Cross-Cluster Replication

Quick steps for testing cross-cluster replication

You can run a local cluster and simulate a remote cluster within a single Kibana directory.

1. Ensure Kibana isn't running so it doesn't load up any data into your cluster. Run yarn es snapshot --license=trial to install a fresh snapshot. Wait for ES to finish setting up and activate the license.
2. Create a "remote" copy of your ES snapshot by running: cp -R .es/8.0.0 .es/8.0.0-2.
3. Start your "local" cluster by running .es/8.0.0/bin/elasticsearch.
4. Run yarn start to start Kibana so that it connects to the "local" cluster.
5. Start your "remote" cluster by running .es/8.0.0-2/bin/elasticsearch -E cluster.name=europe -E transport.port=9400.
6. Index a document into your "remote" cluster by running curl -X PUT http://elastic:changeme@localhost:9201/my-leader-index --data '{"settings":{"number_of_shards":1,"soft_deletes.enabled":true}}' --header "Content-Type: application/json". Note that these settings are required for testing auto-follow pattern conflicts errors (see below).

Now you can create follower indices and auto-follow patterns to replicate the my-leader-index index on the "remote" cluster that's available at 127.0.0.1:9400.

Auto-follow pattern conflict errors

You can view conflict errors by creating two auto-follow patterns with overlapping patterns (e.g. my* and my-*) that will both capture the my-leader-index index on your remote cluster. Run the curl command to create my-leader-index2 on your remote cluster, since auto-follow patterns don't replicate existing indices.

Now, when you open the details flyout of one of the auto-follow patterns you will see a list of recent errors.