Improve documentation for manually joining a cluster

http://www.lagomframework.com/documentation/1.2.x/java/Cluster.html#Joining-without-ConductR

Here’s the current example:

akka.cluster.seed-nodes = [
  "akka.tcp://MyService@host1:2552",
  "akka.tcp://MyService@host2:2552"]

A few notes:

• MyService needs to be the name of the actor system “it should match play.akka.actor-system, which defaults to application”
• As a shortcut for a single-node cluster, you can specify lagom.cluster.join-self = on
• It would be more helpful to integrate this into the documentation for running in production: I didn’t think to look for it here

[edouardKaiser]

Thanks Tim, yes I was talking about Java. Should have been more explicit. I thought about the constructor too of the service with the ActorSystem injection, was curious if there was a better way. All good 👍

[jousby]

Ok thanks. All sorted now ( #500 ).

If it helps with the doco fix… I was rereading the Persistent Entity documentation assuming I had missed some config option along the lines of ‘turn off sharding if using jdbc’. Might be worth putting a pointer in that section to the page on clusters as that is likely the point at which most people will strike this issue.