What are the pros and cons of Solr & ElasticSearch?

Question

Both Solr and ElasticSearch are built upon Lucene.

How do they compare to each other in terms of:

• Features (facet & multi-language support in particular)
• Performance
• Scalability
• Stability
• Manageability

Any experiences you have with either software that you can share?

Thanks.

Answer 1

Well, to make is short and simple:

• Use SOLR if you want to be able to fine tune your performance(by fiddling with the internals), want more control and also a huge community.

• Use elastic search if you want faster deployment, ready to live with a lesser grain of control(there are advanced options though) and get the actual output(during development) that you want to get during deployment.

Both are known to be scalable and stable and offer great performance.

PS: I have read about a person 'getting stuck' with some minor problems/bugs in elastic search. However, there are plenty that are satisfied. :D

Answer 2

I can only speak about Solr since that is what we are using.

I deployed a Solr stack on EC2 and we are handling several million records on each database. I currently have a master/slave setup and a very nice schema defined.

Besides the obvious performance benefits of Solr and the amazing queries you can perform, one thing that's often overlooked is just how easy it is to setup and learn! It took me a day to read Solr 1.4 Enterprise Search Server, and within a week I had built pretty solid Master & Slave AMI's on Amazon EC2, performed load testing, and configured my Map Reduce jobs to continuously pump data into the Solr Master.

My two cents.