Two nodes cluster - can't detect each other

Elastic Stack Elasticsearch
1

Hello,
I'm running a two nodes cluster on AWS (as EC2 instances, didn't use any community image, installed with RPM).
the nodes are 10.90.20.100 and 10.90.20.101, both are reachable to each other in ports 9200 and 9300. Yet, upon restart, the log file doesn't show any indication for mutual discovery, and _cat/nodes also shows that each formed its own cluster. I'm using version 7.5 , in version 6 this configuration worked perfectly. any idea where to look ?

configuration node 1:

cluster.name: es-cluster
node.name: elasticsearch01
network.host: 10.90.20.100
discovery.seed_hosts: ["10.90.20.101"] #i tried all combinations , and listing both and what not
cluster.initial_master_nodes: ["elasticsearch01","elasticsearch02"] #also tried all combinations

configuration node 2:

cluster.name: es-cluster
    node.name: elasticsearch02
    network.host: 10.90.20.101
    discovery.seed_hosts: ["10.90.20.100"] #i tried all combinations , and listing both and what not
    cluster.initial_master_nodes: ["elasticsearch01","elasticsearch02"] #also tried all combinations
2

Version 6 had a bug which permitted you to form two distinct clusters and merge them together later, which risked data loss. It sounds like this is what you're trying to do. This bug is fixed in version 7 so you can no longer merge separate clusters together.

There is a note in the docs describing this problem, and the solution. Does that help you?

1 Like
3

thank you David, I was overlooking that last note on the bootstrapping page, apparently the AWS image I was using had a populated data folder, which apparently prevented it from joining a cluster.
so deleting one node's data folder did the trick.

1 Like
4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.