Are there instructions on how to do it either way? If I have 4 nodes running a database and want to add a node, wouldn't I need to spin up 5 more nodes and then replicate the contents over from the first 4 nodes? That seems really intensive and painful if not using virtualized hardware.
On Fri, Mar 15, 2013 at 3:04 PM, Paul Davis <paul.joseph.da...@gmail.com>wrote: > On Fri, Mar 15, 2013 at 8:35 AM, Tim Tisdall <tisd...@gmail.com> wrote: > > I've been running a CouchDB instance for quite a while now, but only on a > > single machine. I'm looking at moving my hosting and settings things up > in > > a distributed fashion. I was hoping someone could point me in the right > > direction for what solutions are current and what solutions may actually > > make it into the core CouchDB code base (if any). > > > > I came across couchdb-lounge but as none of the code has been touched > for 3 > > years I'm guessing it's a nearly dead project. > > > > I looked at Bigcouch, but I'm a little unsure how you can modify an > > existing system (existing Bigcouch DB) to add or remove nodes. Also, is > > there any automated way of making a single CouchDB database into a > Bigcouch > > instance? > > > > -Tim > > Changing the cluster size is still a bit of an ops heavy situation. > Generally speaking the easiest method when expanding is to just create > a new database once all nodes are up and then replicate the data over. > Contracting a cluster would be similar but requires some fiddling with > zones and placements when creating the target data for replication. >
