Elasticsearch 0.90.10

See issues on GitHub

Release Notes

Breaking changes

  • Stats/Infos API: JvmStats now have standard names for gc and memory pools #4661
  • Cluster Stats API: Expose min/max file descriptors #4681

New features

  • Scripting: Term statistics are now accessible in scripts #3772, #4584
  • Allocation: Added new NodeVersionAllocationDecider, which ensures that relocation is forward compatible from a versioning point of view. This is important when doing rolling upgrades #4588


  • Mapping: Allow omit_norms for _all field #3734
  • Warmers: Allow specifying dedicated norms/terms as warming options in mapping #4079
  • Create index API now uses the new acknowledgement mechanism #4421
  • Parsing of ByteSizeValue like 12gb is now case-independent #4442
  • Add support for TB and PB in ByteSizeValue #4640
  • Aliases API: Allow IndicesAdminClient.getAliases() to return all aliases #4455
  • TransportClient: Expose list of filtered nodes #4571
  • Geo distance calculations: Use haversine for fast and accurate distance measurement #4596
  • Plugin manager: New timeout option, which prevents waiting indefinitely when hanging #4603
  • HTTP layer: Expose headers on HttpRequest #4609
  • Highlighting: FastVectorHighlighter now uses phrase limit to prevent excessive resource wasting on memory and time to highlight with documents that contains lots of matches #4546
  • Build: Shading the joda jars as well to prevent scala compilation problem #4660
  • Query API: Expose flags in simple_query_string query #4490

Bug fixes

  • Multi data path config could cause a shard to be perceived as corrupted #4674
  • Highlighting: Fixed FastVectorHighlighter throwing away some query boosts #4351
  • Query API: Named filter and query now work with parent/child queries #4534
  • Mapping: GeoPointFieldMapper.doXContentBody now honors includeDefaults #4563
  • Mapping: The _all field could loose configuration on serialization #4579
  • Mapping: The _all field used the wrong setting name for storing term vectors #4581
  • Mapping: Specific Term Vector settings could conflict with each other #4582
  • Check if node is still present when collecting attribute shard routings for that specific node #4589
  • Regex: Double wildcards in the the index name could cause a request to hang #4610
  • Allocation: BalancedShardAllocator could trigger an unnecessary relocation under rare circumstances #4630
  • Bulk API: Disabling allow_explicit_index breaks bulk #4668