etcd isn't designed for significantly dynamic environments but we certainly want to make it easy for people to add and remove machines at runtime. We have basic support for runtime cluster configuration but are building out a more complete API. The current proposal that is being built can be found over here: http://thread.gmane.org/gmane.comp.distributed.etcd/168
As you point out there is no magic garbage collection of stale peers. You have to explicitly delete nodes that you don't expect to be participating any longer.
As you point out there is no magic garbage collection of stale peers. You have to explicitly delete nodes that you don't expect to be participating any longer.