Clustering, load balancing and high availability in NX 3.5.0
Load balancing is supported starting from NX 2.0.0 thanks to the NX Advanced Server that provides the following functionalities:
- Multi-node support with the possibility to add and remove a new node to the NX network in any moment.
- Node selection according to a round-robin mechanism which allows for distributing the load of the sessions among all the available nodes, with other node selection methods under development.Possibility to temporarily disable access to a node, e.g. for allowing maintenance operations on the host machine.
- Possibility to disconnect user and then upon reconnection go back to same session.
- With user profiles, ensure that users only go to specific nodes. See the following article for more details on user profiles:
https://www.nomachine.com/AR10C00293
In NX 3.5.0, the NX Server controlling the nodes represents a single point of failure. In case of NX server unavailability, it will be not possible anymore to reconnect sessions and handling incoming requests for new sessions.
Because of the stateful nature of the protocol and of the applications running on the nodes, it is not possible to employ the traditional approaches used to load balance HTTP applications, as Linux Virtual Server and Cisco Local Directory.
High availability is provided with the NoMachine Cluster Server (availablesince NoMachine 4), that allows the setting up of a double front-end in order to provide automatic failovers in case of failure of NoMachine servers. Please refer to the following document for pre-requisites for a multi-node environment with load-balancing of virtual desktops and failover cluster: