The job scheduler will execute randomly on Node 1 or Node 2 at every moment of time. From Pega 8 you can't assign job scheduler to a specific node, only to node type. So, if you want the job to be executed in Node 1 only, you need to make this Node of some specific type (let's say Custom 1) and then choose the same node type in Job Scheduler configuration. The node type is assigned at java starting parameters when a node fires up.
You are right, it will be either Node1 or Node2. But it can be different at any time, like first execution is on Node2, second on Node1, and so on.
I am not sure about exact algorithm how Pega will choose between Node1 or Node2, I believe it will be just whichever node will pick up the task first to process. Some details from Pega manuals are below:
To run the job scheduler on only one node in a cluster, click Any one associated node.
If you configure the Job Scheduler to run on any node, the Pega Platform application chooses the node on which it processes the task.