Elasticsearch version 7.17.4edit

Also see Breaking changes in 7.17.

Bug fixesedit

Cluster Coordination
  • Avoid breaking add/clear voting exclusions #86657
Geo
ILM+SLM
  • Reroute after migrating to data tiers routing #86574 (issue: #86572)
Infra/Core
  • Fix assertDefaultThreadContext enumerating allowed headers #86262
  • Relax data path deprecations from critical to warn #85952
  • Set autoexpand replicas on fleet actions data stream #85511
Infra/Settings
  • Adjust MDP deprecation to warning instead of critical #86102 (issue: #85695)
Machine Learning
  • Fix max_model_memory_limit reported by _ml/info when autoscaling is enabled #86660
  • Improve reliability of job stats in larger clusters #86305
  • Make autoscaling and task assignment use same memory staleness definition #86632 (issue: #86616)
  • Fix edge case which could cause the model bounds to inflate after detecting seasonality #2261
  • Fix cause of "Must provide points at which to evaluate function" log error training classification and regression models #2268
Packaging
  • Fix edge case where user-defined heap settings are ignored #86438 (issue: #86431)
Snapshot/Restore
  • Better failure for source-only snapshots of partially/fully mounted indices #86207
  • Check if searchable snapshots cache pre-allocation is successful in Windows #86192 (issue: #85725)
  • Delay searchable snapshot allocation during shutdown #86153 (issue: #85052)
Stats
  • Correctly calculate disk usage for frozen data tier telemetry #86580 (issue: #86055)
Task Management
  • Use TaskCancelledException in TMNA #86659

Upgradesedit

Packaging
  • Switch to OpenJDK and upgrade to 18.0.1 #86554