Elasticsearch 8.3.0 发布

Aggregations

• Allow serial_diff under min_doc_count aggs #86401
• Allow bucket paths to specify _count within a bucket #85720
• Fix a bug with flattened fields in terms aggregations #87392
• Fix flaky top_metrics test #86582 (issue: #86377)
• Fix: check field existence before trying to merge running stats #86926
• Fix: ordering terms aggregation on top metrics null values #85774
• Serialize interval in auto date histogram aggregation #85473

Audit

• Fix audit logging to consistently include port number in origin.address #86732
• Support removing ignore filters for audit logging #87675 (issue: #68588)

Authentication

• An authorized user can disable a user with same name but different realm #86473
• Fix clearing of lastSuccessfulAuthCache when clear all realm cache API is called #86909 (issue: #86650)

Authorization

• Fix resolution of wildcard application privileges #87293

CAT APIs

• Get hidden indices stats in GET _cat/shards #86601 (issue: #84656)

CCR

• Prevent invalid datastream metadata when CCR follows a datastream with closed indices on the follower #87076 (issue: #87048)
• Remove some blocking in CcrRepository #87235

Cluster Coordination

• Add master_timeout support to voting config exclusions APIs #86670
• Small fixes to clear voting config excls API #87828

Discovery-Plugins

• [discovery-gce] Fix initialisation of transport in FIPS mode #85817 (issue: #85803)

Distributed

• Enforce external id uniqueness during DesiredNode construction #84227

Engine

• Fork to WRITE thread before failing shard in updateCheckPoints #87458 (issue: #87094)
• Removing Blocking Wait for Close in RecoverySourceHandler #86127 (issue: #85839)

Features

• Fix autoGeneratedTimestamp should not be set externally error when retrying IndexRequest #86184 (issue: #83927)

Geo

• Fix Geotile aggregations on geo_shapes for precision 0 #87202 (issue: #87201)
• Fix null_value for array-valued geo_point fields #85959
• Guard for adding null value tags to vector tiles #87051
• Quantize geo queries to remove true negatives from search results #85441 (issue: #40891)

Highlighting

• FastVectorHighlighter should use ValueFetchers to load source data #85815 (issues: #75011, #84690, #82458, #80895)

ILM+SLM

• Make the ILM Move to Error Step Batched #85565 (issue: #81880)

Indices APIs

• Make GetIndexAction cancellable #87681

Infra/Circuit Breakers

• Make CBE message creation more robust #87881

Infra/Core

• Adjust osprobe assertion for burst cpu #86990
• Clean up DeflateCompressor after exception #87163 (issue: #87160)
• Error on direct creation of non-primary system index #86707
• Fix null message in output #86981
• Fix using FilterOutputStream without overriding bulk write #86304
• Hide system indices and their aliases in upgraded clusters #87125
• Refactor code to avoid JDK bug: JDK-8285835 #86614

Infra/Logging

• Temporarily provide SystemPropertiesPropertySource #87149

Infra/Node Lifecycle

• Upgrade folders after settings validation #87319

Infra/Plugins

• Use Windows newlines when listing plugin information on Windows #86408 (issue: #86352)

Infra/REST API

• Fix min node version before state recovery #86482

Infra/Scripting

• Allow to sort by script value using SemVer semantics #85990 (issues: #85989, #82287)
• Script: Fix setter shortcut for unbridged setters #86868
• Script: Load Whitelists as Resource #87539

Infra/Settings

• Permit removal of archived index settings #86107

Ingest

• Execute self-reference checks once per pipeline #85926 (issue: #85790)

Java Low Level REST Client

• Do not retry client requests when failing with ContentTooLargeException #87248 (issue: #86041)

License

• Consistent response for starting basic license #86272 (issue: #86244)

Machine Learning

• Fix ML task auditor exception early in cluster lifecycle #87023 (issue: #87002)
• Fix WordPiece tokenization of unknown words with known subwords #87510
• Fix distribution change check for change_point aggregation #86423
• Fixes inference timeout handling bug that throws unexpected NullPointerException #87533
• Correct logic for restart from failover fine tuning hyperparameters for training classification and regression models #2251
• Fix possible source of “x = NaN, distribution = class boost::math::normal_distribution<…​” log errors training classification and regression models #2249
• Fix some bugs affecting decision to stop optimizing hyperparameters for training classification and regression models #2259
• Fix cause of “Must provide points at which to evaluate function” log error training classification and regression models #2268
• Fix a source of “Discarding sample = nan, weights = …​” log errors for time series anomaly detection #2286

Mapping

• Don’t run include_in_parent when in copy_to context #87123 (issue: #87036)

Network

• Reject openConnection attempt while closing #86315 (issue: #86249)

Recovery

• Fail shard if STARTED after master failover #87451 (issue: #87367)

SQL

• Fix FORMAT function to comply with Microsoft SQL Server specification #86225 (issue: #66560)
• Implement binary format support for SQL clear cursor #84230 (issue: #53359)

Search

• Add status field to Multi Search Template Responses #85496 (issue: #83029)
• Fields API to allow fetching values when _source is disabled #87267 (issue: #87072)
• Fix _terms_enum on unconfigured constant_keyword #86191 (issues: #86187, #86267)
• Fix status code when open point in time without keep_alive #87011 (issue: #87003)
• Handle empty point values in DiskUsage API #87826 (issue: #87761)
• Make sure to rewrite explain query on coordinator #87013 (issue: #64281)

Security

• Make user and role name constraint consistent with max document ID #86728 (issue: #66020)
• Security plugin close releasable realms #87429 (issue: #86286)

Snapshot/Restore

• DONE should mean fully processed in snapshot status #86414
• Distinguish missing and invalid repositories #85551 (issue: #85550)
• Fork after calling getRepositoryData from StoreRecovery #87264 (issue: #87237)
• Fork after calling getRepositoryData from StoreRecovery #87254 (issue: #87237)
• Throw exception on illegal RepositoryData updates #87654
• Upgrade Azure SDK to 12.16.0 #86135

Stats

• Run TransportClusterInfoActions on MANAGEMENT pool #87679

TSDB

• TSDB: fix the time_series in order collect priority #85526
• TSDB: fix wrong initial value of tsidOrd in TimeSeriesIndexSearcher #85713 (issue: #85711)

Transform

• Fix transform _start permissions to use stored headers in the config #86802
• [Transforms] fix bug when unsetting retention policy #87711

Authentication

• Configuring a bind DN in an LDAP or Active Directory (AD) realm without a corresponding bind password is deprecated #85326 (issue: #47191)

Aggregations

• Improve min and max performance while in a random_sampler aggregation #85118

Authentication

• Support configurable claims in JWT Realm Tokens #86533
• Warn on user roles disabled due to licensing requirements for document or field level security #85393 (issue: #79207)
• TokenService decode JWTs, change warn to debug #86498

Authorization

• Add delete privilege to kibana_system for Synthetics #85844
• Authorize painless execute as index action when an index is specified #85512 (issue: #86428)
• Better error message for run-as denials #85501 (issue: #72904)
• Improve “Has Privilege” performance for boolean-only response #86685
• Relax restrictions for role names in roles API #86604 (issue: #86480)
• [Osquery] Extend kibana_system role with an access to osquery_manager… #86609

Autoscaling

• Add support for CPU ranges in desired nodes #86434

Cluster Coordination

• Block joins while applier is busy #84919
• Compute master task batch summary lazily #86210
• Log cluster.initial_master_nodes at startup #86101
• Reduce resource needs of join validation #85380 (issue: #83204)
• Report pending joins in ClusterFormationFailureHelper #85635
• Speed up map diffing (2) #86375

Discovery-Plugins

• Remove redundant jackson dependencies from discovery-azure #87898

Distributed

• Keep track of desired nodes cluster membership #84165

Engine

• Cache immutable translog lastModifiedTime #82721 (issue: #82720)
• Increase force_merge threadpool size based on the allocated processors #87082 (issue: #84943)
• More optimal forced merges when max_num_segments is greater than 1 #85065

Geo

• Support GeoJSON in CartesianPoint for point #85442
• Support geo label position as runtime field #86154
• Support geo label position through REST vector tiles API #86458 (issue: #86044)

Health

• Add a basic check for tier preference and allocation filter clashing #85071
• Add preflight checks to Health API to ensure health is obtainable #86404
• Add tier information on health api migrate tiers user actions #87486
• Health api add indicator doc links #86904 (issue: #86892)
• Health api copy editing #87010
• Return a default user action if no actions could be determined #87079

ILM+SLM

• Make the ILM and SLM history_index_enabled settings dynamic #86493

Indices APIs

• Batch execute template and pipeline cluster state operations #86017

Infra/Core

• Add mapping for tags for the elastic agent #86298
• Expand jar hell to include modules #86622
• Faster GET _cluster/settings API #86405 (issue: #82342)
• Faster string writes by saving stream flushes #86114
• Fleet: Add start_time and minimum_execution_duration attributes to actions #86167
• Force property expansion for security policy #87396
• Refactor array part into a BytesRefArray which can be serialized and … #85826
• Speed up ip v4 parser #86253
• Use varhandles for primitive type conversion in more places #85577 (issue: #78823)

Infra/Scripting

• Script: add ability to alias classes in whitelist #86899

Ingest

• Iteratively execute synchronous ingest processors #84250 (issue: #84274)
• Skip ensureNoSelfReferences check in IngestService #87337

License

• Initialize active realms without logging a message #86134 (issue: #81380)

Machine Learning

• A text categorization aggregation that works like ML categorization #80867
• Add new _infer endpoint for all supervised models and deprecate deployment infer api #86361
• Adds new question_answering NLP task for extracting answers to questions from a document #85958
• Adds start and end params to _preview and excludes cold/frozen tiers from unbounded previews #86989
• Adjust automatic JVM heap sizing for dedicated ML nodes #86399
• Replace the implementation of the categorize_text aggregation #85872
• Upgrade PyTorch to version 1.11 #2233, #2235, #2238
• Upgrade zlib to version 1.2.12 on Windows #2253
• Upgrade libxml2 to version 2.9.14 on Linux and Windows #2287
• Improve time series model stability and anomaly scoring consistency for data for which many buckets are empty #2267
• Address root cause for actual equals typical equals zero anomalies #2270
• Better handling of outliers in update immediately after detecting changes in time series #2280
• Improve normalization of anomaly detection results for short bucket lengths. This corrects bias which could cause our scoring to be too low for these jobs #2285 (issue: #2276)

Mapping

• Intern field names in Mappers #86301
• Replace BYTE_BLOCK_SIZE – 2 with indexWriter#MAX_TERM_LENGTH #85518

Network

• Log node identity at startup #85773

Search

• GeoBoundingBox query should work on bounding box with equal latitude or longitude #85788 (issue: #77717)
• Improve error message for search API url parameters #86984 (issue: #79719)

Security

• Add run-as support for OAuth2 tokens #86680
• Relax username restrictions for User APIs #86398 (issue: #86326)
• User Profile – Add hint support to SuggestProfiles API #85890
• User Profile – Add new action origin and internal user #86026
• User Profile – Support request cancellation on HTTP disconnect #86332
• User Profile – add caching for hasPrivileges check #86543

Snapshot/Restore

• Add parameter to exclude indices in a snapshot from response #86269 (issue: #82937)

Stats

• Add documentation for “io_time_in_millis” #84911

TLS

• Set serverAuth extended key usage for generated certificates and CSRs #86311 (issue: #81067)

TSDB

• Aggregation Execution Context add timestamp provider #85850

Transform

• Prefer secondary auth headers for transforms #86757
• Support range aggregation in transform #86501

Authorization

• Has privileges API for profiles #85898

Geo

• New geo_grid query to be used with geogrid aggregations #86596 (issue: #85727)

Health

• Add support for impact_areas to health impacts #85830 (issue: #85829)
• Add troubleshooting guides to shards allocation actions #87078
• Adding potential impacts to remaining health indicators #86197
• Health api drill down #85234 (issue: #84793)
• New service to keep track of the master history as seen from each node #85941
• Sorting impact index names by index priority #85347

Mapping

• Add support for dots in field names for metrics usecases #86166 (issue: #63530)
• Synthetic source #85649

SQL

• SQ: Allow partial results in SQL queries #85897 (issue: #33148)

Search

• Snapshots as simple archives #86261 (issue: #81210)

TSDB

• TSDB: Implement downsampling on time-series indices #85708 (issues: #69799, #65769)

Infra/CLI

• Upgrade procrun executables to 1.3.1 #86710

Infra/Core

• Upgrade jackson to 2.13.2 #86051

Ingest

• Upgrading to tika 2.4 #86015

Network

• Upgrade to Netty 4.1.76 #86252

Packaging

• Update Iron Bank base image to 8.6 #86796

SQL

• Update dependency – JLine – to v 3.21.0 #83767 (issue: #83575)

Search

• Update to public lucene 9.2.0 release #87162

Snapshot/Restore

• Upgrade GCS Plugin to 1.118.1 #87800

转自 Elasticsearch version 8.3.0 | Elasticsearch Guide [8.3] | Elastic