Elasticsearch version 6.4.0

edit

Elasticsearch version 6.4.0

edit

Known issue with Elasticsearch 6.4.0 and rollover

Elasticsearch can’t start if any of the shards on the node have been rolled over. (issue: #33316)

The error message looks like:

[2018-08-31T09:00:35,520][ERROR][o.e.b.Bootstrap ] Exception
org.elasticsearch.ElasticsearchException: java.io.IOException: failed to read [id:##, file:path_to_index/_state/state-##.st]
...
Caused by: org.elasticsearch.common.xcontent.XContentParseException: [-1:####] [rollover_info] failed to parse field [some_name]

bin/plugin can no longer remove x-pack

bin/plugin remove x-pack no longer works. You should remove the X-Pack plugin before upgrading to 6.4 or any later versions.

Known issues

edit

See the important sections above.

Elasticsearch 6.4.0 fails to start when PEM encoded private key files that have been exported from PKCS#12 keystores are in use. These files can be identified by the existence of lines that start with either Bag Attributes or Key Attributes in the beginning of the key file.

These keys can be converted to a supported format using one of the following methods:

  • Remove the extra lines from the beginning of the file so that the file starts with the line that starts with -----BEGIN
  • Use openssl to convert it, i.e: openssl pkcs8 -in old_format.key -topk8 -nocrypt -out new_format.key assuming the key was not password protected.

Also see 6.4.

Breaking Changes

edit
Plugins
  • Plugins: Remove meta plugins #30670
Search
  • Reject regex search if regex string is too long (#28344) #28542 (issue: #28344)
Task Management
  • Remove metadata customs that can break serialization #30945 (issues: #30731)

Breaking Java changes

edit
Authentication
Discovery-Plugins
  • Allow multiple unicast host providers #31509
Java High Level REST Client
Java Low Level REST Client

Deprecations

edit
Analysis
  • Correct spelling of AnalysisPlugin#requriesAnalysisSettings #32025
  • Deprecate nGram and edgeNGram names for ngram filters #30209
Index APIs
  • Add deprecation warning for default shards #30587
  • Deprecate not copy settings and explicitly disallow #30404 (issues: #28347)
Java High Level REST Client
  • Add high-level client methods that accept RequestOptions #31069
Java Low Level REST Client
  • Client: Deprecate many argument performRequest #30315
Mapping
  • Deprecate unindexed phrases #31072
Scripting
  • Deprecate accepting malformed requests in stored script API #28939 (issue: #27612)
Search
  • In the field capabilities API, deprecate support for providing fields in the request body. #30157
Suggesters
  • Deprecates indexing and querying a context completion field without context #30712 (issue: #29222)

New Features

edit
Aggregations
  • Add WeightedAvg metric aggregation #31037 (issue: #15731)
  • Add a MovingFunction pipeline aggregation, deprecate MovingAvg agg #29594
  • Add missing_bucket option in the composite agg #29465 (issue: #29380)
Analysis
  • Expose lucene’s RemoveDuplicatesTokenFilter #31275
  • Multiplexing token filter #31208
  • Adds a new analysis plugin called analysis_nori that exposes the Lucene Korean analysis module. (#30397)
  • Adding a char_group tokenizer #24186
Authentication
  • Add Kerberos authentication support #32263 (issue: #30243)
Authorization
  • Introduce Application Privileges with support for Kibana RBAC #32309
Java High Level REST Client
  • Add analyze API to high-level rest client #31577 (issue: #27205)
  • Add support for search templates to the high-level REST client. #30473
  • Rest High Level client: Add List Tasks #29546 (issue: #27205)
Machine learning
  • Detectors now support custom rules that enable the user to improve machine learning results by providing some domain-specific knowledge in the form of rule. #119, #31110, #31294 (issue: #31110)
  • Reverse engineer Grok patterns from categorization results #30125
Mapping
Network
Plugins
  • Reload secure settings for plugins - backport (#31383) #31481 (issue: #29135)
SQL
Scripting
  • Add more contexts to painless execute api #30511
Search
Watcher
  • Make watcher settings reloadable #31746

Enhancements

edit
Aggregations
  • Fix wrong NaN check in MovingFunctions#stdDev() #31888
  • Mitigate date histogram slowdowns with non-fixed timezones. #30534 (issue: #28727)
  • Build global ordinals terms bucket from matching ordinals #30166 (issue: #30117)
Analysis
  • Add exclusion option to keep_types token filter #32012 (issue: #29277)
  • Added lenient flag for synonym token filter #31484 (issue: #30968)
  • Consistent encoder names #29492
Audit
Authentication
  • Support RequestedAuthnContext #31238 (issue: #29995)
  • Make native realm usage stats accurate #30824
  • Limit user to single concurrent auth per realm #30794 (issue: #30355)
  • SAML: Process only signed data #30641
CRUD
  • Support for remote path in reindex api #31290 (issue: #22913)
  • Don’t swallow exceptions on replication #31179 (issue: #28571)
Circuit Breakers
  • Enhance Parent circuit breaker error message #32056
  • Split CircuitBreaker-related tests #31659
Core
Discovery-Plugins
  • Add support for AWS session tokens #30414 (issues: #16428)
Distributed
  • Avoid sending duplicate remote failed shard requests #31313
Engine
  • Adjust translog after versionType is removed in 7.0 #32020 (issue: #31945)
  • Enable engine factory to be pluggable #31183
  • Allow to trim all ops above a certain seq# with a term lower than X #30176 (issue: #10708)
  • Do not add noop from local translog to translog again #29637
Geo
  • Add support for ignore_unmapped to geo sort #31153 (issue: #28152)
Highlighting
  • Bypass highlight query terms extraction on empty fields #32090
Index APIs
Ingest
  • Extend KV Processor (#31789) #32232 (issue: #31786)
  • Make a few Processors callable by Painless #32170
  • date_index_name processor template resolution #31841
  • Introduction of a bytes processor #31733
  • Extend allowed characters for grok field names #31653, #31722 (issue: #21745)
  • Ingest: Add ignore_missing option to RemoveProc #31693 (issues: #23086)
  • Enable Templated Fieldnames in Rename #31690 (issue: #29657)
  • Add region ISO code to GeoIP Ingest plugin #31669
  • Extend allowed characters for grok field names #31653 (issue: #21745)
  • Add ingest-attachment support for per document indexed_chars limit #31352
Java High Level REST Client
  • Add Snapshots Status API to High Level Rest Client #32295, #31515
  • Add put watch action #32026, #32191 (issue: #29827)
  • Add Get Snapshots High Level REST API #31980
  • Add X-Pack usage api #31975
  • Check that client methods match API defined in the REST spec #31825
  • Clean Up Snapshot Create Rest API #31779
  • Add cluster get settings API #31706 (issue: #27205)
  • Add get index API #31703 (issues: #27205)
  • Turn GetFieldMappingsResponse to ToXContentObject #31544
  • Add Get Snapshots High Level REST API #31537 (issue: #27205)
  • Add Snapshots Status API to High Level Rest Client #31515 (issue: #27205)
  • Add get field mappings to High Level REST API Client #31423 (issue: #27205)
  • Add delete snapshot High Level REST API #31393 (issue: #27205)
  • Add explain High Level REST API #31387 (issue: #27205)
  • Add get stored script and delete stored script to high level REST API #31355 (issue: #27205)
  • Add Create Snapshot to High-Level Rest Client #31215
  • Add get index templates API #31161 (issue: #27205)
  • Add simulate pipeline API #31158 (issue: #27205)
  • Add validate query API #31077 (issue: #27205)
  • Moved pipeline APIs to ingest namespace #31027
  • List tasks failure to not lose nodeId #31001
  • Add Verify Repository High Level REST API #30934 (issue: #27205)
  • Move list tasks API under tasks namespace #30906 (issue: #29546)
  • Add get mappings support to high-level rest client #30889 (issue: #27205)
  • Fix AliasMetaData#fromXContent parsing #30866 (issue: #28799)
  • Add delete ingest pipeline API #30865 (issues: #27205)
  • Add get ingest pipeline API #30847 (issues: #27205)
  • Add MultiSearchTemplate support to High Level Rest client #30836
  • Add put ingest pipeline API #30793 (issue: #27205)
  • Add cancel task API #30745 (issue: #27205)
  • Add Delete Repository High Level REST API #30666 (issue: #27205)
  • Add synced flush API #30650 (issues: #27205)
  • Add PUT Repository High Level REST API #30501 (issue: #27205)
  • Allow caller to set per request options #30490
  • Add put index template api to high level rest client #30400 (issue: #27205)
  • Add GET Repository High Level REST API #30362 (issue: #27205)
  • Add support for field capabilities to the high-level REST client. #29664 (issue: #27205)
  • Add Cluster Health API #29331 (issue: #27205)
  • Add Get Settings API support to java high-level rest client #29229
  • Add Get Aliases API to the high-level REST client #28799 (issue: #27205)
  • Register ERR metric with NamedXContentRegistry #32320
Java Low Level REST Client
  • Node selector per client rather than per request #31471
  • NodeSelector for node attributes #31296
  • Replace Request#setHeaders with addHeader #30588
  • Preserve REST client auth despite 401 response #30558
  • Add String flavored setEntity #30447
  • Refactor Sniffer and make it testable #29638 (issues: #25701, #27697)
  • Add Request object flavored methods #29623
License
Logging
  • Add x-opaque-id to search slow logs #31539 (issue: #31521)
Machine learning
  • If a machine learning datafeed is configured to use cross-cluster search to retrieve data, the remote clusters must have X-Pack installed and a valid licence for machine learning. If the licence requirements are not met, datafeeds using cross-cluster search will not start. #31247
  • Use default request durability for .ml-state index #32233
  • Return statistics about forecasts as part of the job stats and usage API #31647 (issue: #31395)
  • Add description to ML filters #31330
  • Clean left behind model state docs #30659 (issue: #30551)
  • Hide internal job update options from the REST API #30537
  • Provide tmp storage for forecasting and possibly any machine learning native jobs #30399
  • Improves and uses periodic boundary condition for seasonal component modeling (#84)
  • Improves robustness with respect to outliers in detection and initialization of seasonal components (#90 (issue: #87))
  • Improves behavior when there are abrupt changes in the seasonal components present in a time series (#91 (issue: #6))
  • Adds explicit change point detection and modeling (#92)
  • Improves partition analysis memory usage (#97)
  • Reduces model memory by storing state for periodicity testing in a compressed format (#104, #100)
  • Improves the accuracy of model memory control (#125, #122)
  • Improves adaption of the modeling of cyclic components to very localized features (#138, #134)
  • Reduces the memory consumed by distribution models (#162, #146)
  • Forecasting of large machine learning jobs is now supported by temporarily storing model state on disk (#89)
  • Secures the machine learning processes by preventing system calls such as fork and exec. The Linux implementation uses Seccomp BPF (secure computing with Berkeley Packet Filters) to intercept system calls and is available in kernels since 3.5. On Windows, Job Objects prevent new processes being created and macOS uses the sandbox functionality (#106, #98)
  • Fixes a bug that caused underestimation of the memory used by shared pointers. Also reduces the memory consumed by unnecessary reference counting (#121, https://github.com/elastic/ml-cpp/pull/108, #115)
  • Reduces model memory by storing the state for testing predictive calendar features in a compressed format (#137, #127)
  • Always combine duplicate samples when updating population models (#74)
  • Speeds up trend model component prediction (#73)
  • Encodes distribution model weight style by offset in a fixed size weight array (#54)
Mapping
  • Remove RestGetAllMappingsAction #31129
  • Add a doc value format to binary fields. #30860 (issue: #30831)
Monitoring
  • _cluster/state should always return cluster_uuid #30143
Network
  • Backport SSL context names (#32223, #30953)
  • Remove client connections from TcpTransport #31886 (issue: #31835)
  • Support multiple system store types #31650
  • Use remote client in TransportFieldCapsAction #30838
  • Replace custom reloadable Key/TrustManager #30509
  • Derive max composite buffers from max content len #29448
Packaging
  • Set elasticsearch user to have non-existent homedir #29007 (issue: #14453)
Plugins
  • Verify signatures on official plugins #30800
Ranking
  • Rename ranking evaluation quality_level to metric_score #32168
  • Rename ranking evaluation response unknown_docs section #32166
  • Add Expected Reciprocal Rank metric #31891 (issue: #29653)
  • Add details section for dcg ranking metric #31177
  • Move templated _rank_eval tests #30679 (issue: #30628)
  • Forbid expensive query parts in ranking evaluation #30151 (issue: #29674)
Rollup
  • Rollup now indexes null values, meaning a single "unified" job for heterogeneous data is now the recommended pattern. #31402
  • Rollup Search endpoint now supports the terms query. #30973)
  • Allow rollup job creation only if cluster is X-Pack ready. #30963
  • Rollups no longer allow patterns that match its rollup_index, which can lead to strange errors. #30491
  • A new API allows getting the rollup capabilities of specific rollup indices, rather than by the target pattern. #30401
  • Validation errors thrown while creating a rollup job are now a specialization of the previous ActionRequestValidationException, which makes it easier to catch. The new exception is RollupActionRequestValidationException. #30339
  • Validate timezone in range queries to ensure they match the selected job when searching. #30338
SQL
  • Allow LEFT and RIGHT as function names #32066 (issue: #32046)
  • Add support for single parameter text manipulating functions #31874 (issue: #31604)
  • Remove restriction for single column grouping #31818 (issue: #31793)
  • Make a single JDBC driver jar #31012 (issue: #29856)
  • Remove the last remaining server dependencies from JDBC #30771 (issue: #29856)
  • Whitelist SQL utility class for better scripting #30681 (issue: #29832)
  • Improve compatibility with MS query #30516 (issue: #30398)
  • Reduce number of ranges generated for comparisons #30267 (issue: #30017)
  • Teach the CLI to ignore empty commands #30265 (issue: #30000)
  • JDBC driver prepared statement set* methods #31494 (issue: #31493)
Scripting
Search
  • Force execution of fetch tasks #31974 (issue: #29442)
  • Add second level of field collapsing #31808 (issue: #24855)
  • Remove QueryCachingPolicy#ALWAYS_CACHE #31451
  • Cross-cluster search: don’t proxy requests for already connected node #31273
  • Reject long regex in query_string #31136 (issue: #28344)
  • Cross-cluster search: do not use dedicated masters as gateways #30926 (issue: #30687)
  • Added max_expansion param to span_multi #30913 (issue: #27432)
  • Increase the maximum number of filters that may be in the cache. #30655
  • Improve explanation in rescore #30629 (issue: #28725)
Security
  • Introduce fips_mode setting and associated checks #32326, #32344
  • Tribe: Add error with secure settings copied to tribe #32298 (issue: #32117)
  • Only auto-update license signature if all nodes ready #30859 (issues: #30731)
  • Limit the scope of BouncyCastle dependency #30358
  • Make licensing FIPS-140 compliant #30251
Settings
  • Add notion of internal index settings #31286 (issue: #29823)
  • Move RestGetSettingsAction to RestToXContentListener #31101
  • Harmonize include_defaults tests #30700
  • Fold RestGetAllSettingsAction in RestGetSettingsAction #30561
Snapshot/Restore
  • ECS Task IAM profile credentials ignored in repository-s3 plugin #31864 (issues: #26913)
  • Add write*Blob option to replace existing blob #31729
  • Fixture for Minio testing #31688
  • Do not check for object existence when deleting repository index files #31680
  • Remove extra check for object existence in repository-gcs read object #31661 time an Azure object is accessed or modified #31617
  • Lazy snapshot repository initialization #31606
  • Do not check for S3 blob to exist before writing #31128
  • Remove extra checks from HdfsBlobContainer #31126
  • Allow date math for naming newly-created snapshots #30479 (issue: #7939 )
  • Use simpler write-once semantics for HDFS repository #30439
  • User proper write-once semantics for GCS repository #30438
  • Use stronger write-once semantics for Azure repository #30437
  • Use simpler write-once semantics for FS repository #30435
  • Do not fail snapshot when deleting a missing snapshotted file #30332 (issue: #28322)
  • Repository GCS plugin new client library #30168 (issue: #29259)
  • Fail snapshot operations early on repository corruption #30140 (issues: #29649)
  • Index name added to snapshot restore exception #29604 (issue: #27601)
  • Do not load global state when deleting a snapshot #29278 (issue: #28934)
  • Don’t load global state when only restoring indices #29239 (issue: #28934)
Stats
  • Add _coordinating_only for nodes resolving in nodes API #30313 (issue: #28831)
Store
  • Move caching of the size of a directory to StoreDirectory. #30581
Suggesters
Task Management
  • Make Persistent Tasks implementations version and feature aware #31045 (issues: #30731)
Transport API
  • Implemented XContent serialisation for GetIndexResponse #31675
  • Send client headers from TransportClient #30803
  • Modify state of VerifyRepositoryResponse for backwards compatibility #30762
Watcher
  • Clean up ensureWatchExists use #31926
  • Store username on watch execution #31873 (issue: #31772)
  • Consolidate setting update registration #31762
  • Add secure setting for watcher email password #31620
  • Slack message empty text #31596 (issue: #30071)
  • Move watcher-history version setting to _meta field #30832 (issue: #30731)
  • Only allow x-pack metadata if all nodes are ready #30743 (issues: #30731)
  • Configure HttpClient parallel sent requests #30130
  • Watcher: Make start/stop cycle more predictable and synchronous #30118
ZenDiscovery
  • Preserve response headers on cluster update task #31421 (issues: #31408)
  • Treat ack timeout more like a publish timeout #31303
  • Use system context for cluster state update tasks #31241 (issue: #30603)

Bug Fixes

edit
Aggregations
  • Fix profiling of ordered terms aggs #31814 (issue: #22123)
  • Ensure that ip_range aggregations always return bucket keys. #30701 (issue: #21045)
  • Fix class cast exception in BucketMetricsPipeline path traversal #30632 (issue: #30608)
  • Fix NPE when CumulativeSum agg encounters null value/empty bucket #29641 (issue: #27544)
Allocation
  • A replica can be promoted and started in one cluster state update #32042
  • Ignore numeric shard count if waiting for ALL #31265 (issue: #31151)
  • Move allocation awareness attributes to list setting #30626 (issue: #30617)
  • Auto-expand replicas when adding or removing nodes #30423 (issue: #1873)
  • Auto-expand replicas only after failing nodes #30553
Analysis
  • Call setReferences() on custom referring tokenfilters in _analyze #32157 (issue: #32154)
Audit
  • Fix audit index template upgrade loop #30779
Authentication
  • [Kerberos] Add debug log statement for exceptions #32663
  • Remove Kerberos bootstrap checks #32451
  • Fix building AD URL from domain name #31849
  • resolveHasher defaults to NOOP #31723 (issues: #31697)
  • Check auth scheme case insensitively #31490 (issue: #31486)
  • Fix joining cluster with production license #31341 (issue: #31332)
  • Fix token backwards compatibility with pre 6.0.0-beta2 #31254 (issues: #31195)
  • Compliant SAML Response destination check #31175
  • Clean up code in file stores #30348
  • Fix TokenMetaData equals and hashcode #30347
Authorization
  • Fix role query that can match nested documents #32705
  • Make get all application privileges require "*" permission #32460
  • Revert to old way of merging automata #32254
  • [PKI Realm] Invalidate cache on role mappings change #31510
  • Fix dynamic mapping updates with aliases #30787 (issue: #30597)
  • Include an empty JSON object in a JSON array when FLS filters out all fields #30709 (issue: #30624)
  • Reduce garbage during index resolution #30180
CRUD
  • Bulk operation fail to replicate operations when a mapping update times out #30244
Core
  • Fix content type detection with leading whitespace #32632 (issue: #32357)
  • Disable C2 from using AVX-512 on JDK 10 #32138 (issue: #31425)
  • Create default ES_TMPDIR on Windows #30325
  • Pick inner most parse exception as root cause #30270 (issues: issue}30261[#30261])
Distributed
  • Fix race between replica reset and primary promotion #32442 (issues: #32118, #32304, #32431)
  • ClassCastException when re-throwing "shard not available" exception in TransportShardMultiGetAction #32185 (issue: #32173)
Engine
  • Fail shard if IndexShard#storeStats runs into an IOException #32241 (issue: #29008)
  • IndexShard should not return null stats #31528
Geo
  • Fix handling of points_only with term strategy in geo_shape #31766 (issue: #31707)
  • Fix coerce validation_method in GeoBoundingBoxQueryBuilder #31747 (issue: #31718)
  • Improve robustness of geo shape parser for malformed shapes #31449 (issue: #31428)
  • Fix defaults in GeoShapeFieldMapper output #31302 (issue: #23206)
  • Add support for indexed shape routing in geo_shape query #30760 (issue: #7663)
  • Add validation that geohashes are not empty and don’t contain unsupported characters #30376 (issue: #23579)
Index APIs
  • Copy missing segment attributes in getSegmentInfo #32396
  • Add support for is_write_index in put-alias body parsing #31674
  • Fix writeIndex evaluation for aliases #31562
  • Fix IndexTemplateMetaData parsing from xContent #30917
  • Do not ignore request analysis/similarity settings on index resize operations when the source index already contains such settings. #30216
  • Do not return all indices if a specific alias is requested via get aliases api. #29538 (issues: #27763)
Ingest
  • Fix broken backport of #31578 by adjusting constructor #31587 (issue: #31578)
  • Ingest Attachment: Upgrade Tika to 1.18 #31252
  • Interrupt the current thread if evaluation grok expressions take too long #31024 (issue: #28731)
Java High Level REST Client
  • Ban LoggingDeprecationHandler #32756 (issue: #32151)
  • Move commercial clients from XPackClient #32596
  • Fix CreateSnapshotRequestTests Failure #31630 (issue: #31625)
  • Change bulk’s retry condition to be based on RestStatus #29329 (issues: #28885)
Java Low Level REST Client
  • Avoid setting connection request timeout #30384 (issue: #24069)
License
  • Do not serialize basic license expiration in X-Pack info #30848
Machine learning
  • Move open job failure explanation out of root cause #31925 (issue: #29950)
  • Fix calendar and filter updates from non-master nodes #31804 (issue: #31803)
  • Don’t treat stale FAILED jobs as OPENING in job allocation #31800 (issue: #31794)
  • Rate limit established model memory updates #31768
  • Account for gaps in data counts after job is reopened #30294 (issue: #30080)
  • Ages seasonal components in proportion to the fraction of values with which they’re updated (#88 (issue: #87))
  • Fixes persist and restore, which were missing some of the trend model state. (#103, #99)
  • Stops zero variance data from generating a log error in the forecast confidence interval calculation (#120, #107)
  • Fixes corner case which was failing to calculate lgamma values and fixes the corresponding log errors (#131, #126)
  • Fixes influence count per bucket for metric population analyses, which was wrong and lead to incorrect influencer scoring (#153, #150)
  • Fixes a possible SIGSEGV for jobs with multivariate by fields enabled, which caused the jobs to fail (#174, #170)
  • Corrects the model bounds and typical value calculation for time series models which use a multimodal distribution. This issue could cause "Unable to bracket left percentile =…​" errors to appear in the logs. (#178, #176)
Mapping
  • Make sure that field collapsing supports field aliases. #32648 (issue: #32623)
  • Improve the error message when an index is incompatible with field aliases. #32482
  • Make sure that field aliases count towards the total fields limit. #32222
  • Fix range queries on _type field for singe type indices (#31756) #32161, #31756 (issues: #31476)
  • In NumberFieldType equals and hashCode, make sure that NumberType is taken into account. #31514
  • Get Mapping API to honour allow_no_indices and ignore_unavailable #31507 (issue: #31485)
  • Make sure KeywordFieldMapper#clone preserves split_queries_on_whitespace. #31049
  • Delay _uid field data deprecation warning #30651 (issue: #30625)
Monitoring
  • Fix _cluster/state to always return cluster_uuid #30656
Network
  • Ensure we don’t use a remote profile if cluster name matches #31331 (issue: #29321)
  • Transport client: Don’t validate node in handshake (#30737) #31080 #30737 (issue: #30141)
  • Add TRACE, CONNECT, and PATCH http methods #31079, #31035 (issue: #31017)
Packaging
  • Add temporary directory cleanup workarounds #32615 (issue: #31732)
  • Add package pre-install check for java binary #31343 (issue: #29665)
  • Do not run sysctl for vm.max_map_count when its already set #31285
  • Stable filemode for zip distributions #30854 (issue: #30799)
  • Force stable file modes for built packages #30823 (issue: #30799)
Plugins
  • Template upgrades should happen in a system context #30621 (issue: #30603)
REST API
  • Reject forcemerge requests with a body #30792 (issue: #29584)
  • Respect accept header on no handler #30383 (issue: #30329)
Recovery
  • IndicesClusterStateService should replace an init. replica with an init. primary with the same aId #32374 (issue: #32308)
  • Ensure to release translog snapshot in primary-replica resync #32045 (issue: #32030)
  • Fix missing historyUUID in peer recovery when rolling upgrade 5.x to 6.3 #31506 (issue: #31482)
  • Cancelling a peer recovery on the source can leak a primary permit #30318
  • ReplicationTracker.markAllocationIdAsInSync may hang if allocation is cancelled #30316
  • Do not log warn shard not-available exception in replication #30205
Rollup
  • Move to 128bit document IDs for Rollup. The old IDs were not wide enough and susceptible to hashing collisions. Jobs that are running during cluster upgrade will "self-upgrade" to the new ID scheme, but it is recommended that users fully rebuild Rollup indices from scratch if possible. Any existing collisions are not fixable and so data-loss may affect the rollup index despite the new IDs being used. #32558 (issue: #32372)
  • Histo group configurations should support scaled_float #32048 (issue: #32035)
  • Fix rollup on date fields that don’t support epoch_millis #31890
  • Metric config properly validates itself now #31159
SQL
  • HAVING clause should accept only aggregates #31872 (issue: #31726)
  • Check timeZone argument in AbstractSqlQueryRequest #31822
  • Fix incorrect HAVING equality #31820 (issue: #31796)
  • Fix incorrect message for aliases #31792 (issue: #31611)
  • Allow long literals #31777 (issue: #31750)
  • Fix stackoverflow on getObject and timestamp conversion #31735 (issue: #31734)
  • Fix rest endpoint names in node stats #31371
  • Preserve scoring in bool queries #30730 (issue: #29685)
  • Verify GROUP BY ordering on grouped columns #30585 (issue: #29900)
  • SYS TABLES ordered according to *DBC specs #30530
  • Fix parsing of dates with milliseconds #30419 (issue: #30002)
  • Improve correctness of SYS COLUMNS & TYPES #30418 (issue: #30386)
  • Fix bug caused by empty composites #30343 (issue: #30292)
  • Correct error message #30138 (issue: #30016)
  • Add BinaryMathProcessor to named writeables list #30127 (issue: #30014)
Scripting
  • Painless: Fix Bug with Duplicate PainlessClasses #32110
  • Painless: Fix bug for static method calls on interfaces #31348
  • Deprecate Empty Templates #30194
Search
  • Fix multi level nested sort #32204 (issues: #31554, #31783, #32130)
  • Fix race in clear scroll #31259
  • Fix index prefixes to work with span_multi #31066 (issue: #31056)
  • Cross-cluster search: preserve remote status code #30976 (issue: #27461)
  • Fix NPE in more_like_this when field has zero tokens #30365 (issue: #30148)
  • Fix failure for validate API on a terms query #30319, #29483 (issue: #29033)
  • Fix a bug in FieldCapabilitiesRequest#equals and hashCode. #30181
  • Fix TermsSetQueryBuilder.doEquals() method #29629 (issue: #29620)
  • Add additional shards routing info in ShardSearchRequest #29533 (issue: #27550)
  • Use date format in date_range mapping before fallback to default #29310 (issue: #29282)
Security
  • Enable FIPS140LicenseBootstrapCheck #32903
  • Detect old trial licenses and mimic behaviour #32209
  • Preserve thread context when connecting to remote cluster #31574 (issues: #31462)
Snapshot/Restore
  • Fix repository update with the same settings but different type #31458
  • Delete temporary blobs before creating index file #30528 (issues: #30507)
Store
Suggesters
  • Add proper longitude validation in geo_polygon_query #30497 (issue: #30488)
  • Fix merging logic of Suggester Options #29514
Transport API
  • Fix interoperability with < 6.3 transport clients #30971 (issue: #30731)
  • Fix bad version check writing Repository nodes #30846 (issue: #30807)
Watcher
  • Guard against null in email admin watches #32923 (issue: #32590)
  • Fix null failure in watcher test #31968 (issue: #31948)
  • Fix chain input toXcontent serialization #31721
  • Add ssl.trust email account setting #31684
  • Fix check for currently executed watches #31137
  • Prevent duplicate watch triggering during upgrade #30643
  • Prevent triggering watch when using activate API #30613
  • Ensure trigger service pauses execution #30363
  • Fix watch history template for dynamic slack attachments #30172
  • Ensure mail message ids are unique per watch action #30112
  • Validate xContentType in PutWatchRequest. #31088 (issue: #30057)
ZenDiscovery
  • Fsync state file before exposing it #30929
  • Use correct cluster state version for node fault detection #30810
  • Only ack cluster state updates successfully applied on all nodes #30672

Regressions

edit
Engine
  • Give the engine the whole index buffer size on init. #31105
Snapshot/Restore
  • S3 repo plugin populate SettingsFilter #30652

Upgrades

edit
Core
  • Dependencies: Upgrade to joda time 2.10 #32160
Logging
Network
Search
  • Upgrade to Lucene 7.4.0. #31529