dab5822 | Ryan Ernst | 15 January 2021, 19:03:25 UTC | Remove deprecation log file check in 6.8 (#67465) This test has been muted for a very long time, yet the underlying cause is still unknown, but it only occurs in 6.8. This commit removes the failing portion of the test so that we have some coverage. closes #35990 | 15 January 2021, 19:03:25 UTC |
2f07e90 | Nhat Nguyen | 13 January 2021, 14:02:42 UTC | Fix AllocationIdIT test failure on WindowFS (#67179) This test failed on WindowsFS. We failed to remove the corrupted file if it's being opened (for a short window by ListShardStore action) and the pending delete files were clear when we restarted that node. This commit fixes the issue by shutting down the node before removing the corrupted file to avoid any access to that file. Closes #66893 | 13 January 2021, 21:15:53 UTC |
dad1e91 | David Roberts | 13 January 2021, 16:55:09 UTC | Extend the selective muting of memory tests on Debian 8 (#67456) The selective muting implemented for autoscaling in #67159 is extended to the ML tests that also fail when machine memory is reported as 0. Most of the logic to determine when memory will not be accurately reported is now in a utility method in the base class. Relates #66885 Backport of #67422 | 13 January 2021, 16:55:09 UTC |
ef6f94f | William Brafford | 07 January 2021, 18:05:52 UTC | [6.8] Backport some fixes to CharArraysTest#testConstantTimeEquals() (#67168) * Ensure char array test uses different values (#47238) The test of constantTimeEquals could get unlucky and randomly produce the same two strings. This commit tweaks the test to ensure the two string are unique, and the loop inside constantTimeEquals is actually executed (which requires the strings be of the same length). * Fix CharArraysTests.testConstantTimeEquals() (#47346) The change #47238 fixed a first issue (#47076) but introduced another one that can be reproduced using: org.elasticsearch.common.CharArraysTests > testConstantTimeEquals FAILED java.lang.StringIndexOutOfBoundsException: String index out of range: 1 at __randomizedtesting.SeedInfo.seed([DFCA64FE2C786BE3:ED987E883715C63B]:0) at java.lang.String.substring(String.java:1963) at org.elasticsearch.common.CharArraysTests.testConstantTimeEquals(CharArraysTests.java:74) REPRODUCE WITH: ./gradlew ':libs:elasticsearch-core:test' --tests "org.elasticsearch.common.CharArraysTests.testConstantTimeEquals" -Dtests.seed=DFCA64FE2C786BE3 -Dtests.security.manager=true -Dtests.locale=fr-CA -Dtests.timezone=Pacific/Johnston -Dcompiler.java=12 -Druntime.java=8 that happens when the first randomized string has a length of 0. Co-authored-by: Ryan Ernst <ryan@elastic.co> Co-authored-by: Tanguy Leroux <tlrx.dev@gmail.com> | 07 January 2021, 18:05:52 UTC |
3b4ab52 | Rory Hunter | 07 January 2021, 17:04:41 UTC | Fix missing `QuotaPath.unwrap()` instances (#67165) Closes #67164. There were a couple of places in `QuotaAwareFileSystemProvider` where the supplied paths were not being unwrapped back into platform-specific path objects. | 07 January 2021, 17:14:28 UTC |
b0cfc82 | Mayya Sharipova | 29 December 2020, 20:00:12 UTC | Mute RolloverIT testRolloverWithDateMath (#66875) Relates to #66874 | 29 December 2020, 20:00:12 UTC |
81b0b16 | Sandeep Kanabar | 23 December 2020, 13:58:36 UTC | [DOCS] Fix typo (#66779) Co-authored-by: James Rodewig <40268737+jrodewig@users.noreply.github.com> | 23 December 2020, 13:58:36 UTC |
962b11b | Gordon Brown | 18 December 2020, 22:35:35 UTC | Mute tests failing on Debian 8 due to memory reporting (#66648) See https://github.com/elastic/elasticsearch/issues/66629 for details | 18 December 2020, 22:35:35 UTC |
118c6a1 | asce0705 | 16 December 2020, 18:12:14 UTC | Fix constructors of NoOpResult (#66269) | 16 December 2020, 18:28:05 UTC |
f27e7c6 | Jim Ferenczi | 29 October 2019, 09:42:37 UTC | Fix expectations in SearchAfter integration tests (#48372) This commit fixes the expectations of SearchAfterIT#shouldFail regarding the inner exceptions that should be thrown when testing failures. The exception is sometimes wrapped in a QueryShardException so this change only checks that the toString representation contains the expected message. Closes #43143 | 16 December 2020, 14:34:44 UTC |
5ca5987 | David Turner | 15 December 2020, 17:04:26 UTC | Give a less dangerous example of logger config (#66318) Today the docs use `logger.org.elasticsearch.transport: TRACE` as the example for adjusting the logger config. This is a dangerous thing to suggest since that's one of the most verbose loggers we have. An accidental copy-and-paste of this example into a busy cluster can cause damage. This commit suggests `logger.org.elasticsearch.discovery: DEBUG` instead, which is much more benign. | 15 December 2020, 17:21:12 UTC |
d6d022d | Nhat Nguyen | 13 December 2020, 14:23:07 UTC | Fix TranslogTests#testStats (#66227) If creating the latest translog file and retrieving a translog stats happen within the same millisecond, then the earliestLastModifiedAge will be zero. Closes #66092 | 13 December 2020, 15:59:23 UTC |
758a2ab | Nhat Nguyen | 13 December 2020, 02:15:44 UTC | CCR should check historyUUID in every read request (#65841) Today, CCR only checks the historyUUID of the leader shard when it has operations to replicate. If the follower shard is already in-sync with the leader shard, then CCR won't detect if the historyUUID of the leader shard has been changed. While this is not an issue, it can annoy users in the following situation: 1. The follower index is in-sync with the leader index 2. Users restore the leader index from snapshots 3. CCR won't detect the issue and report ok in its stats API 4. CCR suddenly stops working when users start indexing to the leader index 5. This commit makes sure that we always check historyUUID in every read-request so we can detect and report the issue as soon as possible. Backport of #65841 | 13 December 2020, 02:15:44 UTC |
e6276e8 | Rory Hunter | 07 December 2020, 10:57:59 UTC | Stop serialising licensed flag | 07 December 2020, 10:57:59 UTC |
1d9d6db | Rory Hunter | 04 December 2020, 09:33:15 UTC | Introduce licensed plugins (#65490) Backport of #64850. This PR introduces the concept of "licensed" plugins. Such plugins may only be installed on installations of the default distribution, and this is enforced by the plugin installer. This PR also moves the `quote-aware-fs` plugin to the `x-pack` directory, and marks it as licensed. Note that I didn't move the plugin source under `x-pack/plugin` because all the existing x-pack plugins are actually bundles as modules into the default distribution, whereas the `quota-aware-fs` plugin needs to remain a standalone plugin. | 04 December 2020, 09:33:15 UTC |
04d7f6b | Nhat Nguyen | 02 December 2020, 22:41:33 UTC | Ensure notify when proxy connections disconnect (#65697) TransportService doesn't respond to the pending requests of proxy connections when the underlying connections get disconnected because proxy connections do not override the getCacheKey method. Some CCS requests would never be completed because of this bug. | 03 December 2020, 22:03:34 UTC |
0cd5506 | Mark Vieira | 02 December 2020, 17:27:24 UTC | Always use "elasticsearch" as Gradle project name (#65709) | 02 December 2020, 17:30:01 UTC |
4afc78e | Przemko Robakowski | 30 November 2020, 22:27:39 UTC | Fix TranslogTests.testTotalTests when n=0 (#65632) When n=0 in TranslogTests.testTotalTests we never update earliestLastModifiedAge so it fails comparison with default value of total.getEarliestLastModifiedAge() which is 0. In this change we always check this special case and then select n>0 Closes #65629 | 30 November 2020, 23:33:23 UTC |
a4e3c68 | Howard | 26 November 2020, 20:43:47 UTC | Fix the earliest last modified age of translog stats (#64753) Currently translog's `earliest_last_modified_age` field is always 0 in `_nodes/stats` response. | 30 November 2020, 22:35:53 UTC |
2eee124 | Andrei Stefan | 23 November 2020, 13:54:15 UTC | Extend the interval date comparison (#65361) | 23 November 2020, 13:54:15 UTC |
f544be0 | Rory Hunter | 19 November 2020, 19:47:54 UTC | Reimplement QuotaAwarePath.toFile() (#65191) An issue was observed where calling `QuotaAwarePath.toFile()` caused `UnsupportedOperationException: Path not associated with default file system.` Replacing the implementation with that of the interface's default implementation in a later JDK prevents the exception. As to why, I suspect that whatever implementation was being called before, or the location of that code, caused an incorrect default filesystem to be picked for comparison. | 19 November 2020, 19:49:05 UTC |
4bf4614 | Nhat Nguyen | 18 November 2020, 02:03:07 UTC | Stop renew retention leases when follow task fails (#65168) If a shard follow-task hits a non-retryable error and stops, then we should also stop the retention-leases renewal process associated with that follow-task. | 18 November 2020, 20:54:37 UTC |
767003c | James Rodewig | 18 November 2020, 17:29:05 UTC | [DOCS] Remote reindex is not fwd compatible (#65207) | 18 November 2020, 17:38:58 UTC |
03a48f9 | James Rodewig | 30 July 2020, 13:08:59 UTC | [DOCS] Note remote reindex is not fwd compatible (#60425) | 18 November 2020, 14:23:27 UTC |
9082d4f | Nhat Nguyen | 04 November 2020, 13:21:35 UTC | Increase timeout in testCleanUpShardFollowTasksForDeletedIndices (#64562) If the deleted index has N shards, then ShardFollowTaskCleaner can send N*(N-1)/2 requests to remove N shard-follow tasks. I think that's fine as the implementation is straightforward. The test failed when the deleted index has 8 shards. This commit increases the timeout in the test. Closes #64311 | 10 November 2020, 17:21:44 UTC |
fb582dd | Ryan Ernst | 09 November 2020, 23:11:45 UTC | Use no-kpi subdomains for tests downloading artifacts (#64502) (#64749) This commit converts build code that downloads distributions or other artifacts to use the new no-kpi subdomain, and removes the formerly used no-kpi header. | 09 November 2020, 23:11:45 UTC |
0d448db | Nhat Nguyen | 08 November 2020, 00:38:45 UTC | Do not wrap can_match searchers (#64742) The can_match phase does not work with frozen indices that have document-level security enabled. Also, we should not wrap can_match searchers to reduce the load during the can_match phase. | 08 November 2020, 00:38:45 UTC |
ab0f5c0 | Ryan Ernst | 03 November 2020, 22:14:21 UTC | Require Java 11 for 6.8 compile version (#64556) We lost the ability a while ago to upgrade the java version in 6.8, since we are stuck on a much older version of gradle. Since Java 12 is EOL (past the 6 months original support), it would be good to require an LTS version there. This commit changes the minimum compiler version back down to java 11 instead of 12. | 03 November 2020, 22:14:21 UTC |
5c75c04 | Ryan Ernst | 03 November 2020, 22:14:01 UTC | Fix bats tests expected jar version for poi (#64558) This commit fixes the expected version of the poi jar for the attachment plugin, which changed in 6.8. The bats tests no longer existin in 7.x or master, so the fix is only needed here. | 03 November 2020, 22:14:01 UTC |
088076c | Mark Vieira | 02 November 2020, 17:20:40 UTC | Upgrade to version 3.5 of Gradle Enterprise plugin | 02 November 2020, 17:20:40 UTC |
e8e948e | James Rodewig | 02 November 2020, 14:23:06 UTC | [DOCS] Note heap size must be set to same min and max (#64090) (#64477) Co-authored-by: C.J. Jameson <cjcjameson@users.noreply.github.com> | 02 November 2020, 14:23:06 UTC |
99d703e | Rory Hunter | 30 October 2020, 19:04:17 UTC | Introduce quota-aware filesystem ES plugin (#64331) Partial backport of #63620. Closes #61145. This PR adds a quota-aware filesystem plugin to Elasticsearch. This plugin offers a way to provide user quota limits (specifically, total quota size and available quota size) to Elasticsearch, in an implementation-agnostic manner. | 30 October 2020, 19:04:17 UTC |
46d8c84 | Ryan Ernst | 29 October 2020, 20:19:44 UTC | Upgrade bundled jdk to 15.0.1 and switch back to AdoptOpenJDK This commit updates the bundled jdk to 15.0.1, and at the same time once again switches the bundled jdk back to adoptopenjdk, which has fixed their build problem and regained support for glibc 2.12. closes #64026 | 29 October 2020, 20:19:44 UTC |
f70c85d | junmuz | 22 October 2020, 20:52:15 UTC | Update getting-started.asciidoc for Java version (#63106) Update client documentation to state "at least" Java 1.8 | 22 October 2020, 20:52:15 UTC |
4c12fc1 | Przemyslaw Gomulka | 22 October 2020, 13:10:32 UTC | [6.8] Version bump to 6.8.14 (#63822) version bump after 6.8.13 release | 22 October 2020, 13:10:32 UTC |
1a7f86b | James Rodewig | 22 October 2020, 12:31:33 UTC | [DOCS] Remove 6.8.13 coming tag (#64058) | 22 October 2020, 12:31:33 UTC |
8ad7a4b | Przemyslaw Gomulka | 21 October 2020, 20:47:50 UTC | [DOCS] Release notes for 6.8.13 (#63838) Co-authored-by: lcawl <lcawley@elastic.co> Co-authored-by: James Rodewig <40268737+jrodewig@users.noreply.github.com> | 21 October 2020, 20:47:50 UTC |
51e22e8 | Rene Groeschke | 19 October 2020, 15:52:48 UTC | Explicitly use TLS 1.2 in Gradle (#63880) | 19 October 2020, 15:59:22 UTC |
4fdc088 | Ioannis Kakavas | 16 October 2020, 10:29:33 UTC | Move RestRequestFilter to core (#63791) Move RestRequestFilter to core so that Rest requests outside xpack can use it to filter fields and expand its usage. Backport of #63507 | 16 October 2020, 10:29:33 UTC |
be13c69 | Ryan Ernst | 15 October 2020, 20:05:21 UTC | Remove opensuse 42 from vagrant tests (#63759) Opensuse 42 has not worked in a while. The test image is unmaintained, and cannot be launched. It was removed from CI packaging test runs, but still remained in vagrant tests. This commit removes it from vagrant tests. | 15 October 2020, 20:13:56 UTC |
2f119bc | Nhat Nguyen | 06 October 2020, 01:43:08 UTC | Retry CCR shard follow task when no seed node left (#63225) If the connection between clusters is disconnected or the leader cluster is offline, then CCR shard-follow tasks can stop with "no seed node left". CCR should retry on this error. | 06 October 2020, 15:42:54 UTC |
cd38a51 | Lisa Cawley | 05 October 2020, 19:41:04 UTC | [DOCS] Add experimental tag to rollup APIs (#63206) | 05 October 2020, 20:39:23 UTC |
509507d | Adam Locke | 02 October 2020, 15:00:48 UTC | Removing wording indicating that read-only index block is released automatically. (#62627) | 02 October 2020, 15:00:48 UTC |
e4e4aab | Sylvain Wallez | 30 September 2020, 20:36:44 UTC | Enable support for compressed response within RestHighLevelClient (#63087) Co-authored-by: Hakky54 <hakangoudberg@hotmail.com> | 30 September 2020, 20:36:44 UTC |
6a92ef5 | Przemyslaw Gomulka | 29 September 2020, 14:21:03 UTC | [doc] 6.8 deprecation logging bug - known issue (#62869) | 29 September 2020, 14:21:03 UTC |
90d27a0 | Bogdan Pintea | 21 September 2020, 10:55:04 UTC | SQL: Document Maven Central as a JDBC repository (#62618) * Add Maven Central as a JDBC repository Document Maven Central as a JDBC repository. (cherry picked from commit 2bc4d7eb19a26bf21b11214c4351470b677e1598) | 21 September 2020, 11:09:23 UTC |
957eaf6 | Lisa Cawley | 17 September 2020, 23:01:14 UTC | [DOCS] Add warning about derived keys (#62588) | 17 September 2020, 23:01:14 UTC |
ffe7df0 | Ryan Ernst | 17 September 2020, 19:08:19 UTC | Upgrade the bundled JDK to JDK 15 (#62580) This commit upgrades the bundled JDK to JDK 15. | 17 September 2020, 19:08:19 UTC |
11ba9e5 | Albert Zaharovits | 15 September 2020, 17:57:28 UTC | Ensure authz operation overrides transient authz headers (#61621) AuthorizationService#authorize uses the thread context to carry the result of the authorisation as transient headers. The listener argument to the `authorize` method must necessarily observe the header values. This PR makes it so that the authorisation transient headers (`_indices_permissions` and `_authz_info`, but NOT `_originating_action_name`) of the child action override the ones of the parent action. Co-authored-by: Tim Vernum tim@adjective.org | 15 September 2020, 17:57:28 UTC |
e559f3c | Nhat Nguyen | 05 September 2020, 02:46:51 UTC | CCR should retry on CircuitBreakingException (#62013) CCR shard follow task can hit CircuitBreakingException on the leader cluster (read changes requests) or the follower cluster (bulk requests). CCR should retry on CircuitBreakingException as it's a transient error. | 10 September 2020, 21:35:53 UTC |
d15dad7 | Julie Tibshirani | 08 September 2020, 22:57:34 UTC | Small fixes to breaking changes docs. * Remove accidental block formatting. * Move ngram and shingle changes to the analysis section. | 08 September 2020, 23:44:03 UTC |
8a73888 | Mikołaj Przybysz | 04 September 2020, 13:00:11 UTC | [DOCS] Add line break to get ILM lifecycle API docs (#61892) | 04 September 2020, 13:03:17 UTC |
c38f269 | Nhat Nguyen | 31 August 2020, 12:49:27 UTC | Set timeout of auto put-follow request to unbounded (#61679) If the master node of the follower cluster is busy, then the auto-follower will fail to initialize the following process. This also occurs when an auto-follow pattern matches multiple indices. We should set the timeout of put-follow requests issued by the auto-follower to unbounded to avoid this problem. Closes #56891 | 01 September 2020, 13:59:51 UTC |
2bb45b4 | Ioannis Kakavas | 01 September 2020, 09:33:01 UTC | [6.8] Call ActionListener.onResponse exactly once (#61691) Under specific circumstances we would call onResponse twice, which led to unexpected behavior. Backport of #61584 | 01 September 2020, 09:33:01 UTC |
8abd807 | silvinalucero | 31 August 2020, 23:00:24 UTC | Add hlrc_rest_total_hits_as_int parameter in scroll request (#61685) This change sets the rest_total_hits_as_int option automatically in scroll search requests emitted by the Java high level rest client. This option is already set for normal search requests but was missing for scrolls in 6.x. Closes #61677 | 31 August 2020, 23:00:24 UTC |
f530513 | Albert Zaharovits | 27 August 2020, 20:28:32 UTC | Relax the index access control check for scroll searches (#61446) The check introduced by #60640 for scroll searches, in which we log if the index access control before the query and fetch phases differs from when the scroll context is created, is too strict, leading to spurious warning log messages. The check verifies instance equality but this assumes that the fetch phase is executed in the same thread context as the scroll context validation. However, this is not true if the scroll search is executed cross-cluster, and even for local scroll searches it is an unfounded assumption. The check is hence reduced to a null check for the index access. The fact that the access control is suitable given the indices that are actually accessed (by the scroll) will be done in a follow-up, after we better regulate the creation of index access controls in general. | 27 August 2020, 20:28:32 UTC |
00aecf8 | Albert Zaharovits | 27 August 2020, 18:40:45 UTC | Revert "Relax the index access control check for scroll searches (#61446)" This reverts commit 9a42f3fa673ac4c21856cb20469a6e44727e3c68. | 27 August 2020, 18:40:45 UTC |
9a42f3f | Albert Zaharovits | 27 August 2020, 18:16:01 UTC | Relax the index access control check for scroll searches (#61446) The check introduced by #60640 for scroll searches, in which we log if the index access control before the query and fetch phases differs from when the scroll context is created, is too strict, leading to spurious warning log messages. The check verifies instance equality but this assumes that the fetch phase is executed in the same thread context as the scroll context validation. However, this is not true if the scroll search is executed cross-cluster, and even for local scroll searches it is an unfounded assumption. The check is hence reduced to a null check for the index access. The fact that the access control is suitable given the indices that are actually accessed (by the scroll) will be done in a follow-up, after we better regulate the creation of index access controls in general. | 27 August 2020, 18:21:56 UTC |
88c2a75 | Ioannis Kakavas | 27 August 2020, 09:42:26 UTC | [6.8] Add SAML AuthN request signing tests (#48444) (#61586) - Add a unit test for our signing code - Change SAML IT to use signed authentication requests for Shibboleth to consume Backport of #48444 | 27 August 2020, 09:42:26 UTC |
37825c3 | David Turner | 20 August 2020, 05:44:17 UTC | Add note on the danger of closed indices (#61332) Prior to #33888 it is dangerous to keep closed indices in your cluster long-term: Elasticsearch does not maintain their shard copies so they tend to get lost as the cluster migrates to new nodes. This risk isn't documented today. This commit addresses that gap. | 20 August 2020, 05:44:17 UTC |
fa39e7d | James Rodewig | 19 August 2020, 16:56:57 UTC | [DOCS] Document empty string boolean value as `false` (#61341) | 19 August 2020, 17:06:14 UTC |
a6348af | Jacob Dreesen | 19 August 2020, 13:31:43 UTC | [DOCS] Fix indentation in nested query (#61270) | 19 August 2020, 13:31:43 UTC |
3c690e6 | Lisa Cawley | 18 August 2020, 16:42:04 UTC | [DOCS] Add security updates to release notes (#61287) | 18 August 2020, 16:42:04 UTC |
f64fbbb | Dimitrios Liappis | 18 August 2020, 14:01:23 UTC | Bump current version to 6.8.13 (#61197) | 18 August 2020, 14:01:23 UTC |
427ec41 | Jay Modi | 14 August 2020, 01:28:34 UTC | Canonicalize client name in krb5kdc-fixture (#61123) This commit changes the value for client name canonicalization to true in the krb5.conf template file. This is done as a means to workaround JDK-8246193 which has made it into some builds of JDK8. Closes #61050 Backport of #61119 | 14 August 2020, 01:28:34 UTC |
e079ebd | Ryan Ernst | 13 August 2020, 17:28:33 UTC | Upload sysout/syserr if test jvm crashes (#61071) This commit adds the sysout and syserr files left behind when a test jvm crashes. These files are deleted upon successful test execution, so they will only exist when a failure occurred. | 13 August 2020, 17:28:33 UTC |
56a4b15 | Ryan Ernst | 13 August 2020, 17:22:02 UTC | Ensure hotthreads do not produce node failures (#61073) This commit adds an assertion that no sub-nodes requests within hot threads failed. relates #58842 | 13 August 2020, 17:23:10 UTC |
2569ec0 | Dimitrios Liappis | 12 August 2020, 15:31:07 UTC | Add release notes for 6.8.12 (#61036) | 12 August 2020, 15:31:07 UTC |
676f15d | Yang Wang | 12 August 2020, 06:34:05 UTC | Mute kerberos tests for jdk 8u[262,271) (#60995) (#60998) The Kerberos bug (JDK-8246193) is introduced in JDK 8u262 and fixed in 8u271. This PR mute for any possible releases between these two versions. | 12 August 2020, 06:34:05 UTC |
7a15d2a | Nhat Nguyen | 11 August 2020, 22:34:28 UTC | Handle empty input in AddStringKeyStoreCommand (#39490) This change ensures that we do not make assumptions about the length of the input that we can read from the stdin. It still consumes only one line, as the previous implementation Co-authored-by: Ioannis Kakavas <ikakavas@protonmail.com> | 11 August 2020, 22:34:28 UTC |
9505212 | Nhat Nguyen | 04 August 2020, 03:09:52 UTC | Increase timeout in testFollowIndexWithConcurrentMappingChanges (#60534) The test failed because the leader was taking a lot of CPUs to process many mapping updates. This commit reduces the mapping updates, increases timeout, and adds more debug info. Closes #59832 | 11 August 2020, 21:16:21 UTC |
e61c66c | Nhat Nguyen | 31 July 2020, 13:48:16 UTC | Fix synchronization in ShardFollowNodeTask (#60490) The leader mapping, settings, and aliases versions in a shard follow-task are updated without proper synchronization and can go backward. | 11 August 2020, 18:51:26 UTC |
54f1e4a | debadair | 10 August 2020, 20:27:57 UTC | [DOCS] Fix download links for tar.gz archive. (#60854) | 10 August 2020, 20:27:57 UTC |
65a45d6 | Lisa Cawley | 10 August 2020, 14:50:50 UTC | [DOCS] Add ml-cpp PRs to release notes (#60839) | 10 August 2020, 14:50:50 UTC |
8dbcd49 | Albert Zaharovits | 05 August 2020, 13:49:11 UTC | Use the Index Access Control from the scroll search context (#60640) When the RBACEngine authorizes scroll searches it sets the index access control to the very limiting IndicesAccessControl.ALLOW_NO_INDICES value. This change will set it to the value for the index access control that was produced during the authorization of the initial search that created the scroll, which is now stored in the scroll context. | 05 August 2020, 13:49:11 UTC |
e38ae41 | debadair | 03 August 2020, 19:36:30 UTC | [DOCS] Add info about why we removed test fw docs (#60346) (#60620) * [DOCS] Add info about why we removed test fw docs * Apply suggestions from code review Co-authored-by: James Rodewig <40268737+jrodewig@users.noreply.github.com> Co-authored-by: James Rodewig <40268737+jrodewig@users.noreply.github.com> | 03 August 2020, 19:36:30 UTC |
dfd91e0 | Yang Wang | 31 July 2020, 00:11:43 UTC | Fix SimpleKdcLdapServerTests by overriding java.locale.providers (#60447) The JDK system property makes sure date digits are always represented as ASCII chars. This in turn helps the Kerberos response to be correctly encoded and recognised by the client. Resolves: #57749 | 31 July 2020, 00:11:43 UTC |
0299031 | Yang Wang | 29 July 2020, 06:23:18 UTC | Mute kerberos tests for jdk 8u262 (#60353) The JDK bug (https://bugs.openjdk.java.net/browse/JDK-8246193) is fixed since b26. The tests can be unmuted since we are already using b33. However the same bug is now affecting jdk 8u262, which is the base for current Zulu jdk 8.48. This PR mute the tests for this specific jdk version. Relates: #56507 | 29 July 2020, 06:23:18 UTC |
76bf8d2 | Mark Vieira | 28 July 2020, 19:39:51 UTC | Update IntelliJ compiler options to support version 2020.2 (#60322) Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> | 28 July 2020, 19:43:54 UTC |
673dd65 | Nhat Nguyen | 28 July 2020, 16:25:12 UTC | Ensure listener is done in testPendingConnectListeners We need to ensure that we do not fail on InterruptedException in the connection listener as this can happen when the connection is closed. Closes #59879 | 28 July 2020, 16:25:12 UTC |
50fea0d | weizijun | 27 July 2020, 17:51:21 UTC | Fix wait_for_no_initializing_shards params (#58379) | 27 July 2020, 18:37:10 UTC |
370d261 | Nhat Nguyen | 27 July 2020, 17:49:51 UTC | Set timeout of master requests on follower to unbounded (#60070) Today, a follow task will fail if the master node of the follower cluster is temporarily overloaded and unable to process master node requests (such as update mapping, setting, or alias) from a follow-task within the default timeout. This error is transient, and follow-tasks should not abort. We can avoid this problem by setting the timeout of master node requests on the follower cluster to unbounded. Closes #56891 | 27 July 2020, 17:54:32 UTC |
8f26836 | Nhat Nguyen | 27 July 2020, 13:26:51 UTC | Relax ShardFollowTasksExecutor validation (#60054) If a primary shard of a follower index is being relocated, then we will fail to create a follow-task. This validation is too restricted. We should ensure that all primaries of the follower index are active instead. Closes #59625 | 27 July 2020, 17:54:32 UTC |
f964cc6 | Nhat Nguyen | 24 July 2020, 02:30:48 UTC | Increase timeout in AutoFollowIT (#60004) It can take more than 10 seconds to auto-follow and create a follow-task on a slow CI. This commit increases timeout in AutoFollowIT by replacing assertBusy with assertLongBusy. Closes #59952 | 27 July 2020, 17:54:32 UTC |
66d5a3c | Nhat Nguyen | 20 July 2020, 12:36:12 UTC | Increase timeout testFollowIndexWithConcurrentMappingChanges Fixes #59273 | 27 July 2020, 17:54:32 UTC |
83d108d | Adrien Grand | 27 July 2020, 13:01:49 UTC | Add 6.8.12 version constant. | 27 July 2020, 17:26:22 UTC |
625a433 | lcawl | 27 July 2020, 15:08:37 UTC | [DOCS] Removes coming tag | 27 July 2020, 15:08:37 UTC |
04f24ac | Christoph Büscher | 21 July 2020, 12:52:41 UTC | Muting RemoteClusterConnectionTests#testPendingConnectListeners | 21 July 2020, 12:52:41 UTC |
3dd0f5d | Yannick Welsch | 21 July 2020, 11:55:11 UTC | CCR recoveries using wrong setting for chunk sizes (#59597) The default chunk size for CCR file-based recoveries was wrongly set to 40MB instead of 1MB. | 21 July 2020, 11:57:36 UTC |
9995ab3 | lcawl | 16 July 2020, 16:14:37 UTC | [DOCS] Re-adds coming macro in release notes | 16 July 2020, 16:14:37 UTC |
2bf742a | Adrien Grand | 16 July 2020, 07:23:54 UTC | Add 6.8.11 release highlights. (#59616) | 16 July 2020, 07:23:54 UTC |
5fe92df | Adrien Grand | 16 July 2020, 07:22:59 UTC | Add release notes for 6.8.11. (#59401) | 16 July 2020, 07:22:59 UTC |
9daf2d3 | Yannick Welsch | 15 July 2020, 07:37:34 UTC | Wait for active license in CcrRestIT (#59543) Relates #53966 Closes #59486 | 15 July 2020, 07:40:05 UTC |
00bf386 | Nhat Nguyen | 08 July 2020, 03:32:14 UTC | Throw ESRejected*Exception when too many pending connect listeners (#59036) The backport in #56073 was supposed to change the max pending listeners to 1000 and throw ESRejectedExecutionException instead of RejectedExecutionException when reaching that limit. However, it missed the latter. Closes #53225 | 08 July 2020, 03:32:14 UTC |
7038525 | Lisa Cawley | 07 July 2020, 18:36:50 UTC | [DOCS] Adds ml-cpp PRs to release notes (#59170) | 07 July 2020, 18:36:50 UTC |
35a8445 | Nhat Nguyen | 02 July 2020, 18:34:19 UTC | Avoid flipping translog header version (#58866) An old translog header does not have a checksum. If we flip the header version of an empty translog to the older version, then we won't detect that corruption, and translog will be considered clean as before. Closes #58671 | 06 July 2020, 17:19:31 UTC |
5c71f63 | Przemyslaw Gomulka | 03 July 2020, 11:46:59 UTC | Adding a skip section for a new week based yml tests (#59003) | 03 July 2020, 11:46:59 UTC |
e49df55 | Przemyslaw Gomulka | 02 July 2020, 20:06:15 UTC | Index name expression resolver bwc layer for date parsing (#58503) backports #34507 closes #58481 also closes general parsing issue (no date processors) #58602 | 02 July 2020, 20:06:15 UTC |
fd5a708 | Przemyslaw Gomulka | 02 July 2020, 08:19:46 UTC | [6.8] Week based parsing for ingest date processor backport(#58597) (#58805) Date processor was incorrectly parsing week based dates because when a weekbased year was provided ingest module was thinking year field was not present on a date and was trying to applying the logic for dd/MM type of dates. Date Processor is also allowing users to specify locale parameter. It should be taken into account when parsing dates - currently only used for formatting. If someone specifies 'en-us' locale, then calendar data rules for that locale should be used. The iso8601 ingest format is still using joda implementation closes #58479 | 02 July 2020, 08:19:46 UTC |
138e3c7 | Nhat Nguyen | 01 July 2020, 15:42:06 UTC | Ensure CCR partial reads never overuse buffer (#58620) When the documents are large, a follower can receive a partial response because the requesting range of operations is capped by max_read_request_size instead of max_read_request_operation_count. In this case, the follower will continue reading the subsequent ranges without checking the remaining size of the buffer. The buffer then can use more memory than max_write_buffer_size and even causes OOM. Backport of #58620 | 01 July 2020, 15:42:06 UTC |