On Wed, Aug 14, 2013 at 08:37:18AM +0200, Valerio Pachera wrote: > If I have a cluster with a gateway only and run 'cluster shutdown' > To restart the cluster is necessary to run sheep also on the gateway only > node, otherwise you get > 'Waiting for other nodes to join cluster'. > > I don't consider necessary to have the gateway only back to restart the > cluster. > May be if it was doing cache. > > What do you think? Yes, we should. But currently we lack of code to handle gateway status. For e.g, if one node is forced out as a gateway-only node, you can't see it in the node list. We need more code to take better control of gateway node handling in the later development. Hope things will get better later. Thanks Yuan |