Creating a replica node in a Video Manager cluster

Create a replica node on an IBM® Sametime® Video Manager cluster. This will synchronize the replica node's data with the master data copy through the replication process and thus enable fail over and load balancing in the Sametime Video Manager clustered environment.

About this task

Perform this procedure on the replica node, not on the master Video Manager.

Procedure

  1. From the Sametime Video Manager installed on the IBM WebSphere® Application Server, edit the /root/replication_scripts/replica.config file by completing these steps:
    1. In the master_network_name setting, replace master_host_name with the hostname or IP address of the master to which a replica needs to be registered.
    2. In the master_node_name setting, specify the node name of the master node.
      For example:
      port=2315
      userId=admin
      password=admin
      replica_catalog_name=PROXIAS
      solid_db_driver=solid.jdbc.SolidDriver
      master_network_name=tcp stavvc5.rtp.raleigh.ibm.com 2315
      master_node_name=master_node_01
      tables_file_path=tables.txt
      register_replica=true
  2. On the computer hosting the Sametime Video Manager, stop the WebSphere Application Server.
  3. Run the replication_config.sh script by entering this command:

    #./replication_config.sh

  4. At these prompts, enter the values indicated:
    1. Please enter Solid DB Replication configuration choice [Mm]aster , [Rr]eplica , [Nn]one ? Enter

      r

      The message Configuring Replica... is displayed.

    2. Enter replica node name [make sure node name is unique [ex.replica_node_01] : Enter

      replica_node_03

      The value you specify is displayed.

    3. Enter the fully qualified host name or IP address of the target server to be configured as a replica:

      For example, enter

      sametime_server.renovations.com or localhost

      The value you specify is displayed.

      The script performs these tasks:
      1. Registers with master node and for will register itself with publications on master (publication_name_real_time and publication_name_real_time)
      2. The script creates these stored procedures:
        • sync_refresh_publication_name_real_time
        • sync_refresh_publication_name_config
        • propogate_trx_master_priority_2
        • propogate_trx_master_priority_1
        • check_unsent_sync_msgs
        • check_unreceived_sync_msgs
        • propogate_trx_master_priority
  5. At this prompt, enter one of these values:
    1. Run refresh / propagation scripts with default interval [Yy]es,[Nn]o ?

      If you enter y, it triggers propagation and refresh scripts with default intervals set in the intervals.properties file in the /root/replication_scripts directory.

      If you enter n, the script ends.

  6. Ensure that the replication process starts before you the start the WebSphere Application Server. Replication starts when you enter y in step 5, or when you manually run the propagate_refresh_intervals.sh script in the /root/replication_scripts directory.
    This can be verified with  ps -ef | grep -i propa command  which output similar to this example:
    root      5011     1  0 03:01 pts/2    00:00:07 /bin/sh ./propagate_1.sh interval specified in intervals.properties>
    root      5012     1  0 03:01 pts/2    00:00:02 /bin/sh ./propagate_2.sh interval specified in intervals.properties>
    root     10518   959  0 05:40 pts/2    00:00:00 grep -i propa
    
  7. To set sleep interval for the propagation and refresh scripts, edit the intervals.properties files. The intervals are in seconds. Guidelines for setting intervals :
    1. When configuring intervals in intervals.properties, the interval for sleep_interval_1 should be less than that of sleep_interval_2 because sleep_interval_1 is for the most critical data.
    2. When configuring intervals, configure slightly different intervals for each replica node so that all replica nodes do not overload the master at one time.

      For example, if for replica 1, the intervals are 10 seconds, set the intervals for replica 2 to 40 seconds.

  8. Restart replication on the Sametime Video Manager.
  9. To stop the scripts, enter this command:

    kill.sh

    The script removes the pids from pids.txt file and terminates the processes.