L.s.,
We are at the beginning of a migration to Smartstore. We are reading/ following Migrate existing data on an indexer cluster to SmartStore - Splunk Documentation .
One thing is worrying to us.. bullet 7 in the "Run the migration on the indexer cluster". It says "Stop all the peer nodes."... not so nice on a production cluster.
Is this really necessary? Isn't it possible to do ithis part with the deployment from the MC? So not all peers goes down at the same time (and losing data).
Thanks in advance for the answer.
Jari
Hi @jariw,
In my experience maintenance-mode and stopping all peer nodes is not necessary.
You will already push the configuration from Cluster Manager. Before pushing new configuration, do not forget testing your new indexes.conf configuration on a standalone instance and repFactor=auto setting (as said on item 8.a).
Thanks, this is what i was hoping for.
So the manual is in this case only for safety i suppose. Already tested it on a development environment and then questioned myself why taking the peers down.
Will continue testing.
greetz
Jari
Hi @jariw,
In my experience maintenance-mode and stopping all peer nodes is not necessary.
You will already push the configuration from Cluster Manager. Before pushing new configuration, do not forget testing your new indexes.conf configuration on a standalone instance and repFactor=auto setting (as said on item 8.a).