does round robin policy cause a delay in giving new connections?
I have a 3-node vertica cluster with vertica native connection load-balancer set to "round-robin". This is version 8.0.1.x running on vmware esxi. Our application runs on a separate node on another vmware esxi.
Our app needs to establish a bunch of connections (jdbc) during start-up and start-up is seen getting delayed. If the round-robin is disabled, this would instantly work.
Are there any known issues around this ? Our app is deployed into wildfly (10.1) application server.
Could this be because the load-balancer is doing re-directs which is somehow causing the new connections to be established with a delay (some kind of tcp state becoming stale ?)