Elasticsearch Internal File System Repository for Snapshot

Elastic Stack Elasticsearch
1

Hi All,

Is there anyone who have face this scenario regarding the snapshot using file system as repository.
Schenario:

  1. I have 3 nodes elasticsearch that all nodes have master and data role.
  2. I set up the repository with file share file system method.
  3. What I done is, create a directory /data/snapshot on each nodes, then define path.repo on elasticsearch.yml on each nodes.

The result when I try to verify repository is I got exception for non master elasticsearch node, even the test file is created.
Exception I got

"name": "ResponseError",
"message": "repository_verification_exception\n\tRoot causes:\n\t\trepository_verification_exception: [Julian_Elastic] [[we8RNUYLTdWdNMqErk5tKQ, 'org.elasticsearch.transport.RemoteTransportException: [JLNELKAPP][192.168.12.168:9300][internal:admin/repository/verify]'], [nhCrbSdRRXe8ER3HCZjR2w, 'org.elasticsearch.transport.RemoteTransportException: [JLNELKAPP2][192.168.12.165:9300][internal:admin/repository/verify]']]"
}

And if I run the snapshot it will only stored on master node, but whe the master node restarted and other node promoted as master, I need to reconfigure the repository and the last snapshot wont there because it only save on previous master.

Is there any something that I miss or not configured yet?

Thanks.

Best Regards,
Julian

2

Hello Julian,

We use the same method and it works fine for us. Did you create the directory /data/snapshot on each node? You should create it only on one of the nodes and then mount it under the same path on the other nodes(see docs):

To register a shared file system repository, first mount the file system to the same location on all master and data nodes

This is required to allow all nodes to write their data to the snapshot repository and restore data.

Best regards
Wolfram