This week in Elasticsearch and Apache Lucene: Lucene 5.1.0 has a release branch
Welcome to This Week in Elasticsearch and Apache Lucene! With this new weekly series, we're bringing you an update on all things Elasticsearch and Apache Lucene at Elastic, including the latest on commits, releases and other learning resources.
From tonight's Elasticsearch Study Session in Seoul
With @imotov talking #elasticsearch in Seoul, Korea! @elastic pic.twitter.com/lkuPcumUxR
— Nathan Zamecnik (@zagnut) April 7, 2015
Elasticsearch Core
- Networking: Remove unsafe options (#10360, 2.0.0, 1.6.0)
- Restore: Don't reuse source index UUID on restore (#10367, 2.0.0, 1.6.0, 1.5.1)
- Logging: Add ability to specify a
SizeBasedTriggeringPolicy
for log configuration... (#10373, 2.0.0, 1.6.0) - Logging: Miscellaneous additional logging and cleanups (#10376, 2.0.0, 1.6.0)
- Aggregations: Make sure
size=0
works on the <code>inner_hits level. (#10388, 2.0.0, 1.6.0, 1.5.1) - Mapping: Fix unneccesary mapping refreshes caused by unordered fielddata settings (#10370, 2.0.0, 1.6.0, 1.5.1)
- Snapshot/Restore: separate repository registration (#10354, 2.0.0, 1.6.0, 1.5.1)
- Aggregatios: Make sure inner hits also work for nested fields defined in object field (#10353, 2.0.0, 1.6.0, 1.5.1)
- BlobContainer:
AbstractBlobContainer.deleteByPrefix()
should not list all blobs (#10366, 2.0.0, 1.6.0, 1.5.1) - Dependencies: Update to Lucene 5.1 snapshot r1671277 (#10435, 2.0.0)
- Function scoring: apply
min_score
to sub query score if no function provided (#10326, 2.0.0, 1.6.0, 1.5.1) - Aggregations: Fix bug where parse error is thrown if an inner filter is used in a nested filter/query. (#10309, 2.0.0, 1.6.0, 1.5.1)
- Index API: Add check to
MetaData#concreteIndices
to prevent NPE (#10342, 2.0.0, 1.6.0) - Index expansion: Remove expansion of empty index arguments in
RoutingTable
(#10148, 2.0.0, 1.6.0) - Packging: Add properties files to configure startup and installation scripts (#10330, 2.0.0, 1.6.0)
- Tests: Skip
Version.CURRENT for
static bwc indexes (#10293, 2.0.0, 1.6.0, 1.5.1) - Mappings: Fix doc values representation to always serialize if explicitly set (#10302, 2.0.0)
- Tests: Revamp static bwc test framework to use dangling indexes (#10247, 2.0.0, 1.6.0, 1.5.1)
- Queries: Avoid NPE during query parsing (#10333, 2.0.0, 1.6.0, 1.5.1)
- HTTP: Remove
jsonp
support and associated tests (#9242, 2.0.0) - Packaging:
service.bat
file should explicitly use the Windows find command. (#9532, 2.0.0, 1.6.0, 1.5.1) - Fielddata: Add doc values support to boolean fields. (#7961, 2.0.0)
- Indexed Scripts/Templates: Return error message on 404 (#7335, 2.0.0, 1.6.0)
Elasticsearch Plugin Releases
- The Elasticsearch Groovy Client was updated to Elasticsearch 1.5
Apache Lucene
- Lucene 5.1.0 has a release branch.
- PostingsEnum's POSITIONS flag is consistent with the other flags.
- More improvements for span queries, resulting in a nice jump in Lucene's nightly performance tests (annotation BE).
- The new NRT suggester was committed.
- A faster query-time join based on global ordinals was committed.
- Force merge can require substantial (3X your index size) free temporary disk space.
- A new AutoPrefixPostingsFormat
carefully picks "good" prefix terms during indexing to make range and prefix queries fast. - SearcherFactory now also receives the old reader, making it easier to warm just the new segments on refresh.
- Workaround for a possible IndexWriter deadlock issue.
- The IndexWriter.tryDeleteDocument API is trappy.
- Don't allow overly massive positions to be indexed.
- Allow block join query to restrict the number of children to a min/max range for each matching parent document.
- Can we make Lucene's queries immutable?
- IndexWriter's delete-by-query on *:* (MatchAllDocsQuery) now forwards to IndexWriter.deleteAll instead.
Watch This Space
Stay tuned to this blog, where we'll share more news on the whole ELK ecosystem including news, learning resources and cool use cases!