3 Tier Install - Points of Failure

matttrax

Obviously we can install multiple NST on a single server. That way everyone connects through the same box, but NAV balances the load accordingly (my understanding, please correct me if I am wrong). That produces a single point of failure if that server ever goes down.

For the database tier we can just create a cluster and it will fail-over. Is there a recommendation for a similar practice for the NST. We don't want to create, for example, four different servers and have one set of users log into NST01, one set into NST02, etc. Then have the users go through one by one trying to connect to each one if one doesn't work.

We'd like a single NST server with multiple instances that will fail-over to an identical NST in the event something happens. Is this possible?

kine

1) There is no "why" installing more NSTs on one server. NST is multithreaded, there is gain only if the NST needs more than 3GB RAM and you are on 64bit system.
2) There is no "automatic" load balancing. RTC client selects which server will use.
3) You can use some own module to do that: you can create own application, which will check which server to connect (somehow), write the settings into the .config file for the RTC and after that will run the RTC. You only needs to have tool how to check how many connections are currently on each NST and which NSTs are available... ;-)