This week in Elasticsearch and Apache Lucene: 5.1.0 is released!
Welcome to This Week in Elasticsearch and Apache Lucene! With this 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.
Top News
The real deal on #elasticsearch disk capacity requirements from @acdahlqvist + @peterkimnyc: https://t.co/ol7ZaZX44o. pic.twitter.com/Tmr3xHRRTw
— elastic (@elastic) April 17, 2015
Elasticsearch Core
- Internal: Cleanup JVM info and stats (#10553, 2.0.0)
- Internal: Improve exception handling in transport local execution (#10554, 2.0.0)
- Build: Restrict read permission to
project.basedir/target
if security manager is used (#10572, 2.0.0) - Core: Rename
START
phase into <code>VERIFY_INDEX (#10570, 2.0.0) - Core: Move back to single
EngineConfig
(#10586, 2.0.0) - Cleanup: Remove unused
normsField
from <code>MatchAllQuery (#10592, 2.0.0) - Translog: Remove useless random translog directory selection (#10589, 2.0.0, 1.6.0)
- Aggregations: Fix
_as_string
output to only show when format specified (#10571, 2.0.0, 1.6.0, 1.5.2) - Tests: Forbid tests from writing to CWD (#10605, 2.0.0)
- Tests: Fix shadow engine tests (#10621, 2.0.0)
- Mappings: Same code path for dynamic mappings updates and updates coming from the API (#10593, 2.0.0)
- Build: Add tests progress indicator (#10635, 2.0.0)
- Dependencies: Upgrade to Lucene-5.2-snapshot-1674183. (#10641, 2.0.0)
- Core: Fix possible NPE in
InternalClusterService$NotifyTimeout
, the <code>future field is set from a different thread (#10630, 2.0.0, 1.6.0, 1.5.0) - Aggregations: Added context for
significant_terms
scoring (#10633, 2.0.0) - Test: Add back really old upgrade test (#10639, 2.0.0)
- Internal: Cleanup local code transport execution (#10582, 2.0.0)
- Testing: Integrate better with lucene test framework and mockfilesystems (#10656, 2.0.0)
- Search: Cut over to
IndexSearcher.count()
(#10674, 2.0.0) - Packaging: Standardization of packages structure and install (#10595, 2.0.0, 1.6.0)
- Mappings: Remove dead code after previous refactorings (#10666, 2.0.0)
- Snapshot/Restore: Improve the error message when attempting to snapshot a closed index (#10608, 2.0.0, 1.6.0)
- Internal: Ensure that explanation descriptions are not null on serialization (#10689, 2.0.0, 1.6.0)
- Core: Log sending translog operation batches to nodes (#10544, 2.0.0, 1.6.0)
- Analysis: fix ignoring tokenizer settings in
SynonymTokenFilterFactory
(#10489, 2.0.0, 1.6.0, 1.5.2) - Log only a summary line of filesystem detail for all path.data on node startup (#10527, 2.0.0, 1.6.0)
- Term Vectors: terms filtering (#9561, 2.0.0)
- MLT Query: lenient default parameters (#9412, 2.0.0)
- MLT API: Fix misbehaving user parameters (#8028, 2.0.0)
- Upgrade API: Add
upgrade_only_ancient_segments
option to upgrade API (#10540, 2.0.0, 1.6.0) - Startup: Export the hostname as environment variable (#9474, 2.0.0, 1.6.0)
- Mapping: Fixed an equality check in
StringFieldMapper
(#10359, 2.0.0, 1.6.0, 1.5.2) - Store: Move to one
data.path
per shard (#10461, 2.0.0)
Apache Lucene
- 5.1.0 is released!
FieldType
's copy constructor was failing to copy numeric precision step- Many test framework fixes: add
SuppressReproduceLine
</a>, make <a href="https://issues.apache.org/jira/browse/LUCENE-6431"><code>ExtraFS reproducible, add hashCode and equals toFilterPath
</a>, fix <a href="https://issues.apache.org/jira/browse/LUCENE-6424">other bugs in mock filesystems</a>, check two-phased iterators in <a href="https://issues.apache.org/jira/browse/LUCENE-6419"><code>AssertingQuery, add a test annotation to suppress fsync SolrSynonymParser
was failing to preserve the original token and set position length- Queries and filters now check more carefully for invalid null arguments
- Move
extractTerms
</a> from <code>Query toWeight
- Remove the dangerous
TermQuery(Term, int)
constructor MultiPhraseQuery
defers loading positions until they are neededIndexWriter
and <code>CheckIndex enforce a maximum value for token position to prevent queries from silently not finding some results- Improve testing of span queries
SpanNotQuery
gets two-phase iteratorsLongBitSet
can now handle very large bit sets correctlyBooleanQuery.extractTerms
skips terms from non-scoring clausesTermsQuery.extractTerms
should not throw an exception- Upgrade to forbidden APIs v1.8
StreamingQuadPrefixTree
should improve indexing memory and disk used in certain cases- Try to make
IOUtils.spins
a bit more robust - More baking for auto-prefix
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!