Elasticsearch version 7.8.1edit

Also see Breaking changes in 7.8.

Known issuesedit

  • SQL: If a WHERE clause contains at least two relational operators joined by AND, of which one is a comparison (<=, <, >=, >) and another one is an inequality (!=, <>), both against literals or foldable expressions, the inequality will be ignored. The workaround is to substitute the inequality with a NOT IN operator.

    We have fixed this issue in Elasticsearch 7.10.1 and later versions. For more details, see #65488.

  • Parsing a request when the last element in an array is filtered out (for instance using _source_includes) fails. This is due to a bug in Jackson parser. Fixed in Elasticsearch 8.6.1 (#91456)
  • The deprecated index.mapper.dynamic setting can break your cluster. It can only be set using the Update index settings API. Symptoms include nodes failing to start or shards failing to allocate. Do not use this setting in versions prior to 7.17.22. The bug is fixed in 7.17.22. (issue: #109160)

Breaking changesedit

License
  • Display enterprise license as platinum in /_xpack #58217

New featuresedit

SQL

Enhancementsedit

Authorization
Features/Indices APIs
  • Change "apply create index" log level to DEBUG #56947
  • Make noop template updates be cluster state noops #57851 (issue: #57662)
  • Rename template V2 classes to ComposableTemplate #57183 (issue: #56609)
Machine Learning
  • Add exponent output aggregator to inference #58933
Snapshot/Restore
  • Allow read operations to be executed without waiting for full range to be written in cache #58728 (issues: #58164, #58477)
  • Allows SparseFileTracker to progressively execute listeners during Gap processing #58477 (issue: #58164)
  • Do not wrap CacheFile reentrant r/w locks with ReleasableLock #58244 (issue: #58164)
  • Use snapshot information to build searchable snapshot store MetadataSnapshot #56289
Transform

Bug fixesedit

CCR
  • Ensure CCR partial reads never overuse buffer #58620
Cluster Coordination
  • Suppress cluster UUID logs in 6.8/7.x upgrade #56835
  • Timeout health API on busy master #57587
Engine
  • Fix realtime get of numeric fields from translog #58121 (issue: #57462)
Features/ILM+SLM
  • Fix negative limiting with fewer PARTIAL snapshots than minimum required #58563 (issue: #58515)
Features/Indices APIs
Features/Stats
  • Fix unnecessary stats warning when swap is disabled #57983
Geo
  • Fix max-int limit for number of points reduced in geo_centroid #56370 (issue: #55992)
  • Re-enable support for array-valued geo_shape fields. #58786
Infra/Core
  • Week based parsing for ingest date processor #58597 (issue: #58479)
Machine Learning
  • Allow unran/incomplete forecasts to be deleted for stopped/failed jobs #57152 (issue: #56419)
  • Fix inference .ml-stats-write alias creation #58947 (issue: #58662)
  • Fix race condition when force stopping data frame analytics job #57680
  • Handle broken setup with state alias being an index #58999 (issue: #58482)
  • Mark forecasts for force closed/failed jobs as failed #57143 (issue: #56419)
  • Better interrupt handling during named pipe connection #1311
  • Trap potential cause of SIGFPE #1351 (issue: #1348)
  • Correct inference model definition for MSLE regression models #1375
  • Fix cause of SIGSEGV of classification and regression #1379
  • Fix restoration of change detectors after seasonality change #1391
  • Fix potential SIGSEGV when forecasting #1402 (issue: #1401)
Network
  • Close channel on handshake error with old version #56989 (issue: #54337)
Percolator
  • Fix nested document support in percolator query #58149 (issue: #52850)
Recovery
SQL
  • Fix behaviour of COUNT(DISTINCT <literal>) #56869
  • Fix bug in resolving aliases against filters #58399 (issues: #57270, #57417)
  • Fix handling of escaped chars in JDBC connection string #58429 (issue: #57927)
  • Handle MIN and MAX functions on dates in Painless scripts #57605 (issue: #57581)
Search
  • Ensure search contexts are removed on index delete #56335
  • Filter empty fields in SearchHit#toXContent #58418 (issue: #41656)
  • Fix exists query on unmapped field in query_string #58804 (issues: #55785, #58737)
  • Fix handling of terminate_after when size is 0 #58212 (issue: #57624)
  • Fix possible NPE on search phase failure #57952 (issues: #51708, #57945)
  • Handle failures with no explicit cause in async search #58319 (issues: #57925, #58311)
  • Improve error handling in async search code #57925 (issue: #58995)
  • Prevent BigInteger serialization errors in term queries #57987 (issue: #57917)
  • Submit async search to not require read privilege #58942
Snapshot/Restore
  • Fix Incorrect Snapshot Shar Status for DONE Shards in Running Snapshots #58390
  • Fix Memory Leak From Master Failover During Snapshot #58511 (issue: #56911)
  • Fix NPE in SnapshotService CS Application #58680
  • Fix Snapshot Abort Not Waiting for Data Nodes #58214
  • Remove Overly Strict Safety Mechnism in Shard Snapshot Logic #57227 (issue: #57198)
Task Management
  • Cancel persistent task recheck when no longer master #58539 (issue: #58531)
  • Ensure unregister child node if failed to register task #56254 (issues: #54312, #55875)
Transform

Upgradesedit

Infra/Core