Add an additional Search node to a cluster 

You can add another Search node to a cluster where there is an existing Search node for load-balancing purposes or if you want to have a backup node for redundancy.


Before starting

IBM WAS Network Deployment (Application Server option) must installed on the new node.


Procedure

To add a second Search node to a cluster, complete these steps.

  1. Add a new node to the dmgr cell...

    1. Log in to the new node.

    2. Open a command prompt and change to the bin directory of the local WAS profile:

        app_server_root/profiles/<profile_name>/bin

        where <profile_name> is the name of the applicable WAS profile on this node.

    3. Run the addNode command to add the node to the dmgr cell:

        addnode [dmgr_host] [dmgr_port] [-username uid] [-password pwd] [-localusername localuid] [-localpassword localpwd]

        where:

        • dmgr_host is the host name of the dmgr.

        • dmgr_port is the SOAP port of the dmgr. The default is 8879.

        • uid and pwd are the DM administrator username and password respectively.

        • localuid and localpwd are the username and password respectively for the node's WAS administrator.

    4. Open the addNode.log file and confirm that the node was successfully added to the dmgr cell.

        The file is stored in the following location:

        app_server_root/profiles/<profile_name>/log/addNode.log

  2. Copy the relevant JDBC files from the dmgr node to the new node, placing them in the same location as the JDBC files on the dmgr.

      For example, if you copied the db2jcc.jar file from the C:\IBM\SQLLIB directory on the dmgr, you need to copy the same file to the C:\IBM\SQLLIBdirectory on the new node. Use the following table to determine which files to copy.

      Table 1. JDBC files

        Database type

        JDBC files

        DB2

        db2jcc.jar

        db2jcc_license_cu.jar sql

        Oracle

        ojdbc6.jar

        Microsoft SQL Server

        sqljdbc4.jar


  3. Ensure that the shared folders that are used for the application content stores in the cluster are accessible from the new node.

  4. Add the node as an additional member to the cluster.

    1. Log in to the dmgr Integration Solutions Console.

    2. Select Servers>Clusters>cluster_name>Cluster members>New.

    3. Specify the following information about the new cluster member:

        Member name

          The name of the server instance that is created for the cluster. The dmgr creates a new server instance with this name.
          Note: Each member name in the same cluster must be unique.

        Select node

          The node where the server instance resides.

    4. Click Add Member to add this member to the cluster member list.

    5. Click Next to go to the summary page where you can examine detailed information about the cluster member. Click Finish to complete this step or click Previous to modify the settings.

    6. Click Save to save the configuration.

    7. Select Server>Servers>Clusters>cluster_name>Cluster members. In the member list, click the new member that you added in the previous step.

    8. On the detailed configuration page, click Ports to expand the port information of the member. Make a note of the WC_defaulthost and WC_defaulthost_secure port numbers.

        For example, the WC_defaulthost port number is typically 9084, while the WC_defaulthost_secure port number is typically 9447.

    9. Select Environment>Virtual Hosts>default_host>Host Aliases>New.

        Enter the following information for the host alias for the WC_defaulthost port:

        Host name

          The IP address or DNS host name of the node where the new member resides.

        Port:

          The port number for WC_defaulthost. For example, 9084.

    10. Click OK to complete the virtual host configuration, and then click Save to save the configuration.

    11. Repeat the previous two substeps to add the host alias for the WC_defaulthost_secure port.

    12. Select System administration>Nodes.

    13. In the node list page, select all the nodes where the target cluster members reside, and then click Synchronize to perform a synchronization between the nodes.

  5. Use the WAS admin console, stop the Search application on the new node.

  6. Stop all the previously existing nodes that are running the Search application.

  7. Copy the index directory from one of the existing nodes to the new node.

  8. Restart all the nodes that are running Search.


What to do next

Configure IBM HTTP Server to connect to the new node. For more information, see Configure IBM HTTP Server and Define IBM HTTP Server.

IBM AIX or Linux only: Configure the path variables to point to the Search application. For more information, see Setting AIX or Linux path variables for Search.

If you experience interoperability failure, you might be running two servers on the same host with the same name. This problem can cause the Search and News applications to fail. For more information, go to the following web page: http://publib.boulder.ibm.com/infocenter/wasinfo/v6r1/index.jsp?topic=/com.ibm.websphere.base.doc/info/aes/ae/rtrb_namingprobs.html


Parent topic

Administer Search

Related concepts
Backing up the Search index
Configure IBM HTTP Server


Related tasks


Restore the Search index
Setting AIX or Linux path variables for Search
Add a node to a cluster
Define IBM HTTP Server

+

Search Tips   |   Advanced Search