I have a 2-node ES cluster. I recently enabled scripting using the following:
script:
inline: true
indexed: true
I know this requires me to restart the ES service, so I decided to follow this rolling restart guide: https://www.elastic.co/guide/en/elasticsearch/guide/current/_rolling_restarts.html
So I tried to restart the ES service on the node with all the primary shards and verified that the replica shards on the other node had been promoted as primaries. The problem is the restarted node is also reporting primary shards and can't join the cluster. I'm assuming it's because a cluster can't have 2 masters.
So my question is, how did my cluster get 2 masters? Is it because of split-brain (1 node restarted too fast thinking it's still the master) or is it because the different settings on the 2 nodes making it impossible to join the same cluster?