I'll confirm that it's worked for me in the past, but you should always test changes like this in your lab/qa environment and not rely on some random person on the internet.
On Wed, Mar 14, 2018 at 7:13 AM, Cyril Scetbon <cyril.scet...@free.fr> wrote: > And you confirm that if we use snitches like EC2Snitch or GPFS we’ll only > have to update the seed list in Cassandra.yaml if this node is a seed ? > > — > Cyril Scetbon > > On Mar 13, 2018, at 10:08 PM, Jeff Jirsa <jji...@gmail.com> wrote: > > If you're just trying to change an IP, you can just stop the node, change > the IP and restart the node and it'll be fine (change it everywhere). > > Replacing a node is different: replacing is when a node dies, and you're > replacing it with a new node that doesnt have any data. The > -Dcassandra.replace_address option tells the starting instance it needs to > look for a dead host and get all of the data that host should have had. > > > > On Tue, Mar 13, 2018 at 6:57 PM, Cyril Scetbon <cyril.scet...@free.fr> > wrote: > >> Hey, >> >> I always thought that changing the IP address of a node requires to use >> the same procedure as for a died node, which part of it consists in >> starting Cassandra the -Dcassandra.replace_address option as indicated at >> https://docs.datastax.com/en/cassandra/3.0/cassandra/operati >> ons/opsReplaceNode.html >> >> However, it’s said at https://docs.datastax.com/en/d >> se/5.1/dse-admin/datastax_enterprise/operations/opsChangeIp.html that we >> can simply start the new node after having done some changes in >> configuration files that could be impacted (seed list in cassandra.yaml, >> cassandra-topology.properties). Is it a feature of the DSE ? Is it >> something that works with the community version ? How does it work exactly >> ? Does the replacement happen because it has the same data as the replaced >> node and something like an id is found in the local files ? The token list ? >> >> Thank you >> — >> Cyril Scetbon >> >> > >
