also nodetool disablegossip to stop other nodes sending requests to the one you are about to shut down.
> I can shut down my cluster, but I don't want to have the nodes ignore > it due to some schema misoconfiguration etc when I get it up again. if you do a rolling restart the *cluster* will not lose any writes, but individual nodes will. This is by design. Hinted Handoff, Read Repair and the Consistency Level will take care of things. Also, +1 for using cassandra 1.1 Cheers ----------------- Aaron Morton Freelance Developer @aaronmorton http://www.thelastpickle.com On 29/05/2012, at 3:32 AM, Pierre Chalamet wrote: > Hi, > > Using nodetool for each node one by one: > 1. disablethrift > 2. drain > 3. Shutdown your daemon > 4. Modify the config > 5. Restart the node > > You won't lose the data on your nodes - clients might see a node down, it is > usually not a problem if your c* client is smart enough. You also won't lose > updates while performing your operations if the cl is quorum (and have enough > replicas). > > > ------Original Message------ > From: osishkin osishkin > To: user@cassandra.apache.org > ReplyTo: user@cassandra.apache.org > ReplyTo: osish...@gmail.com > Subject: commitlog_sync_batch_window_in_ms change in 0.7 > Sent: May 28, 2012 15:53 > > I'm experimenting with Cassandra 0.7 for some time now. > > I want to increase the value of commitlog_sync_batch_window_in_ms, > without losing previous data. > I can shut down my cluster, but I don't want to have the nodes ignore > it due to some schema misoconfiguration etc when I get it up again. > > I apologize if this was asked before, but I did not see a clear guide > for achieving something like this. > Can someone please help? > > Thank you > > > - Pierre
