We used to include nodes in the bootstrapping process only based on the cluster_manager role.
We should instead only do it on the MAIN_ORCHESTRATOR and flag for every subsequent non-MAIN-ORCHESTRATOR that the cluster is already bootstrapped and that there is no need to attempt that again.
Issue
We used to include nodes in the bootstrapping process only based on the
cluster_manager
role. We should instead only do it on theMAIN_ORCHESTRATOR
and flag for every subsequent non-MAIN-ORCHESTRATOR that the cluster is already bootstrapped and that there is no need to attempt that again.