Why there is no full load-balancing if I use replicated caches with HotRod an JDG client-server mode

Issue

• I see that there is no fair round-robin load balancing with replicated caches, how can it be configured?
• We set the client-timeout to somewhat <1sec but still see longer response times in case of node failure when getting data from the cache
• A new JDG node with replicated caches in a cluster does not gets loaded while existing nodes are