You can configure multiple Policy Servers in a cluster for failover operation. Clustering servers enable failover from one group of servers to another.
Note: If you are using a hardware load balancer to expose Policy Servers as multiple virtual IP addresses (VIPs), we recommend that you configure those VIPs in a failover configuration. Clustering is redundant because the hardware load balancer performs the same function more efficiently.
Policy Server clusters are defined as part of a Host Configuration Object. When a SiteMinder Web Agent initializes, the settings from the Host Configuration Object are used to setup communication with Policy Servers.
Note: The following procedure assumes that you are creating an object. You can also copy the properties of an existing object to create an object.
To configure a cluster
The Host Configuration Objects page appears.
The Create Host Configuration page appears.
Note: By default, the Create a new object of type Host Configuration option is selected.
Note: Click Help for descriptions of settings and controls, including their respective requirements and limits.
The Add Cluster page appears.
Note: To add another Policy Server to the cluster, click Add to Cluster. To delete a Policy Server from the cluster, click the minus sign next to Port. To change the sequence of Policy Servers in the cluster, click the up and down arrows.
The cluster is added.
Note: To modify a cluster, click the right-facing arrow in the first column. To delete a cluster, click the minus sign in the last column. To add another cluster, click Add, and repeat steps 7 and 8.
SM automatically calculates the Failover Threshold value that is based on both the number of Policy Servers configured in the cluster and the Failover Threshold Percent you specify. The Failover Threshold value (displayed in the column to the right of the list of servers in each cluster) is the total number of Policy Servers that must be active in the Cluster to consider the Cluster available based on your configuration. If the number of Active Servers in the Cluster falls below the displayed Failover Threshold value, the Cluster fails over to the next available Cluster in the list. If the calculation of the Failover Threshold value that is based on your configuration would result in a non-whole number, the Policy Server rounds the value up to the nearest integer.
For example, consider a Cluster with 5 Policy Servers configured.
With a Failover Threshold Percentage ...;
...in the range between [0% - 19%], the value of required active servers (Failover Threshold) is rounded up to 1 .
...of 20%, the calculated value for Failover Threshold is 1 .
in the range between [21% - 39%], the value for Failover Threshold is rounded up to 2 .
of 40%, the calculated value for Failover Threshold is 2 .
in the range between [41% - 59%], the value for Failover Threshold is rounded up to 3 .
of 60%, the calculated value for Failover Threshold is 3 .
in the range between [61% - 79%], the value for Failover Threshold is rounded up to 4 .
of 80%, the calculated value for Failover Threshold is 4 .
in the range between [81% - 99%], the value for Failover Threshold is rounded up to 5 .
of 100%, the calculated value for Failover Threshold is 5 .
When you set the Failover Threshold Percentage, it applies to all clusters that use the Host Configuration Object.
Note: If the percentage of active servers in a cluster falls below the specified percentage, the cluster fails over to the next available cluster in the cluster list.
The Host Configuration Object is configured with the Policy Server.
Important! Configuration Values specifies a single Policy Server and a simple failover operation that are only used when no clusters are specified. If you decide to delete all clusters in favor of a simple failover operation, be sure to delete all Policy Server information.
|
Copyright © 2012 CA Technologies.
All rights reserved.
|
|