Restricting high availability synchronization traffic to a VLAN
In a high availability (HA) deployment, traffic related to maintaining the HA configuration flows between the two HA nodes. The traffic is of the following types:
-
Config synchronization
-
Config propagation
-
Connection mirroring
-
Load balancing persistency config synchronization
-
Persistent session synchronization
-
Session state synchronization
Proper flow of this HA-related traffic between the two nodes is critical for the functioning of the HA deployment. Typically, the HA-related traffic is small in volume but can become high during a failover. It becomes high if a stateful connection failover is enabled and the node that was primary before the failover was handling many connections.
By default, the HA-related traffic flows through the VLANs to which the NSIP address is bound. To accommodate a potential surge in this traffic, you can separate the HA-related traffic from the management traffic and restrict its flow to a separate VLAN. This VLAN is called the HA SYNC VLAN.
Note:
In an HA setup, if the
tag all parameter is enabled, then the HA packets are also tagged with the VLAN traffic. However, there might be issues with the high availability communication. Therefore, it is recommended to configure NSVLAN or HA SYNC VLAN for HA traffic.
-
For NSVLAN configuration, see Configuring NSVLAN.
-
For HA SYNC VLAN configuration, see Configure HA SYNC VLAN.
Points to consider before configuring an HA SYNC VLAN
-
The HA SYNC VLAN operates at Layer 2 (L2) only.
-
The HA SYNC VLAN cannot be configured on the following types of VLANs:
-
Layer 3 (L3) VLANs
-
VLANs bound to a Traffic Domain (TD)
-
VLANs associated with a partition
-
-
HA synchronization and node-to-node communication always use the NetScaler® IP (NSIP) address, regardless of syncvlan, nsvlan, or default VLAN configuration. The NSIP serves as the primary communication channel for HA-related traffic between nodes.
-
The configuration of an HA SYNC VLAN is neither propagated nor synchronized. In other words, the HA SYNC VLAN is node specific and is configured independently on each node.
-
HA SYNC VLAN configuration is removed when you clear the configuration in only FULL mode.
-
In typical environments where both dedicated management and data traffic interfaces exist, we recommend setting HA MON to OFF for the management interfaces that are part of the HA SYNC VLAN. This setting ensures that the node is not marked as DOWN only because the management interface is down.
-
Management interfaces (for example, 0/1 and 0/2) must not be part of the HA SYNC VLAN, so that HA related traffic does not flow through management interfaces.
-
High availability heartbeat messages can also be enabled on data interfaces after:
-
Disabling messages on management interfaces
-
Enabling messages on HA SYNC VLAN interfaces
For more information on disabling high availability heartbeat messages on interfaces, see Managing high availability heartbeat messages on a NetScaler appliance. -
Configure HA SYNC VLAN
To configure an HA SYNC VLAN on a NetScaler node, specify a configured VLAN with the HA SYNC VLAN parameter of the local node entity.
Configure an HA SYNC VLAN on a local node by using the CLI
At the command prompt, type:
-
set ha node –syncvlan <VLANID> -
show node
Parameter Description:
syncvlan (Sync VLAN) - VLAN on which HA-related traffic is sent. It includes traffic for synchronization, propagation, connection mirroring, load balancing persistency, configuration synchronization, persistent session synchronization, and session state synchronization. However, HA heartbeats can use any interface.
Configure an HA SYNC VLAN on a node by using the GUI
-
Navigate to System > High Availability.
-
Set the Sync VLAN parameter while modifying the local node.