First, my bad for a typo in my previous message, so correcting that first..
If the idea is just to control several devices which are registered to different nodes, then those can be controlled via any cti manager of a cluster, you need to connect to several nodes just for that reason, having multiple applications is good idea if you plan to have a backup jtapi application.
The above should actually be read as:
If the idea is just to control several devices which are registered to different nodes, then those can be controlled via any cti manager of a cluster,
you need not connect to several nodes just for that reason, having multiple applications is good idea if you plan to have a backup jtapi application.
Now coming to your query,
It means for one cluster I need to run one instance of our service.
So I need to create JTAPI application user in each subsriber and need to add devices of respective subscriber into JTAPI application user.
Is the CTI manager of subscriber 1 will automatically able to monitor the devices of subsber2.....subscriber5 or I need to add other subscriner's devices into each JTAPI application user.
Just to ensure we are on same page, when we say subscribers(subscriber1, subscriber2...), these are different nodes of single cluster.
Now, for JTAPI Application user to monitor a device, it must have that device in its control list, it doesnt matter whether the device is registered with subscriber1 or subscriber2 or so on.
To summarize:
If you plan to run a single instance of your application, you can run that with the provider string you specified in last message and your user needs to have all the devices it want to control in its control list, it doesnt matter if the device is actually registered to which of the subscriber.
If you still plan to run multiple JTAPI applications, I think its important to know what are we trying to achieve with it, are we trying to balance load between different application instances? or are we trying to load balance CTI Managers on different subscribers? or are we having multiple application for redundancy to ensure availability in case of app failure? IMO, the configuration, setup and application design would differ based on that.