Elasticsearch version 7.15.1

edit

Elasticsearch version 7.15.1

edit

Also see Breaking changes in 7.15.

Known issues

edit
  • Frozen tier: (Windows only) The frozen data tier relies on multiple caching mechanisms to speed up access to searchable snapshot files. One of these caches uses sparse files to avoid creating large files on disk when it is not strictly required. A bug prevented files from being created with the right options to enable sparse support on Windows, leading Elasticsearch to create potentially large files that can end up consuming all the disk space.

    This issue is fixed in Elasticsearch versions 7.15.2 and later. There is no known workaround for earlier versions. Filesystems that enable sparse files by default are not affected. For more details, see #79371.

Enhancements

edit
Search
  • Fail early if an index request has the wrong type #77417
Packaging
  • Use grep in Docker regexes instead of bash #78154 (issue: #77862)

Bug fixes

edit
Aggregations
  • Fix bug with significant terms background count #76730 (issue: #76729)
  • Fix aggregation memory leak for CCS #78404 (issue: #72309)
Allocation
  • Match IPv6 addresses without compressed id from settings #77404 (issue: #72091)
  • Change single node disk watermark deprecation to a warning #77166 (issue: #73733)
CompositeAggs
  • Revert the change to use segment ordinals in composite terms aggregations #78848 (issues: #74559, #78836)
Data streams
  • Properly handle wildcards in data stream deletion requests #78463 (issue: #78422)
Geo
  • Make feature factory more resilient to invalid polygons #77200 (issue: #77071)
  • Support for aggregation names with dots in vector tiles #77477 (issue: #77361)
  • Add vector tile endpoint support for source parameters in the body of requests #77377 (issues: #77264, #77365)
  • Change geo_shape mapping created before version 6.6.0 to use the legacy mapper #77881 (issue: #77857)
ILM+SLM
  • Implement DiffableStringMap's get and containsKey in terms of the wrapped innerMap #77965
Infra/Core
  • Ensure sys. index patterns are transformed before use #78492
  • AutoCreate should work for non-primary system indices #77045 (issue: #65970)
Infra/Node Lifecycle
  • Ensure Node Shutdown doesn’t stall when all nodes in the cluster have a copy of a shard #78578
  • Handle master failure in NodeSeenService #77220 (issue: #76689)
Machine Learning
  • Fix minor autoscaling edge case check #78102
Mapping
  • NullPointer fix for field capabilities API #76742 (issue: #76716)
Search
  • Add support for QueryString-parsed wildcard queries on runtime keyword fields #76867 (issue: #76838)
  • Propagate original indices in NodeTermsEnumRequest #77776 (issue: #77508)
  • Choose postings format from FieldMapper instead of MappedFieldType #77234 (issue: #77183)
Snapshot/Restore
  • Fix issue where snapshot can get stuck if queued after a delete operation where a shard is in a MISSING state #78587
  • Fix AzureBlobStore#convertStreamToByteBuffer chunking in Windows #78772
Stats
  • Change the name of the shards field in _node/stats API to shard_stats #78531 (issues: #75433, #78311)
  • Detect when a deserialized Map is immutable before changing it in IndexDiskUsageStats (#77219) #77390 (issues: #77219, #77327)
TLS
  • Fix issuerDN for created certificates #77828 (issue: #69403)

Upgrades

edit
Packaging