Shipping Lucene 10.0, 25 years in the making
summary
Apache Lucene 10.0 lands after a monthslong push that surfaces specific wins and snafus: sparse indexing and measurable CPU/IO improvements, minimum Java bumped to Java 21, and compatibility rules that drop support for v8 indexes. Release work uncovers thousands of compile/test failures, 80% unit-test breakage early on, four release candidates, a focused team of ~4–5 engineers getting CI green in ~20 working days, and shipping to production in two weeks.
AI-written summary · about
chapters (14)
- 0:00Introduction and speakers
- 1:22What is Apache Lucene
- 3:30Why major releases and breaking changes
- 7:00How Lucene releases work
- 9:05Lucene release history and innovation
- 12:38Why Lucene succeeded for 25 years
- 16:12Elasticsearch integration with Lucene
- 20:10Lucene 10 release journey begins
- 22:17Upgrading Elasticsearch to Lucene 10
- 24:44Release candidate challenges
- 27:50Final hurdles to Lucene 10 release
- 29:36Community involvement and quality
- 31:02Shipping Lucene 10 to production
- 32:04Q&A and closing remarks
related talks
41mApache Lucene and Java 9+ #bbuzz
22mWhat's coming next with Apache Lucene?
40mApache Lucene 6: What's coming next? #bbuzz
43mWhat's new in Apache Solr 9.0
41mApache Lucene 5 - New Features & Improvements for Apache Solr
36mLucene Server: From Erlang to Java and Back Again
56mJDK 9 & 10: Mission Accomplished, What Next For Java
37mFast Decompression Lucene Codec #bbuzz