master_not_discovered_exception: master_not_discovered_exception / MasterNotDiscoveredException

What does it mean?

The cluster could not elect/find an active master node. Without a master, shard allocation, mapping changes and cluster state updates cannot happen; the cluster becomes unable to process operations.

Why does it happen?

A majority (quorum) of master-eligible nodes is unreachable, a network partition (split brain prevention), incorrect discovery.seed_hosts / cluster.initial_master_nodes configuration, no bootstrap on a new installation, or too few master-eligible nodes.

How to fix it

Verify that master-eligible nodes are running and can see each other: GET _cat/nodes?v and GET _cluster/state. Check the discovery.seed_hosts and cluster.initial_master_nodes settings (on first setup). Verify network/firewall and inter-node (transport 9300) connectivity. Keep enough master-eligible nodes (preferably 3) running. Inspect the logs for master election messages.