Lucene Vs Elasticsearch Indexing performance

Hi, we are designing and developing a search application which will be installed in a single system. so, we chose lucene. but elasticsearch contains lot of features and easy management. but we also need performance. so, the dilemma starts. so, we created an index in lucene and elasticsearch with similar data type mapping and tested with millions of documents. in the end we got lucene is approximately 7 to 10 times faster than elasticsearch. May be some configuration issue. but i tried monitoring both tests. one of the main impact for performance in ES seems to be translog. but data loss prevention (made by translog) in our application is not a must. so, please suggest can i use elasticsearch for my scenario?

after proper configuring 3-4 times

Did you use bulk requests when indexing into Elasticsearch?

Yes, I have used elasticsearch's bulk processor with two concurrent request and other default settings. the test involves one date_field (so, keyword analyzer, indexed,SortedNumericDocValuesField field in lucene) , one text field (so, standard analyzer and simple IndexableField in lucene) and other fields (approx 10 with uuid field) are keyword (so, keyword analyzer, indexed,SortedDocValuesField field in lucene). logs sent to index is single threaded and it is an update query with matching uuid in lucene.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.