| Certainly, trying to connect before the agent creation will cause the remoting attempt to fail and exit. The reconnect flag wont cover the initial bug, as it is interesting to keep on reconnecting while the agent is still registered (but stop doing so if the agent gets removed as happened after a master restart) Is there some rule or recommendation about the procedure and order of the "agent endpoint register" and "remoting start" order? It seems weird to kick a process that will have to retry until an endpoint becomes available |
