You can use route monitors to make the HA state dependent on the internal routing table, whether or not the table contains any dynamically learned or static routes. In an HA configuration, a route monitor on each node watches the internal routing table to make sure that a route entry for reaching a particular network is always present. If the route entry is not present, the state of the route monitor changes to DOWN.
When a NetScaler appliance has only static routes for reaching a network, and you want to create a route monitor for the network, you must enable monitored static routes (MSR) for the static routes. MSR removes unreachable static routes from the internal routing table. If MSR is disabled on static routes, an unreachable static route can remain in the internal routing table, defeating the purpose of having the route monitor.
Route Monitors are supported both in non-INC and INC mode.
|Route Monitors in HA in non-INC mode
||Route Monitors in HA in INC mode
|Route monitors are propagated by nodes and exchanged during synchronization.
||Route monitors are neither propagated by nodes nor exchanged during synchronization.
|Route monitors are active only in the current primary node.
||Route monitors are active on both the primary and the secondary node.
|The NetScaler appliance always displays the state of a route monitor as UP irrespective of the whether the route entry is present or not in the internal routing table.
||The NetScaler appliance displays the state of the route monitor as DOWN if the corresponding route entry is not present in the internal routing table.
|A route monitor starts monitoring its route after 180 seconds in the following cases [This is done to allow dynamic routes to get learnt, which may take 180 secs]:
- set route6 command for v6 routes
- set route msr enable/disable command for v4 routes.
- adding a new route monitor
Route monitors are useful in a non-INC mode HA
configuration where you want the non-reachability of a gateway from
a primary node to be one of the conditions for HA
Consider an example of a non-Inc mode HA setup
in a two-arm topology that has NetScaler appliances NS1 and NS2 in
the same subnet, with router R1 and switches SW1, SW2, and
Because R1 is the only router in this setup,
you want the HA setup to failover whenever R1 is not reachable from
the current primary node. You can configure a route monitor (say,
RM1 and RM2, respectively) on each of the nodes to monitor the
reachability of R1 from that node.
With NS1 as the current primary node, the execution flow is as follows:
- Route monitor RM1 on NS1 monitors NS1's internal routing table for the presence of a route entry for router R1. NS1 and NS2 exchange heartbeat messages through switch SW1 or SW3 at regular intervals.
- If switch SW1 goes down, the routing protocol on NS1 detects that R1 is not reachable and therefore removes the route entry for R1 from the internal routing table. NS1 and NS2 exchanges heartbeat messages through switch SW3 at regular intervals.
- Detecting that the route entry for R1 is not present in the internal routing table, RM1 initiates a failover. If route to R1 is down from both NS1 and NS2, failover happens every 180 seconds till one of the appliances is able to reach R1 and restore the connectivity.