Multiple Datanodes and gateway problem

[aabergaa]

Not sure if this is the best spot to ask a question, but I'm not on the hub much so thought I'd give it a try.

When I launch my master running the resourcemanager and namenode, followed by another container for a data node things seem alright. I then launch a second data node - with different "hostname", and the first one gets dropped as LOST. It still has its services running, but the resource manager doesn't see it anymore. I can restart services on the first and the second one will disappear. After some investigating, I tagged this to them being setup to run off the gateway when they connect to the RM? for instance, a new one will add on 192.168.56.1:50010 - the gateway address of the docker network. I start the next one, watch the logs, and it also adds and registers as a new node, also at 192.168.56.1:50010. Repeat, Repeat, Repeat whether starting new or restarting "LOST" datanodes.

Am I not starting this properly or is there an undocumented secret to get them to associate to their docker ip address rather than the docker network' gateway address?