Configuring high availability nodes in different subnets

The following figure shows an HA deployment with the two systems located in different subnets:

Figure 1. High Availability over a Routed Network

localized image

In the figure, the systems NS1 and NS2 are connected to two separate routers, R3 and R4, on two different subnets. The NetScaler appliances exchange heartbeat packets through the routers. This configuration could be expanded to accommodate deployments involving any number of interfaces.

Note:

If you use static routing on your network, you must add static routes between all the systems to ensure that heartbeat packets are sent and received successfully. (If you use dynamic routing on your systems, static routes are unnecessary.)

If the nodes in an HA pair reside on two separate networks, the primary and secondary node must have independent network configurations. This means that nodes on different networks cannot share entities such as SNIPs, VLANs, and routes. This type of configuration, where the nodes in an HA pair have different configurable parameters, is known as Independent Network Configuration (INC) or Symmetric Network Configuration (SNC).

The following table summarizes the configurable entities and options for an INC, and shows how they must be set on each node.

NetScaler entities Options
IPs (NSIP/SNIPs) Node-specific. Active only on that node.
VIPs Floating.
VLANs Node-specific. Active only on that node.
Routes Node-specific. Active only on that node. Link load balancing routes are floating.
ACLs Floating (Common). Active on both nodes.
Dynamic routing Node-specific. Active only on that node. The secondary node should also run the routing protocols and peer with upstream routers.
L2 mode Floating (Common). Active on both nodes.
L3 mode Floating (Common). Active on both nodes.
Reverse NAT (RNAT) Node-specific. RNAT with VIP, because NATIP is floating.

As in configuring HA nodes in the same subnet, to configure HA nodes in different subnets, you log on to each of the two NetScaler appliances and add a remote node representing the other appliance.

Adding a Remote Node

When two nodes of an HA pair reside on different subnets, each node must have a different network configuration. Therefore, to configure two independent systems to function as an HA pair, you must specify INC mode during the configuration process.

When you add an HA node, you must disable the HA monitor for each interface that is not connected or not being used for traffic. For CLI users, this is a separate procedure.

To add a node by using the command line interface

At the command prompt, type:

  • add ha node <id> <IPAddress> -inc ENABLED
  • show ha node

Example

> add ha node 3 10.102.29.170 -inc ENABLED
 Done
> add ha node 3 1000:0000:0000:0000:0005:0600:700a:888b
  Done

To disable an HA monitor by using the command line interface

At the command prompt, type:

  • set interface <ifNum> [-haMonitor ( ON | OFF )]
  • show interface <ifNum>

Example

> set interface 1/3 -haMonitor OFF
 Done

To add a remote node by using the GUI

  1. Navigate to System > High Availability and, on the Nodes tab, add a new remote node.
  2. Make sure to select the Turn off HA monitor on interfaces/channels that are down and Turn on INC (Independent Network Configuration) mode on self mode options.

Removing a Node

If you remove a node, the nodes are no longer in high availability configuration.

To remove a node by using the command line interface

At the command prompt, type:

rm ha node <id>

Example

> rm ha node 2
  Done

To remove a node by using the GUI

Navigate to System > High Availability and, on the Nodes tab, delete the node.

Note:

You can use the Network Visualizer to view the NetScaler appliances that are configured as a high availability (HA) pair and perform high availability configuration tasks.

Configuring high availability nodes in different subnets