Added the xpack.ml.max_model_memory_limit setting, which can be dynamically updated. For more information, see Machine learning settings.
Added checks and error messages for the ml.enabled and ml.max_open_jobs node attributes. These are reserved for internal use and their values should be set by using xpack.ml.enabled and xpack.ml.max_open_jobs machine learning settings.

Mapping

Don’t detect source’s XContentType in DocumentParser.parseDocument() #26880

Network

Add additional low-level logging handler #26887
Unwrap causes when maybe dying #26884

Plugins

Adjust SHA-512 supported format on plugin install #27093

REST

Cat shards bytes #26952

Security

Improved the error messages that are returned by the setup-passwords command.

Watcher

Added verification that the required templates exist before Watcher starts. For more information, see Scripts and templates.
Added the xpack.watcher.history.cleaner_service.enabled setting. You can use this setting to enable or disable the cleaner service, which removes previous versions of Watcher indices (for example, .watcher-history*) when it determines that they are old. For more information, see Watcher Settings.

Bug Fixes

edit

Aggregations

Create weights lazily in filter and filters aggregation #26983
Fix IndexOutOfBoundsException in histograms for NaN doubles (#26787) #26856 (issue: #26787)
Scripted_metric _agg parameter disappears if params are provided #19863 (issue: #19768)

CAT API

Fix NPE for /_cat/indices when no primary shard #26953 (issue: #26942)

Cache

Reduce the default number of cached queries. #26949 (issue: #26938)

Core

Timed runnable should delegate to abstract runnable #27095 (issue: #27069)
Stop invoking non-existent syscall #27016 (issue: #20179)
MetaData Builder doesn’t properly prevent an alias with the same name as an index #26804

Ingest

date processor should not fail if timestamp is specified as json number #26986 (issue: #26967)
date_index_name processor should not fail if timestamp is specified as json number #26910 (issue: #26890)

Internal

Upgrade Lucene to version 7.0.1 #26926

Java High Level REST Client

Make ShardSearchTarget optional when parsing ShardSearchFailure #27078 (issue: #27055)

Java REST Client

rest-client-sniffer: configurable threadfactory #26897

Machine Learning

Fixed a scenario where models were incorrectly combined. This problem occurred when anomaly detectors were considered to be the same despite having different partition field values.
Cleaned up the job closure process for situations where the job was still in the process of opening.

Mapping

wrong link target for datatype murmur3 #27143

Network

Check for closed connection while opening #26932

Packaging

Fix handling of Windows paths containing parentheses #26916 (issue: #26454)

Percolator

Also support query extraction for queries wrapped inside a ESToParentBlockJoinQuery #26754

Plugin Analysis Phonetic

Fix beidermorse phonetic token filter for unspecified languageset #27112 (issue: #26771)

Plugin Repository Azure

Use Azure upload method instead of our own implementation #26751

REST

Fix inconsistencies in the rest api specs for cat.snapshots #26996 (issues: #25737, #26923)
Fix inconsistencies in the rest api specs for *_script #26971 (issue: #26923)
exists template needs a template name #25988

Reindex API

Fix update_by_query’s default size parameter #26784 (issue: #26761)

Search

Avoid stack overflow on search phases #27069 (issue: #27042)
Fix search_after with geo distance sorting #26891
Fix serialization errors when cross cluster search goes to a single shard #26881 (issue: #26833)
Early termination with index sorting should not set terminated_early in the response #26597 (issue: #26408)
Format doc values fields. #22146

Security

Enabled PKI realms to obtain the password for the truststore from either the truststore.secure_password or the truststore.password setting. For more information, see PKI realm settings.
Fixed document level security such that if your role has authority to access a root document, you also have access to its nested documents.
Fixed an issue that caused LDAP authentication requests to be slow and to require multiple binds when authenticating in user search mode.

Sequence IDs

Fire global checkpoint sync under system context #26984

Settings

Emit settings deprecation logging on empty update #27017 (issue: #26419)
Fix filtering for ListSetting #26914

Stats

Keep cumulative elapsed scroll time in microseconds #27068 (issue: #27046)

Suggesters

Fix division by zero in phrase suggester that causes assertion to fail #27149

« Elasticsearch version 6.0.0 Elasticsearch version 6.0.0-rc1 »