Process design for High Availability
For High Availability to work, there are also requirements for the Process design. Here are a couple of points to take into account:
In HA configuration, Agent's local G drive cannot be used. The G drive is Virtual Machine dependent and cannot be used between Agents installed in different Virtual Machines. The proper HA configuration is to use the F drive that is shared between Agents.
Shared State sharing between the Agents.
Using Remote Subprocess calls for load balancing is not considered a viable option. The latency caused by Remote Subprocess calls most likely will negate the benefits gained from HA configuration. The better option is to analyse the integration architecture and consider whether different types of Subprocesses can be assigned to specific Agent Groups in order to not mix for example interface related and heavy batch load related Subprocesses.
Whenever sharing a resource in between Agents in HA configuration takes place, special attention is to be directed to how the resource is shared to enable the best possible high availability, what the performance of the resource is and are there limitations that risk the high availability functionality.
The next article is Introduction to Multicloud Deployment