• Feature Releases
  • OpenSearch
Cross-cluster replication for OpenSearch®

NetApp, the intelligent data infrastructure company, is announcing the launch of cross-cluster replication for OpenSearch® on the NetApp Instaclustr Managed Platform. OpenSearch is the leading open source technology for search, broadly adopted by customers that require reliable, scalable and secure solutions for managing their data.

The OpenSearch cross-cluster replication feature automatically replicates data across clusters for failover and reduced latency use cases. With this release, cross-cluster replication is available now as an enterprise add-on for all Run in Your Own Account (RIYOA) clusters on the Instaclustr Managed Platform and can be configured on all new and existing clusters.

This functionality allows NetApp customers to replicate data from an OpenSearch cluster in one region to a cluster in another region. Depending on the individual use case, this can provide a range of benefits that include:

  • In the event that a cluster is affected by an outage in one region, the customer can switch over to the secondary cluster to reduce the risk of downtime and continue to serve read requests.  
  • The ability to replicate data across geographical regions for reduced latency globally. 
  • Assisting in migrations where data can be safely and reliably replicated from one cluster to another.
  • Improve performance by having dispersed clusters replicate data to a centralised cluster that handles search queries instead of querying across the entire network. 
  • Improve indexing performance and reduce costs by directing all read queries to cheaper replicated clusters and sending all write requests to a more powerful primary cluster. 
  • Configuring a read and write cluster for low-latency applications and a separate read-only cluster for internal users who make more intensive queries without impacting the customer-facing application. 

The following considerations should be taken into account when using cross-cluster replication:

  • Replication is performed through an active-passive model, with follower clusters actively pulling data from a leader cluster for synchronization. 
  • Replicated indices on the follower cluster are strictly read-only. However, the follower cluster can continue to be used to create, read and write to indices that are not replicated from the leader. 
  • Followers can replicate from at most 1 leader cluster, while a leader cluster can be used to replicate to multiple follower clusters with a limit of 5 clusters. 
  • Replication adds additional load to the leader clusters that increases proportional to the number of follower clusters.  Benchmarking has shown that up to 3 follower clusters can be configured for replication with an impact of <10% increase to the leader cluster’s CPU usage. To limit the performance impact of replication on the leader, we recommend creating no more than 3 followers per leader. 

Cross-cluster replication is recommended for customers who have OpenSearch as a missioncritical part of their data layer and require resilience in the face of a full region outage. Data is replicated using a leader-follower model meaning that new data on the leader cluster is always actively replicated to the follower cluster.  

Of the new cross-cluster replication functionality, Ben Slater, Chief Product Officer of NetApp Instaclustr said “In the years of working with customers using NetApp Instaclustr, we know how important replication is to the resilience of critical workloads. By enabling cross-cluster replication, we are empowering our customers with the tools they need to make their applications resilient to full region outages and enabling lower latency solutions for many applications with a global user base. 

Our support documentation has a stepbystep guide on the simple steps required to set up cross cluster replication for OpenSearch from scratch. For existing customers with running OpenSearch clusters, please contact Support for assistance to add a follower cluster.   

For more information on the cross-cluster replication functionality, existing customers can contact their NetApp Customer Success representative. Otherwise, our Sales team would be happy to walk you through the details.