In AWS, I attempted to create a node then save it as an AMI. When testing deployment, the new node would not join the cluster, as it complained "a node with a different name but duplicate ID exists".
Obviously a side affect of cloning into an AMI, but is there a "formal" way of scrubbing an elasticsearch node so when deployed it can cleanly join a cluster?
The original plan was to build a coordinating node/kibana instance then put it behind an auto-scaling group .. but currently any node past the first will have issues.