To Configure Load Balancer Health Checking

  1. In the tree component, select the HTTP Load Balancers node.
  2. In the right pane, click the load balancer to edit.
  3. Click the Targets tab.

    The HTTP Load Balancer Targets page appears.

  4. In the Health Checker field, click Edit Health Checker.

    The Edit Load Balancer Health Checker page appears.

  5. From the Policy Algorithm drop-down list, select the load balancer policy for the cluster.
    Round robin

    The load balancer assigns requests to server instances in the cluster equally. For example, if you have two instances in the cluster, the requests are assigned to each instance in turn.

    Weighted round robin

    The load balancer assigns requests based on the weights you specify for the instances on the cluster's Instances tab. For example, if you have two instances in a cluster and assign a weight of 1 to the first instance and three to the second instance, for every four requests one request goes to the first instance and three go to the second instance.

    User-defined

    The load balancer assigns requests based on the load balancer policy you created.

    If you select user-defined, the Policy Module field is activated.

  6. (User-defined policy only) In the Policy Module field, type the absolute path to the shared library that implements the user-defined policy.

    If you do not specify a path, the default is the Web Server's load balancer, which is located at /plugins/lbplugin/include/loadbalancer.

  7. Click the Enabled check box. If this check box is not selected, any load balancer using this server will consider this server as unavailable.
  8. In the Disable Timeout field, type the time (in minutes) that this server should take to reach a quiescent state after having been disabled.
  9. In the URL field, type the relative URL of a listener to ping and determine the state of health of the listener.
  10. In the Interval field, type the time (in seconds) between each health check.
  11. In the Timeout field, type the time (in seconds) that the health checker waits to receive a response from a listener.

    If the health checker does not receive a response in this interval, the listener is considered unhealthy.

  12. Click Save.