How is the shared IP address handled in NoMachine Enterprise Server Cluster

[dplNX]

Hi,

I managed to have a look at 3.5. Setting-up Two Servers in a Failover Cluster and wondered how the shared ip address is handled between enterprise 1 and 2 as there is no mentioning in the prerequisites of pacemaker (heartbeat) for that matter.

Does that mean that part has been integrated into NoMachine and is handled by the nxserver itself ?

Thank you for your assistance in advance.

[Britgirl]

Yes, the “pacemaker (heartbeat)” is handled directly by nxserver. The shared IP is assigned to the host that is presently working as the master, so that the master host is working as the frontend of the cluster. In the case the master goes down (and the nxserver keeps the other host under control and checks the availability by itself), the shared IP is assigned to the slave host, so that the slave host becomes the new frontend of the cluster. The cluster product includes also load balancing functionality, but this load balancing, at the moment, is unfortunately limited, since all sessions are basically working only on the host that is master and thus only a limited set of functions can be spread across the two. This is going to change with the next version, that will be available soon, as part of an 8 update, where, after the session negotiation, connections will be “redirected” to the host that is less loaded, so providing the best load balancing. When will this be made available? W are presently working on the final tests. And this will be included in the subscription users already purchased.

[Author]

Posts