A Load Balancer such as the Citrix NetScaler has monitors built in, to automatically redirect Client requests to the back-end servers that are up and running. But, what happens if the front-end is down? Load Balancing can fail when a Virtual Server or Vserver fails, or when the VIP is unable to handle excessive traffic. Protecting the LB setup against failure helps ensure the availablility of the websites and applications.

Configuring Connection-Based Spillover

You can use connection-based spillover to configure a maximum threshold for the number of active client connections on a virtual server. When the client connections exceed the configured threshold limit, new client connections are diverted to the backup virtual server.

To configure connection-based spillover, follow the steps described in the section Diverting Excess Traffic to a Backup Load Balancing Virtual Server. In the Method list, select Connection.

Global Server Load Balancing (GSLB) virtual servers do not support connection-based spillover.

Configuring Dynamic Spillover

Dynamic spillover depends on the maximum client setting configured on the services. If the number of client connections at the virtual server exceeds the sum of the maximum client values, the new connections are diverted to the services of the backup virtual server. To configure dynamic spillover, you must enable it on a virtual server. You must configure the services with appropriate maximum client values. If the value for maximum client is set to 0, the spillover limit is treated as infinity, and spillover never occurs. To configure dynamic spillover, follow the steps described in the section Diverting Excess Traffic to a Backup Load Balancing Virtual Server. In the Method list, select Dynamic connection.

Configuring Bandwidth-Based Spillover

Bandwidth-based spillover allows you to configure a bandwidth threshold value. When the bandwidth on the primary virtual server exceeds the bandwidth threshold value, the NetScaler diverts new connections to a backup virtual server. For example, if you create a primary virtual server, a backup virtual server, and two services, and bind the services to the virtual servers, you can configure bandwidth spillover on the primary virtual server. You can also configure the backup virtual server with a threshold value. When the threshold for the backup virtual server is reached, the NetScaler diverts new client connections to the next backup virtual server. To configure bandwidth-based spillover, follow the steps described in the section Diverting Excess Traffic to a Backup Load Balancing Virtual Server. In the Method list, select Bandwidth.

Configuring Stateful Connection Failover

The NetScaler enables TCP or UDP connections to survive a failover event in high availability (HA) mode. This functionality is called connection failover. After adding a secondary node and configuring HA, you can choose to configure either stateless or stateful connection failover. You can configure connection failover on any LB virtual server.

To configure a stateful connection failover by using the NetScaler command line

At the NetScaler command prompt, type:

  • set lb vserver <vServerName> -connFailover <Value>

Example

set lb vserver Vserver-LB-1 -connFailover stateful

To configure a stateful connection failover by using the GUI configuration utility

  1. In the navigation pane, expand Load Balancing, and then click Virtual Servers.
  2. In the Load Balancing Virtual Servers page, select the virtual server for which you want to configure connection failover (for example, Vserver-LB-1), and click Open.
  3. On the Advanced tab, in the Connection Failover drop-down list, select Stateful.
  4. Click OK.




    Detailed instructions can be found in the NetScaler Traffic Management Guide




    NetScaler Documentation can be found here




    Download the Citrix Load Balancer




    The Citrix Community is powerful!