Hello All,

I am trying to implement hibernate search for the spring based web app, we deploy our application in two different data centers. We wanted to use the hibernate search for for efficient and advanced search. As indices are stored and particular to a location in a data center, So requests coming to data center will not be in control and hence the indices update for every Create , Delete and Update operation will be particular to data center and indices will be updated for that.

Kindly share your experiences, if anyone has implemented for distributed architecture.

Thanks,
Pavan

Hi Pavan,

Hibernate Search supports multiple alternatives for such a design, but choosing the right one will need a more detailed analysis.

One option is to use the master/slave approach described here:
- https://docs.jboss.org/hibernate/stable ... #_back_end

For the backend of this solution you can use either JGroups or JMS queues. Both are very flexible so you should be able to configure them for reliable cross-datacenter replication.

You can also use Infinispan to store the index. Infinispan is a distributed data grid with support for multi-data center replication. The Infinispan team maintains a module to store Hibernate Search indexes in it.

Another option is to use the new Elasticsearch integration. This is an experimental new feature of the latest versions (it's available in version 5.6, 5.7 and 5.8) and supports delegating both index state and query operations to an Elasticsearch cluster.

HTH
Sanne