I am trying to understand high available for HornetQ, I would like to implement an architecture high availability and scalable.
I was reading the docs and the examples but I could find a way to implement a real ha and scalable system without add some other software component.
I will try to explain what I expect from my architecture.
I would like to implement some consumer and producer jms messages both in different servers.
So, the producer will have to connect to the message system (HornetQ). This is my first problem, how could I do this task scalable and high available? At the examples, the producer should always to connect to one message system, you could say to me that I should use the reconect like the example "Application layer failover", but this approach isn't scalable, because if I add other node to my hornetq cluster? How could I update my client about the new HornetQ?
I forgot to say that I would like to use stand alone mode.
I am thinking that I should use LVS, am I right? Or am I missing something? If I use LVS, will I have problem to use JNDI?
" How could I update my client about the new HornetQ?"
If you set the system as a backup, it will send a message to the live node, and the live node will update clients.
And what LVS means?