New features:

• Mapping Overhaul – More user friendly, cluster aware (#34)
• Cluster Health API: A simple way to know if the cluster is healthy or not (#42)
• Discovery: Support local (JVM level) discovery (#2)
• Optimize API: Allow to optimize index/indices (#9)
• Query DSL: moreLikeThis & moreLikeThisField (#43)
• Search API: Set different boost for indices when searching across indices (#23)
• Terms API: Allow to get terms for one or more field (#21)
• Transport: Support local (JVM level) transport (#3)

Enhancements:

• Query http listeners (#10)
• Merge bytebuffer and memory stores into a single memory store options (#22)
• Accept 1 / 0 as true/false (#26)
• query.sort should be an array, not an object (#30)
• Boolean Type: Support also cases when a number/string value are passed (#33)
• Count API: Also accepts /_count uri to search over all indices (#17)
• Create Mapping API: Automatically create indices (#12)
• Discovery/Jgroups: Upgrade to 2.9.0 (#7)
• Flush API: Add refresh flag (refresh after flush) (#14)
• Gateway: Store cluster meta data in JSON (and not binary) (#36)
• HTTP: Rest API should support receiving HTTP chunks (#8)
• Optimize API: Add onlyExpungeDeletes, flush and refresh parameters (#15)
• Query: support negative queries (#44)
• Query DSL:
  • Terms Filter (#1)
  • Bool query/filter to be valid JSON (#6)
  • queryString – allow to escape the string (#41)
• Search API:
  • Fields listed should exists within a “fields” Json object (#13)
  • Also accepts /_search uri to search over all indices (#16)

Bug fixes:

• flush_index hangs when no indices exist (#19)
• Facet query crashes the cluster (#28)
• JSON object properties are not positional (#27)
• Terms results differs between one node and multiple (#38)