You can deploy a pair of NetScaler virtual appliances with multiple NICs in an active-passive high availability (HA) setup on Azure. Each NIC can contain multiple IP addresses.
An active-passive deployment requires:
- An HA Independent Network Configuration (INC) configuration
- The Azure Load Balancer (ALB) in Direct Server Return (DSR) mode
All traffic goes through the primary node. The secondary node remains in standby mode until the primary node fails.
Diagram: Example of an active-passive deployment architecture
In an active-passive deployment, the ALB floating public IP (PIP) addresses are added as the VIP addresses in each NetScaler node. In HA-INC configuration, the VIP addresses are floating and SNIP addresses are instance specific.
ALB monitors each NetScaler instances by sending health probe at every 5 seconds and redirects traffic to that instance only that sends health probes response on regular interval. So in an HA setup, the primary node responds to health probes and secondary does not. If the primary instances misses two consecutive health probes, ALB does not redirect traffic to that instance. On failover, the new primary starts responding to health probes and the ALB redirects traffic to it. The standard NetScaler HA failover time is three seconds. The total failover time that might take for traffic switching can be maximum of 13 seconds.
You can deploy a NetScaler pair in active-passive HA mode in two ways by using:
- Citrix NetScaler Standard HA template: use this option to configure an HA pair with the default option of three subnets and six NICs.
- Windows PowerShell commands: use this option to configure an HA pair according to your subnet and NIC requirements.
This topic describes how to deploy a a NetScaler pair in active-passive HA setup by using the Citrix template. If you want to use PowerShell commands, see Configuring an HA Setup with Multiple IP Addresses and NICs by Using PowerShell Commands.
You can quickly and efficiently deploy a pair of NetScaler instances in HA-INC mode by using the standard template. The template creates two nodes, with three subnets and six NICs. The subnets are for management, client, and server-side traffic, and each subnet has two NICs for both the VPX instances.
You can get the NetScaler 12.0 HA Pair template at the Azure Marketplace:
To use the template:
1. From Azure Marketplace, select and initiate the Citrix solution template. The template appears.
2. Ensure deployment type is Resource Manager and select Create.
3. The Basics page appears. Create a Resource Group and select OK.
4. The General Settings page appears. Type the details and select OK.
5. The Network Setting page appears. Check the vnet and subnet configurations, edit the required settings, and select OK.
6. The Summary page appears. Review the configuration and edit accordingly. Select OK to confirm.
7. The Buy page appears. Select Purchase to complete the deployment.
It might take a moment for the Azure Resource Group to be created with the required configurations. After completion, select the Resource Group to see the configuration details, such as LB rules, back-end pools, health probes, and so on, in the Azure portal. The HA pair appears as VPX0 and VPX1.
If further modifications are required for your HA setup, such as creating more security rules and ports, you can do that from the Azure portal.
Next, you need to configure the load-balancing vserver with the ALB public IP (PIP) address, on each node. To find the ALB PIP, select ALB > Frontend IP configuration.
See the Resources section for more information about how to configure the load-balancing vserver.
The following links provide additional information related to HA deployment and virtual server (vserver) configuration: