With Elasticsearch, I’m trying to set up a simple master-1 node with data-1 and data-2 nodes but when I look at the log I get this error :
I get this error http://puu.sh/FZQwv/3f31394ba7.png
All nodes are fresh CentOS 7 machines that can ping each other. For all the clusters I’ve set:
cluster.name: playground
node.name: <their name master-1/data-1/data-2>
network.host: [_local_,_site_]
http.port: 9200
discovery.seed_hosts: ["<internal IP of master>"]
cluster.initial_master_nodes: ["master-1"]
(at the end):
node.master: (only true for master-1)
node.data: true
node.ingest: false
node.ml: false
Does anyone knows what’s going on?
3
Answers
Error messages saying they have different uuid between local cluster and remote cluster.
If each node has own cluster, you can’t build a new cluster. You need to delete an existing cluster that to join a new cluster.
Follow steps:
elasticsearch.yml
.Thanks! It worked for me.
I stopped the service on all nodes and wiped the lib directory.
It seems like the lib directory has some information about the cluster state and etc.
# rm -rf /var/lib/elasticsearch/*
After that, I started the elasticsearch on all nodes and they’re joined in the cluster successfully.
I also noticed that a new cluster name was generated.
// Before wiping out the directory on master node
// After wiping out the directory on master node
I’m having exactly the same issue on my es cluster deployed in ubuntu 18.04.
However, this cmd does not work for me
Instead, I explicitly delete the nodes directory, then all my nodes can join the cluster successfully