Elasticsearch 1.0.0


See issues on GitHub

Release Notes

Upgrade notes:

We provide an extensive list of breaking changes between 0.90 and 1.0. So please make sure, you are testing your application, before moving into production.

Breaking changes:

  • Query/Get/Update APIs: Allow to control where single fields should be extracted from (source, stored fields or fielddata) #4492

New features:

  • Allow to change transport.publish_port setting to support systems as OpenShift, where public communication ports may be different than internal ones #4794

Enhancements:

  • Cluster state: Bulk process of shard started or failed should not execute on already processed events in order to prevent a possible starvation #5061
  • Cluster state: Add local flag support to all read operations that are usually executed on the master node #3345
  • REST API: Unify RestRequest.paramAsBoolean() and <code>RestRequest.paramAsBooleanOptional #4817
  • REST API: The RestRequest class now uses the same logic for all methods, which try to read boolean values from parameters #4808
  • REST cluster pending tasks API: Support master_timeout parameter #4806
  • Create Index API: Remove omit_term_freq_and_positions option for index creation #4722
  • Field data: Allow to configure circuit breaker with a percentage (based on the total heap) #4616
  • Mapping: Custom _all fields are now possible by specifying additional fields in the mapping, the data should be copied to, thus their name <code>copy_to fields #4520
  • Lucene internals: Use patched version of ReferenceManager to prevent infinite loop in <code>ReferenceManager#accquire() #5043
  • Lucene internals: Prevent possibly undefined behaviour in SearchContext due to the initial state of a <code>DocSetIterator #5049
  • Allocation: Add explanations for all AllocationDeciders #4934
  • Filtering: Do not cache a range filter that uses the now date math expressions #4846
  • Testing: Run REST tests against multiple nodes #5003
  • Testing: Add SearchType randomization to ElasticsearchIntegrationTest #4793

Bug fixes:

  • Aggregations: date_histogram now parses time zones as numeric value correctly #5057
  • Aggregations: Fixed an data-structure reuse issue, which could lead to return different counts when invoked twice in a row #5021
  • Aggregations: Added camel-case support #5009
  • Aggregations: Fix performance issues by returning correct hash codes #5004
  • Aggregagtions: DateHistogramBuilder supports <code>min_doc_count #4848
  • Query API: Parent/child queries now allow for different similarity #4977
  • Query API: multi_match query now fails if an unknown <code>type is used #4964
  • Query API: Return MatchNoDocsQuery if query string is emtpy #3952
  • Mapping: Prevent IndexOutOfBoundsException when indexing empty JSON document in <code>_all field mapping, and to compute boosts only in case of available tokens #4771
  • Mapping: Trying to enable or disable norms on a field after index creation now returns an error #4761
  • Mapping: Serialization and parsing of norms.enabled and <code>omit_norms were inconsistent #4760
  • cat API: Timestamp column regression in cat/health and <code>cat/count endpoints #4782
  • cat API: Fixed calculation of diskRation in <code>cat/allocation endpoint #4670
  • Bulk Testing: BulkRequestTests could fail on Windows due to line ending differences #4785
  • Testing: Add tracking of pages to MockPageCacheRecycler #4814
  • Plugins: Upgrading analysis plugins failed #5030
  • Allocation: Trying to allocate a shard on a non-data node caused a NullPointerException #4833
  • Hot Threads API: Failed to detect hot threads due to not-alive threads not returning needed data #4775
  • Filter: Ensure to never cache a filter that wraps a parent/child filter #4757
  • Scroll API: Scrolling with has_child filter returns correct hits on further request #4703
  • Start: Logging of immediate startup exceptions to files or standard output was reversed #4805
  • Packaging RPMs: Add correct timeout to shutdown with KILL signal to prevent unwanted hard shutdowns #5020
  • Rivers: Make sure rivers get started when index templates are configured #4577

This release also includes the changes in 1.0.0-beta1, 1.0.0-beta2, 1.0.0-rc1 and 1.0.0-rc2.