cluster breaks using zookeeper

[MarWestermann]

Hi,
I have two containers on openshift running vert.x clustered by zookeeper ( Frontend and Backend ). The frontend calls the backend over the eventbus. Everything is working as long as I first start the Backend (which listens on the eventbus) before I start the frontend (which sends to the eventbus). When I restart the backend pod the frontend is not working any more. But there is not an exception that nobody is listening on the eventbus address.

If I restart the backend I get a log in the frontend: WARNING: Error removing subs

vertx.version: 3.4.2
zookeeper.version: 3.4.6