Elasticsearch version 7.10.1edit

Also see Breaking changes in 7.10.

Known issuesedit

  • In Elasticsearch 7.10.0 there were several regressions around loading nested documents. These have been addressed in Elasticsearch 7.10.2.

    • With multiple layers of nested inner_hits, we can fail to load the _source. (#66577)
    • With nested inner_hits, the fast vector highlighter may load snippets from the wrong document. (#65533)
    • When _source is disabled, we can fail load nested inner_hits and top_hits. (#66572)
  • Snapshot and restore: If an index is deleted while the cluster is concurrently taking more than one snapshot then there is a risk that one of the snapshots may never complete and also that some shard data may be lost from the repository, causing future restore operations to fail. To mitigate this problem, set snapshot.max_concurrent_operations: 1 to prevent concurrent snapshot operations:

    PUT _cluster/settings
      "persistent" : {
        "snapshot.max_concurrent_operations" : 1

    This issue is fixed in Elasticsearch versions 7.13.1 and later. It is not possible to repair a repository once it is affected by this issue, so you must restore the repository from a backup, or clear the repository by executing DELETE _snapshot/<repository name>/*, or move to a fresh repository. For more details, see #73456.

  • 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)

Bug fixesedit

  • Fix NPE in toString of FailedShard #64770
  • Stop renew retention leases when follow task fails #65168
  • Propogate rejected execution during bulk actions #64842 (issue: #64450)
Cluster Coordination
  • Allow null tiebreakers inside ordinals/sequences #65033 (issue: #64706)
  • Fix "resource not found" exception on existing EQL async search #65167 (issue: #65108)
  • Fix aggressive/incorrect until policy in sequences #65156
  • Fix SetSingleNodeAllocateStep for data tier deployments #64679
  • Watcher understands hidden expand wildcard value #65332 (issue: #65148)
  • Fix handling of null values in geo_point #65307 (issue: #65306)
  • Fix date math hidden index resolution #65236 (issue: #65157)
  • Fix Painless casting bug in compound assignment for String #65329
  • Revert null-safe behavior to error at run-time instead of compile-time #65099 (issue: #65098)
Machine Learning
  • Extract dependent variable’s mapping correctly in case of a multi-field #63813
  • Fix bug with data frame analytics classification test data sampling when using custom feature processors #64727
  • Fix custom feature processor extraction bugs around boolean fields and custom one_hot feature output order #64937
  • Protect against stack overflow while loading data frame analytics data #64947
  • Fix a bug where the peak_model_bytes value of the model_size_stats object was not restored from the anomaly detector job snapshots #1572
  • Correctly serialize search-as-you-type analyzer #65359 (issue: #65319)
  • Unused boost parameter should not throw mapping exception #64999 (issue: #64982)
  • Fix the return type problem in the sign function #64845
  • Fix cacheability of custom LongValuesSource in TermsSetQueryBuilder #65367
  • SourceValueFetcher should check all possible source fields #65375
  • Fix Broken Error Handling in CacheFile#acquire #65342 (issue: #65302)
  • Fix Two Snapshot Clone State Machine Bugs #65042