Jonas, I'd be weary of 'autoscaling' a cluster. While it seems nice in theory, rebalancing data when a node is added or removed from the cluster can be expensive. I'd recommend having a human involved, at least until you have a bit more experience administering a cluster.
To answer your question, yes, you'll have to write the code yourself to remove the node. Reid On Sep 13, 2012, at 1:30 PM, Jonas Tehler <jo...@tehler.se> wrote: > > Hi > > I'm setting up a Riak cluster on some EC2 servers running Ubuntu. The goal is > to use autoscaling to make sure that I always have the right number of nodes > online. > > When I kill one node (shut down the server, it's gone permanently) and then > start a new server to replace it (with a new name) the new server can't join > the cluster because it can't communicate with the dead node. It works if I > manually mark the dead node as down and remove it form the cluster but I > would like that to happen automatically (perhaps after a timeout). > > Do I have to write my own code that detects this situation and removes the > dead node from the cluster or is there a way to configure Riak to do this for > me? > > / Jonas > _______________________________________________ > riak-users mailing list > riak-users@lists.basho.com > http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com _______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
