I'm using v4.1.3 of selenium grid and trying to figure out how the load balancing works. Right now I have two nodes, one is an older Windows desktop with capacity for 8 browser sessions and one is a newer Windows desktop with capacity for 12.
If I start execution of test cases one by one until all 20 slots are in use, the first 8 start up using the node running on the slower Windows desktop, and after that it starts sending them to the faster desktop until it's maxxed out.
I have the hub and a node running on the faster desktop, and a node running on the slower desktop, pointed to the hub on the faster machine.
Is this how it's supposed to distribute test cases? I expected the hub to distribute one test case to the slow machine, then one to the fast machine, then another to the slow machine, etc until all of the slots are used.