Skip to main content

Elasticsearch and S3 Buckets (Windows)

This guide walks through how to configure Elasticsearch to store snapshots in an S3 bucket, and how to upload a snapshot to a configured S3 bucket. This guide uses the default Elasticsearch host and port (localhost:9208). If your Elasticsearch instance is running on a different host or port, replace localhost:9208 in all commands with your actual Elasticsearch address (e.g., elasticsearch.example.com:9200).

Prerequisites​

  • The access and secret key to the S3 bucket
  • Command-line access to your Elasticsearch installation directory

Step 1: Add S3 Credentials to Keystore​

[!NOTE] Replace CLIENT-NAME-HERE in the following commands with your desired client identifier (e.g., my-s3-client). Use the same name consistently throughout these steps.

Launch an elevated command prompt and navigate to the .\bin directory for Elasticsearch (by default, located in C:\Entrinsik\elasticsearch\elasticsearch-8.2.2\bin\ and run the following commands to add your S3 access key to the Elasticsearch keystore:

.\elasticsearch-keystore add s3.client.CLIENT-NAME-HERE.access_key

When prompted, paste your S3 access key and press Enter. Then add your S3 secret key:

.\elasticsearch-keystore add s3.client.CLIENT-NAME-HERE.secret_key

When prompted, paste your S3 secret key and press Enter. Restart the Elasticsearch service to load the new credentials.

Step 2: Create S3 Snapshot Repository​

Register your S3 bucket as a snapshot repository:

curl -L -X PUT "http://localhost:9208/_snapshot/REPO-NAME-HERE" \
  -H "Content-Type: application/json" \
  -d "{\"type\":\"s3\",\"settings\":{\"bucket\":\"BUCKET-NAME-HERE\",\"client\":\"CLIENT-NAME-HERE\"}}"

Parameters:

  • REPO-NAME-HERE — repository name (can be customized)
  • BUCKET-NAME-HERE — your S3 bucket name
  • CLIENT-NAME-HERE — must match the keystore client name from Step 1

Step 3: Create A Snapshot​

Create and upload a snapshot with the following command:

curl -L -X POST "http://localhost:9208/_snapshot/REPO-NAME-HERE/SNAPSHOT-NAME-HERE" \
  -H "Content-Type: application/json" \
  -d "{}"

Parameters:

  • REPO-NAME-HERE — repository name (this should be the same as the one set in the previous step)
  • SNAPSHOT-NAME-HERE — snapshot (must be unique)

This will initiate a snapshot backup to S3. Monitor progress with:

curl -L -X GET "http://localhost:9208/_snapshot/REPO-NAME-HERE/SNAPSHOT-NAME-HERE"

Verification​

List all snapshots in your repository:

curl -L -X GET "http://localhost:9208/_snapshot/REPO-NAME-HERE/_all"

The SNAPSHOT-NAME-HERE snapshot should appear in the response with a SUCCESS state.