This Week in Elasticsearch and Apache Lucene - October 13th 2015
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
Elasticsearch 2.0.0-rc1 released! https://t.co/pjuyE2ouji Not long until GA...
— elastic (@elastic) October 7, 2015
Elasticsearch Core
- Java API: Deprecates
defaultRescoreWindowSize
(#13995, 2.2.0, 2.1.0) - REST: Add missing REST spec for
detect_noop
(#14004, 3.0.0, 2.2.0, 2.1.0, 2.0.0) - Mapping: Enforce metadata fields are not passed in documents (#14003, 3.0.0, 2.1.0, 2.0.0)
- Core: Don't pull translog from shadow engine (#14000, 3.0.0, 2.2.0, 2.1.0, 2.0.0)
- Packaging: Make license checks a bit less lenient (#14020, 2.0.0)
- Packaging: change permissions/ownership of config dir (#14017, 3.0.0, 2.2.0, 2.1.0, 2.0.0)
- Internal: move render search template methods to cluster admin client (#14030, 3.0.0, 2.1.0, 2.0.0)
- Cleanup: Lucene utility class. (#14042, 3.0.0, 2.2.0)
- Aggregations: Added correct generic type parameter on
ScriptedMetricBuilder
(#14018, 3.0.0, 2.2.0, 2.1.0) - Build: Remove unused
forbidden-api
file (#14052, 3.0.0) - Search Templates: Update mustache.java to version
0.9.1
(#14053, 3.0.0) - Packaging: Remove Guava as a dependency (#14054, 3.0.0)
- Internal: Fold
IndexAliasesService
into IndexService (#14044, 3.0.0) - Mapping: Ensure more specific analyzer is used independent of the mapping order (#14060, 1.7.3)
- Mapping: Deprecate
index.analysis.analyzer.default_index
in favor of index.analysis.analyzer.default (#14027, 2.0.0) - Build: Update to
maven-assembly-plugin 2.6
(#14073, 3.0.0, 2.2.0, 2.1.0) - Cleanup: Remove unused
clear(IndexReader)
method from IndexFieldData (#14072, 3.0.0, 2.2.0, 2.1.0, 2.0.0) - Engine: Never wrap searcher for internal engine operations (#14071, 2.2.0, 2.1.0, 2.0.0)
- Build: Allow tests to run from eclipse on windows (#14064, 3.0.0, 2.0.0)
- Engine: Remove
Engine.Create
(#13955, 3.0.0) - Packaging: Fix package repo path to only consist of major version (#13971, 3.0.0, 2.2.0, 2.1.0, 2.0.0)
- Internal: Remove and forbid use of
com.google.common.collect.Iterators
(#13916, 3.0.0) - Internal: Simplify similarity module and friends (#13942, 3.0.0)
- REST: Expose nodes operation timeout in REST API (#13981, 3.0.0, 2.2.0, 2.1.0, 2.0.0, 1.7.3)
- Internal: Remove and forbid use of
com.google.common.net.InetAddresses
(#13905, 3.0.0) - Internal:
TransportNodesAction
shouldn't hold to cluster state (#13948, 3.0.0, 2.2.0, 2.1.0, 2.0.0, 1.7.3) - Internal: Remove and forbid use of
com.google.common.collect.EvictingQueue
(#13903, 3.0.0) - Plugin Cloud GCE: Rename cloud-gce plugin to
discovery-gce
plugin (#13815, 3.0.0) - Core: When shard becomes active again, immediately increase its indexing buffer (#13918, 3.0.0, 2.2.0, 2.1.0)
- Tests: Remove
UpdateTests
dependency on groovy. (#13947, 3.0.0, 2.2.0) - Search: Add response into
ClearScrollResponse
(#13835, 3.0.0) - Internal: Replace Guava cache with simple concurrent LRU cache (#13879, 3.0.0)
- Core: Remove and ban
ImmutableMap
(#13939, 3.0.0) - Plugins: Parse Java system properties in
plugin.bat
(#13989, 3.0.0, 2.2.0, 2.1.0, 2.0.0) - Snapshot/Restore: Snapshot restore and index creates should keep index settings and cluster blocks in sync (#13931, 3.0.0, 2.2.0, 2.1.0)
- Packaging: Remove option to configure custom config file via
CONF_FILE or -Des.default.conf
(#13772, 3.0.0, 2.2.0, 2.1.0, 2.0.0) - Plugin Cloud GCE: Add
_gce_
network host setting (#13612, 3.0.0, 2.2.0, 2.1.0, 2.0.0) - Fielddata: Add
GeoPoint in StreamInput/StreamOutput
(#13632, 3.0.0, 2.1.0, 2.0.0) - : Rename the
RenderSearchTemplateAction
to be a cluster level action (#13791, 3.0.0, 2.1.0, 2.0.0) - Internal: Remove support for pre 2.0 indices (#13799, 3.0.0)
- : Tests: Avoid unicode strings in alternative query builder tests (#13798, 3.0.0)
- CRUD: Send response for update request when it timed out (#13101, 2.0.0)
- Plugins: Dont leave leftover files on unsuccessful installs (#12851, 3.0.0, 2.2.0, 2.1.0, 2.0.0)
- Internal: This commit removes
com.google.common.io
(#13302, 3.0.0)
Apache Lucene
- Upgrade Morfologik to version 2.0.1, to provide an explicit path to the input dictionary, preventing hacks like this
- Fix a sneaky
ClassPathException
when merging fields that had all terms filtered out by aFilterCodecReader
ComplexPhraseQuery
failed to disclose itsslop
in itstoString
methodMultiFieldQueryNodeProcessor
should use an ArrayList
not aLinkedList
and we should scrutinize all otherLinkedList
usages in LuceneBoostQuery
now adds parens around the boosted query, for the future Lucene 6.0 onlyOfflineSorter
should use Lucene'sDirectory
abstraction instead of secretly trying to consume temp directory spaceComplexPhraseQuery
still has highlighting bugs- Properly implementing the innocent sounding
equals
method is far from simple - Can we improve on Lucene's
PriorityQueue
when requesting a very large set of results? SynonymFilter
ignores itsignoreCase
argument?- Nested
SpanQuery
instances can still highlight the wrong terms - At long last we can deprecate the
Filter
class, now that its capabilities are fully folded intoQuery
and all internal usage in Lucene has been cutover - We should add a
DimensionalFormat
to Lucene's codec, to enable fast numeric and spatial searching on arbitrary byte[] - Upgrade ANTLR to version 4.5.1 for numerous bug fixes
SpanQueryParser
adds lots of new query syntax, e.g. for proximity queries on top of Lucene's classic query parserTermAutomatonQuery
should rewrite to simpler equivalent queries when possible, simplifying a future query parser integrationTermQuery
should clone the incoming term- Lucene's replication module should use
System.nanoTime
instead ofSystem.currentTimeMillis
for a better chance that the clock is monotonic
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!