Master nodes do not detect the other masters after service restart

Elastic Stack Elasticsearch
8

Thanks, that helps. Going back to my earlier post:

have discovered [] which is not a quorum; discovery will continue using [10.40.250.108:9300, 10.40.250.109:9300] from hosts providers
Either these addresses are wrong (are they?) ...

These addresses are indeed wrong:

      "transport_address": "10.40.250.108:9302",
...
      "transport_address": "10.40.250.109:9303",
...
      "transport_address": "10.40.250.107:9300",

Note the discrepancy in ports. I am guessing you're running multiple nodes on these machines and not ensuring that the masters have a fixed, known port, so it's picking the first free one above 9300 (the default behaviour) and this prevents discovery from working properly. Fix the ports of your master nodes by setting transport.port: 9300 on those nodes and transport.port: 9301-9400 on the data nodes.