On Fri, Oct 2, 2015 at 9:08 AM, Stephen Baynes < [email protected]> wrote:
> Will need to do some thinking about this. I wonder about shiping a backup > of a good system keyspace and restore it on each node before it starts for > the first time - but will that end up with each node having the same > internal id? > (As a general statement, automated multi-threaded creation of schema in a distributed database is a risky activity I do not advise... single thread helps a lot...) If you do a rolling restart, it will pick one of the two names. Unfortunately it will then leave the SSTables from the loser name where they are. It's unclear whether restart with join_ring=false is enough to converge the schemas... But if so, you'd want to : 1) rolling restart nodes with join_ring=false so that their schema picks the winner 2) stop the node and move the SSTables from the loser directory into the winner directory (named appropriately, etc., watch out for name collision etc.) 3) start the node with join_ring=false again 4) repair it 5) instruct the node to join the ring (or just restart it one more time) =Rob
