You can map the IP address of the database server. In this example, MySQL is used as the database server, so <MYSQL-DB-SERVER-IP> is the actual IP address of the database server.

   <IP-of-MYSQL-DB-SERVER>	carbondb.mysql-wso2.com

7. Configure the HTTP/HTTPS proxy ports to communicate through the load balancer by editing the <PRODUCT_HOME>/repository/conf/tomcat/catalina-server.xml file as follows.

   <Connector protocol="org.apache.coyote.http11.Http11NioProtocol"
   port="9763"
   proxyPort="80"
   	...
   	/>
   <Connector protocol="org.apache.coyote.http11.Http11NioProtocol"
   port="9443"
   proxyPort="443"
   	...
   	/>

8. Create the second worker node by getting a copy of the WSO2 product you just configured as a worker node and change the following in the <PRODUCT_HOME>/repository/conf/axis2/axis2.xml file. This copy of the WSO2 product can be moved to a server of its own.
   <parameter name="localMemberPort">4300</parameter>

Testing the cluster

1. Restart the configured load balancer.
2. Start the manager node. The additional -Dsetup argument creates the required tables in the database. 
   sh <PRODUCT_HOME>/bin/wso2server.sh -Dsetup

3. Start the two worker nodes. The additional -DworkerNode=true argument indicates that this is a worker node. This parameter basically makes a server read-only. A node with this parameter will not be able to do any changes such as writing or making modifications to the deployment repository etc. This parameter also enables the worker profile, where the UI bundles will not be activated and only the back end bundles will be activated once the server starts up. When you configure the axis2.xml file, the cluster sub domain must indicate that this node belongs to the "worker" sub domain in the cluster.

   It is recommendation is to delete the  <PRODUCT_HOME>/repository/deployment/server  directory and create the following directory structure in order to avoid any SVN conflicts that may arise.

   • <PRODUCT_HOME>/repository/deployment/server

     • <PRODUCT_HOME>/repository/deployment/server/synapse-configs/

       • <PRODUCT_HOME>/repository/deployment/server/synapse-configs/default

   As depicted above you need to create an empty directory named server in the <APIM_Worker_HOME>/repository/deployment directory. Thereafter, create a subdirectory named synapse-configs under the server directory, and create another subdirectory named default under the synapse-configs directory. This is done to avoid any SVN conflicts that may arise.

   sh <PRODUCT_HOME>/bin/wso2server.sh -DworkerNode= true

4. Check for ‘member joined’ log messages in all consoles.

   Additional information on logs and new nodes

   When you terminate one node, all nodes identify that the node has left the cluster. The same applies when a new node joins the cluster. 

   If you want to add another new worker node, you can simply copy worker1 without any changes if you are running it on a new server (such as xxx.xxx.xxx.184). If you intend to use the new node on a server where another WSO2 product is running, you can use a copy of worker1 and change the port offset accordingly in the carbon.xml file. You also have to change localMemberPort in axis2.xml if that product has clustering enabled. Be sure to map all host names to the relevant IP addresses when creating a new node. The log messages indicate if the new node joins the cluster.

5. Access management console through the LB using the following URL: https://xxx.xxx.xxx.xx1:443/carbon

6. Test load distribution via http://xxx.xxx.xxx.xx1:80/.