Apache Lucene and Solr 5.0 released
More than two years have passed since we saw the last major version of Java search engine Apache Lucene and NoSQL server Apache Solr. The long-awaited releases of version 5.0 are here.
The Apache Software Foundation have recently announced the releases of Apache Lucene 5.0 and Apache Solr 5.0, with both projects incorporating new features and major component changes.
Solr’s new version highlights
Much of the work on Solr 5.0 has focused on improvements in usability. New APIs have been added, too; however, implementation details are only displayed on a need-to-know basis. The scripts included in version 4.10 for starting, stopping and running Solr instances have been extended in their function options: We now have scripts available to enable the indexing of documents and for deleting Solr collections.
In addition, scalability and stability have been improved with the update. Prior to version 5.0, an entire cluster state was written to a single file, which was then monitored by each node and updated as required. Now, each collection comes standard with its own cluster state, which is intended to significantly increase scalability. New configuration options also allow a look “under the hood”, adding the opportunity to throttle replication bandwidth consumption.
Last but certainly not least, Solr 5.0 now supports distributed IDF, which can be enabled via the config.
With all the new stuff coming out of Solr’s latest version, we shouldn’t forget that Lucene has also been updated. Version 5.0 offers a stronger security index, reduced heap usage and automatic IO throttling in ConcurrentMergeScheduler.
FieldCache is gone,
NormsFormat now gets its own dedicated
NormsConsumer/Producer and compound file handling has been made simpler, now falling under codec control.
The list of new changes is vast, however it’s worth checking out Michael McCandless’ blog post, which explains all new features in detail.