| c1310c4 | Armin Braun | 09 November 2022, 16:13:28 UTC | Fix corrupted Metadata from index and alias having the same name (#91456) (#91469) This fixes a bug introduced in https://github.com/elastic/elasticsearch/pull/87863 which added a Metadata copy constructor with separate name collision checks that assumed index name and alias names were already validated in `IndexMetada`. => fixed corrupted states by actually adding the validation to `IndexMetadata` to make it impossible to instantiate a broken `IndexMetadata` in the first place. | 09 November 2022, 16:13:28 UTC |
| a1ed654 | Keith Massey | 09 November 2022, 15:31:20 UTC | Adding more detail about ingest.geoip.downloader.endpoint (#91182) (#91468) | 09 November 2022, 15:31:20 UTC |
| 784d892 | Abdon Pijpelink | 09 November 2022, 15:14:22 UTC | Added reference to terms_set query in regular terms query documentation (#91204) (#91466) * Added reference to terms_set query in regular terms query documentation * Update docs/reference/query-dsl/terms-query.asciidoc Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> Co-authored-by: Etki <etki@ayte.team> | 09 November 2022, 15:14:22 UTC |
| 51fe478 | István Zoltán Szabó | 09 November 2022, 08:52:07 UTC | [8.5][DOCS] Adds ML-related highlights to 8.5 (#90506) Co-authored-by: David Roberts <dave.roberts@elastic.co> Co-authored-by: Lisa Cawley <lcawley@elastic.co> | 09 November 2022, 08:52:07 UTC |
| 7412932 | Mark Vieira | 08 November 2022, 19:59:51 UTC | Use runtime java home for third party audit task (#91412) (#91432) We actually do indeed need to use the runtime Java for this task because depending on the target version, we may be loading classes for a later Java version. For example, if we use dependencies that are MR jars, then we in fact do load different classes depending on the runtime Java version. To test this properly then we need to use the target Java version to actually execute the task so it can both load the newer classes, and any classes from any newer APIs present in the later Java version. | 08 November 2022, 19:59:51 UTC |
| 4ad150c | Nikola Grcevski | 08 November 2022, 18:56:43 UTC | Backport 91297 to 8.5 (#91429) | 08 November 2022, 18:56:43 UTC |
| 1eeee3b | Abdon Pijpelink | 08 November 2022, 18:19:30 UTC | [DOCS] Clarify that lookup runtime sub-fields can't be used in queries and aggs (#91410) (#91426) | 08 November 2022, 18:19:30 UTC |
| 75b754e | debadair | 08 November 2022, 16:38:51 UTC | [DOC] Update CCS version matrix (#91371) (#91419) * [DOC] Update CCS version matrix * Add two extra columns to table definition Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> | 08 November 2022, 16:38:51 UTC |
| d945bd1 | Julie Tibshirani | 08 November 2022, 15:46:01 UTC | Link to kNN search guide in dense_vector docs (#91372) Before it linked to script_score and approximate kNN separately, but now we have a single page that describes both approaches. This change also removes a link to the deprecated _knn_search API. | 08 November 2022, 16:23:34 UTC |
| 895d1f1 | David Kilfoyle | 08 November 2022, 15:16:53 UTC | [DOCS] Add downsampling reference to rollup docs (#91295) (#91411) | 08 November 2022, 15:16:53 UTC |
| a38dd9e | Luigi Dell'Aquila | 08 November 2022, 15:12:15 UTC | EQL: Fix EqlSearchRequest serialization (bwc) (#91402) (#91406) | 08 November 2022, 15:12:15 UTC |
| 175d6bd | Abdon Pijpelink | 08 November 2022, 13:30:39 UTC | Fix confusion in runtime_mapping (#90999) (#91403) Co-authored-by: saryeHaddadi <51054901+saryeHaddadi@users.noreply.github.com> | 08 November 2022, 13:30:39 UTC |
| 5cd0392 | Iraklis Psaroudakis | 08 November 2022, 13:11:58 UTC | [CI] Mute reference/cluster/nodes-stats (#91399) (#91401) relates #91081 | 08 November 2022, 13:11:58 UTC |
| 5013a52 | Przemysław Witek | 08 November 2022, 11:05:50 UTC | [Transform] Make transform `_preview` request cancellable (#91313) (#91388) | 08 November 2022, 11:05:50 UTC |
| c8b4a1f | Iraklis Psaroudakis | 08 November 2022, 10:54:41 UTC | [CI] Mute reference/cluster/nodes-stats (#91390) relates #91081 | 08 November 2022, 10:54:41 UTC |
| 0c3baab | Mark Vieira | 07 November 2022, 19:15:50 UTC | [8.5] Update bundled JDK to Java 19.0.1 (#91025) (#91030) | 07 November 2022, 19:15:50 UTC |
| eaf20c1 | Mark Vieira | 07 November 2022, 16:50:44 UTC | Revert "Upgrade Gradle Enterprise plugin to 3.11.3 (#91323) (#91326)" This reverts commit 8ad1454e9af01878221ef073a6e5593f52081b0c. | 07 November 2022, 16:50:44 UTC |
| 149e447 | Mark Vieira | 04 November 2022, 22:17:46 UTC | Fix build scan publication after Gradle Enterprise plugin upgrade | 07 November 2022, 16:22:16 UTC |
| 73dd560 | David Kyle | 07 November 2022, 10:40:38 UTC | [ML] Do not count interim buckets towards the total bucket count (#91288) (#91341) | 07 November 2022, 10:40:38 UTC |
| 37f1518 | Lisa Cawley | 05 November 2022, 01:51:19 UTC | [DOCS] Remove coming tag from release notes (#91330) (#91331) | 05 November 2022, 01:51:19 UTC |
| f019865 | Lisa Cawley | 05 November 2022, 01:16:03 UTC | [8.5] [DOCS] Clarify description of geo_results (#91237) (#91318) * [DOCS] Clarify description of geo_results (#91237) * [DOCS] Fix formatting | 05 November 2022, 01:16:03 UTC |
| 8ad1454 | Mark Vieira | 04 November 2022, 18:55:21 UTC | Upgrade Gradle Enterprise plugin to 3.11.3 (#91323) (#91326) | 04 November 2022, 18:55:21 UTC |
| 991fdde | Tanguy Leroux | 04 November 2022, 17:59:36 UTC | Adjust TransportActionProxyTests#testSendLocalRequest (#91321) (#91324) Relates #91289 | 04 November 2022, 17:59:36 UTC |
| b4b138c | Francisco Fernández Castaño | 04 November 2022, 16:13:28 UTC | Fix SnapshotBasedIndexRecoveryIT#testNodeDisconnectsDoNotOverAccountRecoveredBytes (#90849) (#91319) The clock used to check for timeouts does not provide enough granularity, and sometimes the RESTORE_FILE_FROM_SNAPSHOT is retried while the test expects to fail right away. This commit configures the nodes with a timeout of 0, effectively configuring it to avoid retrying recovery requests. Closes #90665 | 04 November 2022, 16:13:28 UTC |
| 55cde1a | Tanguy Leroux | 04 November 2022, 14:39:08 UTC | Fix TransportResponse reference counting in DirectResponseChannel (#91289) (#91315) In #76474 we fixed a circuit breaker leak in TransportActionProxy by incrementing a reference on the TransportResponse that is later decremented by the OutboundHandler. This works well for all cases except when the request targets the node which is also the proxy node. In that case the reference is incremented but will never be decremented as the local execution (using TransportService#localNodeConnection and DirectResponseChannel) bypasses the OutboundHandler. This change fixes the ref counting by also decrementing the TransportResponse in DirectResponseChannel. This will also have the consequence to correctly decrement used bytes of the request circuit breaker when GetCcrRestoreFileChunkResponse are executed on a node that is also a proxy node. | 04 November 2022, 14:39:08 UTC |
| 297a0e2 | David Kilfoyle | 04 November 2022, 13:32:15 UTC | [DOCS] Add time range info to TSDS docs (#91291) (#91314) * [DOCS] Add time range info to TSDS docs * Fixup | 04 November 2022, 13:32:15 UTC |
| fd1aadf | Luca Cavanna | 04 November 2022, 10:51:06 UTC | Revert "Validate field names when subobjects are disabled (#90950)" (#91096) This reverts commit 9147be6a6e8c042235f20e303ed221fb498946b1. In hindsight, there is a slight possibility that existing users are relying on Elasticsearch accepting field names that contain only dots when subobjects are disabled. The original change assumed that such fields were not accepted in mappings, but they actually are, as #90950 has surfaced. As a result, we end up rejecting field names that the mappings would happily accept, which is inconsistent. | 04 November 2022, 10:51:06 UTC |
| 396f062 | Przemyslaw Gomulka | 03 November 2022, 17:02:46 UTC | Use System dependent file separator backport (#91273) (#91282) In windows path separator is different than in linux, therefore a File.separator should be used instead of just / closes #91265 | 03 November 2022, 17:02:46 UTC |
| 9afc9d9 | Abdon Pijpelink | 03 November 2022, 10:06:27 UTC | Update move-to-step.asciidoc (#91114) (#91266) (cherry picked from commit 4192c5b327e871727af783f22bbc41ae857575e6) Co-authored-by: charliek17 <charliek17@users.noreply.github.com> | 03 November 2022, 10:06:27 UTC |
| b62014a | Nikola Grcevski | 02 November 2022, 18:41:26 UTC | Allow legacy index settings on legacy indices (#90264) (#91257) Add code to ignore unknown index settings on old indices, to allow rolling-upgrades to work. Co-authored-by: David Turner <david.turner@elastic.co> Co-authored-by: David Turner <david.turner@elastic.co> | 02 November 2022, 18:41:26 UTC |
| 6c2fd60 | Ignacio Vera | 02 November 2022, 15:42:32 UTC | Fix handling indexed envelopes crossing the dateline in mvt API (#91105) (#91254) This commit makes sure an envelope crossing the dateline generates a mutipolygon when creating a mvt feature so it is render properly. | 02 November 2022, 15:42:32 UTC |
| 6bc5f51 | David Turner | 02 November 2022, 15:31:53 UTC | Ensure cleanups succeed in JoinValidationService (#90601) (#91253) Today we sort of assume that cleanups succeed in the `JoinValidationService`. A failure in these places might explain the leaks seen in #90576 and #89712. It's not obvious that anything can fail here but let's make sure. | 02 November 2022, 15:31:53 UTC |
| 4ff8d7f | David Turner | 02 November 2022, 15:07:47 UTC | Add overflow behaviour test for RecyclerBytesStreamOutput (#90638) (#91252) Relates #90632 | 02 November 2022, 15:07:47 UTC |
| ebb6d52 | Nikola Grcevski | 02 November 2022, 14:06:20 UTC | Fix APM configuration file delete (#91058) (#91250) When we launch Elasticsearch with the APM monitoring agent, we create a temporary configuration file to securely pass the API key or secret. This temporary file is cleaned up on Elasticsearch Node creation. After we renamed the APM module, the delete logic didn't get updated, which means we never delete the file anymore. This commit: - fixes the APM module pattern match when we delete - adds additional delete safety net on failed node start - adds tests for ensuring the naming dependency isn't broken again. | 02 November 2022, 14:06:20 UTC |
| 25b7586 | David Kyle | 02 November 2022, 10:10:09 UTC | [ML] Allow NLP truncate option to be updated when span is set (#91224) (#91244) | 02 November 2022, 10:10:09 UTC |
| 69cf4d2 | Yang Wang | 02 November 2022, 03:14:00 UTC | Ensure TermsEnum action works correctly with API keys (#91170) (#91242) An API key's permission is bounded by its owner user's permission. When checking for DLS access, both the key's permission and the owner user's permission must be consulted. The access is granted only when it is granted by both. This PR ensures this logic is correctly enforced by the termsEnum action. | 02 November 2022, 03:14:00 UTC |
| 28528b7 | Yang Wang | 02 November 2022, 01:46:33 UTC | Ensure PKI's delegated_by_realm metadata respect run-as (#91173) (#91241) When delegated PKI authentication is used, the delegatee's realm name is added as a metadata field. This realm name should be the effective subject's realm instead of that of the authenticating subject. This PR ensures this is the case. | 02 November 2022, 01:46:33 UTC |
| 3f24a51 | Yang Wang | 02 November 2022, 00:49:37 UTC | Fix NPE in auditing authenticationSuccess for non-existing run-as user (#91171) (#91240) When run-as fails because the target user does not exist, the authentication is created with a null lookup realm. It is then rejected at authorization time. But for authentication, it is treated as success. This can lead to NPE when auditing the authenticationSuccess event. This PR fixes the NPE by checking whether lookup realm is null before using it. Relates: https://github.com/elastic/elasticsearch/pull/91126#discussion_r1005472501 | 02 November 2022, 00:49:37 UTC |
| 4be14ee | Christos Soulios | 01 November 2022, 18:26:54 UTC | Prune changelogs after 8.5.0 release | 01 November 2022, 18:26:54 UTC |
| c08ae75 | Christos Soulios | 01 November 2022, 18:26:01 UTC | Bump versions after 8.5.0 release | 01 November 2022, 18:26:04 UTC |
| 6c268ff | Lisa Cawley | 01 November 2022, 15:25:55 UTC | [8.5] [DOCS] Fix typo in knn-search.asciidoc (#91206) (#91212) | 01 November 2022, 15:25:55 UTC |
| f190f8e | Abdon Pijpelink | 01 November 2022, 14:32:38 UTC | [DOCS] Update the feature state example in the snapshot and restore docs (#90328) (#91222) (cherry picked from commit 0249d1650f278253eff5721ed43cd53fa0dbab6e) Co-authored-by: Anthony McGlone <102866938+anthonymcglone2022@users.noreply.github.com> | 01 November 2022, 14:32:38 UTC |
| fb075e8 | Abdon Pijpelink | 01 November 2022, 14:32:18 UTC | Fix typo in stop-tokenfilter.asciidoc (#91128) (#91208) Since ignore_case is set to true in our custom stop words filter, the matching will be case-insensitive. (cherry picked from commit a03fba9d77c7ab781fca98eaffe901a17419772e) Co-authored-by: Siniša Subašić <68671543+sinisuba@users.noreply.github.com> | 01 November 2022, 14:32:18 UTC |
| 7d35e8f | David Kilfoyle | 01 November 2022, 14:06:17 UTC | [Docs] Remove feature flag from downsampling page (#91228) (#91230) | 01 November 2022, 14:06:17 UTC |
| 6f7deee | Iraklis Psaroudakis | 01 November 2022, 13:04:31 UTC | 2 snapshot threads for large blob test (#91226) In 8.5, the default snapshot thread pool has 1 thread when the "Timeout with large blobs" runs with the -Dtests.configure_test_clusters_with_one_processor=true pipeline. This makes the 1sec timeout response to never be sent out (because the 1 thread is busy with the slow blob), while with 2 threads it can be sent out. Fixes #90353 | 01 November 2022, 13:04:31 UTC |
| 94d3fc8 | Julie Tibshirani | 27 October 2022, 16:00:56 UTC | Remove experimental marking from kNN search (#91065) This commit removes the experimental tag from kNN search docs and makes some docs improvements: * Add a prominent warning about memory usage in the kNN search guide * Link to the performance tuning guide from the main guide * Clarify the memory requirements section in the tuning guide | 31 October 2022, 19:17:19 UTC |
| 9d47898 | Gordon Brown | 31 October 2022, 18:13:07 UTC | Unmute testCacheUnderConcurrentAccess (#91186) (#91214) This test has been muted for a very long time. I haven't been able to reproduce the failure, so I'm going to unmute it and see if it starts failing on CI. Relates https://github.com/elastic/elasticsearch/issues/51914 | 31 October 2022, 18:13:07 UTC |
| 1485381 | Justin Cranford | 31 October 2022, 15:44:45 UTC | Add STDOUT redirect to security auto-conf skip. (#91165) (#91202) | 31 October 2022, 15:44:45 UTC |
| d8c3320 | Simon Cooper | 31 October 2022, 14:54:09 UTC | Change timezone rounding test to a timezone that is unlikely to change in the future (#91155) (#91198) | 31 October 2022, 14:54:09 UTC |
| 77b936e | Mary Gouseti | 28 October 2022, 12:35:08 UTC | Mute reference/cluster/nodes-stats/line_2751 (#91174) (#91175) Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> | 28 October 2022, 12:35:08 UTC |
| 1110476 | Adam Locke | 27 October 2022, 12:36:48 UTC | Update threadpool.asciidoc (#90098) (#91156) * Update threadpool.asciidoc Starting from 8.0 the value of the `node.processors` setting is bounded by the number of available processors https://github.com/elastic/elasticsearch/pull/44894 * Update docs/reference/modules/threadpool.asciidoc Co-authored-by: Adam Locke <adam.locke@elastic.co> (cherry picked from commit fe0036fdbfcb5e44e349123bb5caed78ff076bd2) Co-authored-by: Frederic Dartayre <30438455+fdartayre@users.noreply.github.com> | 27 October 2022, 12:36:48 UTC |
| e5dae5b | Yang Wang | 26 October 2022, 22:52:46 UTC | [Doc] Improve doc for certutil parameter applicability (#91124) (#91145) The http command does not take most of the parameters. This PR ensures it is consistently documented for all parameters. | 26 October 2022, 22:52:46 UTC |
| 24e7262 | Abdon Pijpelink | 26 October 2022, 12:39:36 UTC | Refine geo-point and geo-shape docs (#90913) (#91135) * Refine geo-point and geo-shape docs While reviewing the docs for another issue, some deprecated references to prefix-trees were discovered, leading to interest in bringing the docs a little more up-to-date. * Update docs/reference/mapping/types/geo-point.asciidoc Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> * Update docs/reference/mapping/types/geo-shape.asciidoc Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> Co-authored-by: Abdon Pijpelink <abdon.pijpelink@elastic.co> (cherry picked from commit c19f642d94a7bce62303c604f55d6b7783df39f2) Co-authored-by: Craig Taverner <craig@amanzi.com> | 26 October 2022, 12:39:36 UTC |
| ebad5d3 | Slobodan Adamović | 26 October 2022, 11:36:58 UTC | Fix carriage return removal when reading a long line from terminal (#91131) (#91133) Fixes an edge case issue where the carriage return (`\r`) character hasn't been removed if it's the last character in the buffer and the buffer is full. Resolves #89227 | 26 October 2022, 11:36:58 UTC |
| 64f49e4 | Dimitris Athanasiou | 26 October 2022, 09:59:13 UTC | [8.5][ML] Fix trained model deployment stats after allocations change (#91122) When the number of allocations of a particular assignment gets changed as the result of a rebalance, we update the params stored in the local task via `TrainedModelDeploymentTask.updateNumberOfAllocations`. These are then used to report stats for the deployment. This commit fixes a bug where we wrongly report: - no `cache_size` - `threads_per_allocation` as equal to local `number_of_allocations` for a deployment after there was a change to the number of allocations of an assignment. Backport of #91095 | 26 October 2022, 09:59:13 UTC |
| 57160b6 | Carlos Delgado | 26 October 2022, 08:30:13 UTC | [8.5] Add privileges for connectors index creation (#91026) (#91055) | 26 October 2022, 08:30:13 UTC |
| ee8f51d | Benjamin Trent | 25 October 2022, 12:50:29 UTC | Bump versions after 7.17.7 release | 25 October 2022, 12:50:31 UTC |
| 52d8235 | Carlos Delgado | 25 October 2022, 11:27:02 UTC | [8.5] Add privileges for crawler logs indices in Enterprise Search service account (#91094) (#91108) | 25 October 2022, 11:27:02 UTC |
| 89d05ce | Adam Locke | 24 October 2022, 20:06:39 UTC | Update docker image (#90730) (#91099) Co-authored-by: Flavio <flavio.fernandes6@gmail.com> | 24 October 2022, 20:06:39 UTC |
| e228956 | Adam Locke | 24 October 2022, 19:29:10 UTC | Fix itemized list (#90855) (#91097) Co-authored-by: Stéphane Campinas <stephane.campinas@gmail.com> | 24 October 2022, 19:29:10 UTC |
| c94b470 | Ignacio Vera | 24 October 2022, 14:58:18 UTC | [8.5] Upgrade to lucene-9.4.1 (#91074) (#91090) | 24 October 2022, 14:58:18 UTC |
| 223092e | Przemysław Witek | 24 October 2022, 12:08:11 UTC | [Transform] Expand the docs section regarding mappings deduction in transform's dest index (#91077) (#91078) | 24 October 2022, 12:08:11 UTC |
| 7723cac | Christos Soulios | 21 October 2022, 09:50:20 UTC | [8.5][DOCS] Add release notes for 8.5.0 (#91029) Add release notes for v8.5.0 after BC6 has been cut | 21 October 2022, 09:50:20 UTC |
| e05d40d | Joe Gallo | 20 October 2022, 14:36:30 UTC | Ingest CompoundProcessor asynchronous time stats can be incorrect (#91033) | 20 October 2022, 18:06:11 UTC |
| fd58815 | Joe Gallo | 19 October 2022, 13:39:48 UTC | Off-by-one innerExecute current processor count (#90995) | 20 October 2022, 18:06:11 UTC |
| 61dca20 | Dimitris Athanasiou | 20 October 2022, 17:37:05 UTC | [ML] Delete 90903 change log entry (#91054) (#91056) As #90903 made the cut for `8.5.0`, the issue is no longer a bug. After udpating the PR to `non-issue`, this commit also deletes the release note from the changelog. | 20 October 2022, 17:37:05 UTC |
| 4e2c7ff | Iraklis Psaroudakis | 20 October 2022, 15:58:41 UTC | Explain disk headroom settings more in docs (#90763) (#91053) Relates to #81406 | 20 October 2022, 15:58:41 UTC |
| ee39ddc | Francisco Fernández Castaño | 20 October 2022, 15:32:34 UTC | Fix ReactiveStorageIT#testScaleWhileShrinking (#90865) (#91051) Sometimes Windows throws an AccessDeniedException when some of the attributes for a file that has been deleted is read. This caused issues in this test since it prevented gathering the FS stats from one of the nodes where some of the relocated index files were deleted concurrently. Closes #88842 | 20 October 2022, 15:32:34 UTC |
| 7612b2b | Armin Braun | 20 October 2022, 13:42:34 UTC | Fix integer overflow in AsyncTaskIndexService (#91044) (#91048) If source is larger than int_max we overflow on the multiplication. Since we didn't use the same variable we will increment the breaker once and then increment it again with the broken overflown int on release. | 20 October 2022, 13:42:34 UTC |
| 474c81f | Armin Braun | 19 October 2022, 12:23:53 UTC | Fork TransportClusterStateAction to MANAGEMENT (#90996) (#91015) For large cluster states the serialization of the response can take a longer time than is acceptable on a transport thread. Also, moving this to management creates a bit of a natural bottleneck via the management pool's size in case a master node is hit by a large number of concurrent requests which might also be helpful in limiting the considerable memory consumption of this transport action. | 19 October 2022, 12:23:53 UTC |
| 187aaad | Nikola Grcevski | 18 October 2022, 21:01:30 UTC | TaskManager traceRequest on/off for CancellableTask (#90972) (#91000) In certain situations, like ML model loading jobs we don't want to enable the APM tracer as per the original design. For that purpose, the register method contains the traceRequest flag, which is enabled by default unless explicitly disabled. We disable tracing for ML model loading, but we didn't correctly when the Task was of CancellableTask type. This PR fixes that control flow bug and adds tests for both control flow paths. | 18 October 2022, 21:01:30 UTC |
| 58e714c | Armin Braun | 18 October 2022, 12:15:59 UTC | Be more paranoid about resources in OutboundHandler (#90980) (#90982) Make outbound handler more paranoid about releasing resources on various potential (unexpected and technically buggy) exceptions. | 18 October 2022, 12:15:59 UTC |
| 31114de | Christos Soulios | 18 October 2022, 11:15:19 UTC | [CI] Fix RollupStepTest.testHashcodeAndEquals() (#90958) (#90979) The test was failing because the random generated value was the same as the previous one, so equals() and hashCode() were the same. Fixes #90843 | 18 October 2022, 11:15:19 UTC |
| 793b469 | Leaf-Lin | 18 October 2022, 00:02:02 UTC | [DOCS] Add CCR limitation (#87348) (#90970) * Add CCR limitation closes https://github.com/elastic/elasticsearch/issues/86121 * Add restored index auto follow pattern restriction https://github.com/elastic/elasticsearch/issues/87055 * Moving content to existing CCR page + several changes * Remove sections to consolidate limitation information * Delete separate file * Remove restored indices from list of things that aren't replicated Co-authored-by: Adam Locke <adam.locke@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> Co-authored-by: Adam Locke <adam.locke@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> | 18 October 2022, 00:02:02 UTC |
| 618b142 | Andrei Dan | 17 October 2022, 18:19:01 UTC | [CCR] Data streams: order follower backing indices by original name (#90850) (#90963) This adds a test to simulate the case when the follower data stream has an early generation shrunken backing index and the leader data stream is rolled over. Currently the rolled over backing index (the new leader write index) will end up not being the leader index in the follower data stream, and even more, the shrunken index will be ordered as the write index in the follower data stream. This attempts to mitigate this a bit by sorting the follower backing indices by the original names (ignoring any characters before the standardised `.ds-` prefix). | 17 October 2022, 18:19:01 UTC |
| ac84eb9 | Lisa Cawley | 17 October 2022, 17:41:48 UTC | [8.5] [DOCS] Add ML CPP PRs to release notes (#90961) (#90962) | 17 October 2022, 17:41:48 UTC |
| de482be | Jack Conradson | 17 October 2022, 16:58:37 UTC | Fix true/false accumulation bug in boolean source fallback (#90895) (#90959) The SourceValueFetcherSortedBooleanDocValues previously accumulated true/false value totals without resetting between documents. This means that any counts on the same segment would continue to increase true/false value totals over multiple documents. This change fixes that bug by resetting true/false value totals per document. | 17 October 2022, 16:58:37 UTC |
| 9147be6 | Luca Cavanna | 17 October 2022, 14:58:32 UTC | Validate field names when subobjects are disabled (#90950) Validation of field names is almost entirely performed as part of the DotExpandingXContentParser which splits dotted paths and validate each part separately. There are situations though where the parser is not used, for instance when subobjects are disabled, for which we can't rely on its validation. This commit introduces validation for field names that only contain dots to DocumentParser, as well as a couple of tests to verify the current behaviour. Relates to #21862 | 17 October 2022, 14:58:32 UTC |
| 5c14f9c | Mary Gouseti | 17 October 2022, 12:10:50 UTC | Test fix: make DataStreamIT not sensitive to midnight (#90595) (#90910) (#90944) | 17 October 2022, 12:10:50 UTC |
| 59c7c6b | Valeriy Khakhutskyy | 17 October 2022, 09:31:56 UTC | Mute tests for #90843 (#90845) (#90939) Mute tests for #90843 Backport of #90845 | 17 October 2022, 09:31:56 UTC |
| f154e5d | James Baiera | 14 October 2022, 19:40:18 UTC | Set the enrich maintenance cluster lifecycle listener only once (#90486) (#90927) This PR updates the code to only set the enrich maintenance task's lifecycle listener one time when it is first activated. Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> | 14 October 2022, 19:40:18 UTC |
| 4eeb9ea | Andrei Dan | 14 October 2022, 18:07:38 UTC | Fix testRedNodesWithoutAnyBlockedIndices (#90917) (#90923) | 14 October 2022, 18:07:38 UTC |
| cfc1a1f | Adam Locke | 14 October 2022, 18:01:42 UTC | Removed note saying DLS/FLS disable shard request cache (#90885) (#90921) Co-authored-by: Arpit Bhardwaj <arpitbhardwaj09@gmail.com> | 14 October 2022, 18:01:42 UTC |
| 3c6ca08 | Mary Gouseti | 14 October 2022, 14:49:18 UTC | Disk indicator troubleshooting guides (#90504) (#90911) | 14 October 2022, 14:49:18 UTC |
| e2a9d06 | Mary Gouseti | 14 October 2022, 13:28:41 UTC | Fix the details calculation of the disk indicator (#90715) (#90869) (#90900) This PR aligns the node counts in the details with the symptom and diagnosis. All of them now are based on cluster state of the coordinating node. Any information coming from the health node regarding nodes that are not known to the coordinating node will be ignored. | 14 October 2022, 13:28:41 UTC |
| 9c5d6c0 | Iraklis Psaroudakis | 14 October 2022, 12:10:40 UTC | Clarify watcher action uses PagerDuty API v2 (#90800) (#90908) Add a note to help avoid the confusion stemming from the fact that the watcher attributes still have names which seem to match the PagerDuty Events API v1 despite the fact that we are actually now using v2 of that API. Delete outdated screenshots Co-authored-by: Iraklis Psaroudakis <kingherc@gmail.com> Co-authored-by: Tom Sparrow <793763+sparrowt@users.noreply.github.com> | 14 October 2022, 12:10:40 UTC |
| 5bf848c | Nik Everett | 26 September 2022, 20:34:00 UTC | Graph tests - more data on failure (#90379) Get more data when the graph timeout test fails. Relates to #90286. | 14 October 2022, 12:05:48 UTC |
| 6793e14 | Dimitris Athanasiou | 14 October 2022, 11:26:48 UTC | [ML] Require correct tier processors when multiple AZs are present (#90903) (#90907) When we added an autoscaling decider for processors we assumed wrongly that when there are multiple availability zones ML would get the processors it requests for the tier multiplied by the number of AZs. This assumption was incorrect. This commit fixes this issue which should correct autoscaling behaviour in clusters with multiple AZs. | 14 October 2022, 11:26:48 UTC |
| 41f2157 | Julie Tibshirani | 12 October 2022, 21:53:53 UTC | Add guide for tuning kNN search (#89782) This 'how to' guide explains performance considerations specific to kNN search. It takes inspiration from the 'tune for search speed' guide. | 13 October 2022, 17:25:48 UTC |
| 9c4d874 | Andrei Dan | 13 October 2022, 13:40:26 UTC | [8.5] Report impact and diagnosis data nodes w/o disk space and no blocked indices (#90772) (#90873) This adds a dedicated impact and a diagnosis for the case where data nodes are out of disk space but not read-only blocked indices are in the cluster. (cherry picked from commit dac276902b8c148e50a4968977f7f4fcea12265b) Signed-off-by: Andrei Dan <andrei.dan@elastic.co> | 13 October 2022, 13:40:26 UTC |
| 5a5bb66 | David Kilfoyle | 13 October 2022, 13:15:29 UTC | [DOCS] Add Downsampling docs (#88571) (#90853) This adds documentation for downsampling of time series indices. | 13 October 2022, 13:15:29 UTC |
| 2fc2d54 | Mark Vieira | 12 October 2022, 21:12:56 UTC | Upgrade bundled JDK to Java 19 (#90571) (#90854) | 12 October 2022, 21:12:56 UTC |
| 16b0dd2 | Dimitris Athanasiou | 11 October 2022, 13:25:04 UTC | [ML] Detect ML nodes by AZ from consistent state during rebalance (#90770) (#90787) This commit ensures the ML nodes by availability zone considered during rebalancing trained model deployments is consistent and based off a specific cluster state. | 11 October 2022, 13:25:04 UTC |
| 4535391 | Armin Braun | 11 October 2022, 12:06:30 UTC | Fix quadratic complexity in SnapshotStatus serialization (#90795) (#90803) The loop here is obviously O(N^2) and it shows as taking multiple seconds to build the by-index map in x-content serialization for a 25k shards snapshot. This makes the logic O(N) and fixes some potential issues with consistency around concurrent access by using a more appropriate access pattern to the field caching the by-index map. | 11 October 2022, 12:06:30 UTC |
| d63cc81 | Josh Dover | 11 October 2022, 11:38:20 UTC | Remove incorrect outputs mapping from fleet-agents (#90714) (#90798) | 11 October 2022, 11:38:20 UTC |
| 4541114 | Andrei Dan | 11 October 2022, 09:29:27 UTC | [8.5] Health API symptom copy (#90761) (#90788) * Health API symptom copy (#90761) Update the copy for a few symptoms and impacts in the health API. * Update yaml test min version | 11 October 2022, 09:29:27 UTC |
| 0450906 | Julie Tibshirani | 11 October 2022, 01:50:37 UTC | Make HitContext fields final (#90784) HitContext's SourceLookup field can no longer be swapped out, so we can make it final. | 11 October 2022, 01:52:22 UTC |
| dab20b2 | Yang Wang | 11 October 2022, 00:15:22 UTC | [Doc] Correctly cite setup-passwords CLI for minimal security guide (#90579) (#90779) The minimal security guide is meant to use setup-passwords instead of reset-password. This is consistent with the 7.x version. In longer term, we might actually want to replace setup-passwords with reset-password. But it requires rewriting of the associated the guide while this PR fixes the immediate problem of incorrect usage of CLI command. Relates: #88393 | 11 October 2022, 00:15:22 UTC |
| f29ead8 | Yang Wang | 10 October 2022, 22:36:07 UTC | [Doc] Document DLS limitation for aggregation (#90701) (#90778) This PR clarifies the limitation for DLS in terms of aggreation. | 10 October 2022, 22:36:07 UTC |
| 49f244b | Andrei Dan | 10 October 2022, 11:38:39 UTC | Rename the fields reported under details by the disk indicator (#90717) (#90756) Currently, we report the count of affected nodes and indices as part of the disk indicator using a leaky abstraction. Namely, we use the status we assign to nodes internally to nodes based on their disk usage (red, yellow, green, unknown). However, these statuses don't have an explicit meaning outside the implementation details e.g. a red node would probably convey it's a node experiencing disk issues but not what kind This proposes being explicit in what we return to our health API users e.g. ``` "details": { "indices_with_readonly_block": 2, "nodes_with_enough_disk_space": 0, "nodes_with_unknown_disk_status": 0, "nodes_over_high_watermark": 0, "nodes_over_flood_stage_watermark": 2 } ``` (cherry picked from commit b55f5fd77b7a37d79db62652d7faf82431e47e23) Signed-off-by: Andrei Dan <andrei.dan@elastic.co> | 10 October 2022, 11:38:39 UTC |
