Elasticsearch 6.1.4配置S3备份：path.repo设置及外部访问疑问

Let me walk you through setting up Elasticsearch 6.1.4 to back up your indices to S3, including the required path.repo configuration, and address your question about external access.

Step 1: Set up path.repo in elasticsearch.yml #

The path.repo parameter defines local directories on each Elasticsearch node that the cluster uses to store temporary files during backup/restore operations. Here's how to configure it:

• Open your elasticsearch.yml file (usually located in /etc/elasticsearch or your ES installation's config directory)
• Add the following line, replacing the path with a directory you want to use (ensure the path is consistent across all nodes in your cluster):

  path.repo: ["/mnt/elasticsearch_backups"]
  

• Create the directory on each node and set proper permissions for the Elasticsearch user:

  sudo mkdir -p /mnt/elasticsearch_backups
  sudo chown elasticsearch:elasticsearch /mnt/elasticsearch_backups
  

• Restart each Elasticsearch node to apply the configuration change.

Step 2: Install the S3 Repository Plugin #

Elasticsearch doesn't include S3 support out of the box, so you need to install the official plugin on every node:

bin/elasticsearch-plugin install repository-s3

After installation, restart each node again to activate the plugin.

Step 3: Register the S3 Backup Repository #

Once the plugin is installed, you can register your S3 bucket as a snapshot repository using the Elasticsearch REST API (this is the recommended method over hardcoding credentials in config):
Send a PUT request to http://<your-es-host>:9200/_snapshot/my_s3_backup_repo with the following JSON body:

{
  "type": "s3",
  "settings": {
    "bucket": "your-s3-bucket-name",
    "region": "your-aws-region", // e.g., us-east-1
    "access_key": "your-aws-access-key",
    "secret_key": "your-aws-secret-key",
    "base_path": "elasticsearch/backups" // Optional: organizes backups under a prefix in the bucket
  }
}

To verify the repository is working, send a GET request to http://<your-es-host>:9200/_snapshot/my_s3_backup_repo/_verify—you should get a response confirming all nodes can access the bucket.

Can path.repo be accessed externally? #

No, the path.repo directories are strictly local to each Elasticsearch node and cannot be directly accessed from outside the node's filesystem.

This path is only used internally by Elasticsearch to stage temporary data during snapshot operations before it's uploaded to S3. External systems or users can't reach this directory over the network—all backup/restore interactions happen via Elasticsearch's REST API, not by accessing the local path.repo directly.

内容的提问来源于stack exchange，提问作者don ali