As with any software that stores data, it is important to routinely back up your data. Elasticsearch’s replica shards provide high availability during runtime; they enable you to tolerate sporadic node loss without an interruption of service.
However, replica shards do not protect an Elasticsearch cluster from catastrophic failure. You need a backup of your cluster— a copy in case something goes wrong.
Elasticsearch offers two features to support high availability for a cluster:
- Snapshot and restore, which you can use to back up individual indices or entire clusters. You can automatically store these backups in a repository on a shared filesystem.
- Cross-cluster replication (CCR), which you can use to copy indices in remote clusters to a local cluster. You can use cross-cluster replication to recover from the failure of a primary cluster or serve data locally based on geo-proximity.