Snapshot and restore

Hello,

Can someone please tell me what is the proper process to back up an elasticsearch cluster?

I have installed the curator plugin and have created a new snapshot in a registered repository for my snapshots.

curator --timeout 86400 --host <es_host> snapshot --repository <my_registered_repo> indices --all-indices

My snapshot then created is called curator-20151110123220.

I can see that this snapshot is present in the registered repository by doing a curl _XPUT 'localhost:9200/_snapshot/<my_registered_repo>/_all' because it returns snapshots which include curator-20151110123220.

I would like to upload this new snapshot to an s3 bucket which I have created.

I use the following command that should backup to an s3 bucket?

curl -XPUT http://esearch.logs.dev.aws.mycompany.service/_snapshot/<my_registered_repo> -d'

{
 "type": "s3",
 "settings": {
   "region": "",
    "bucket": "esearchbackup",
      "access_key":"xxxxxxxxxx",
      "secret_key":"xxxxxxxxxx"
   }

}'

I have ensured that my aws s3 credentials have full access and can write to the bucket.

When I run this command, should it then upload all contents of my_registered_repo to the s3 bucket called esearchbackup because when I look in this bucket I don't see my snapshot?

Does this look like the correct procedure to backup an elasticsearch cluster ? also can you tell me where on a node are the snapshots held?

Thanks

Snapshots are stored wherever you defined the repository to be.

Otherwise that looks sane, check https://www.elastic.co/guide/en/elasticsearch/guide/current/backing-up-your-cluster.html for more context :slight_smile: