260387d | Deb Adair | 30 June 2017, 21:37:05 UTC | [DOCS] Removed painless topics from reference & cross referenced separate book. | 30 June 2017, 22:03:59 UTC |
6e980af | James Baiera | 30 June 2017, 21:57:56 UTC | Upgrading HDFS Repository Plugin to use HDFS 2.8.1 Client (#25497) Hadoop 2.7.x libraries fail when running on JDK9 due to the version string changing to a single character. On Hadoop 2.8, this is no longer a problem, and it is unclear on whether the fix will be backported to the 2.7 branch. This commit upgrades our dependency of Hadoop for the HDFS Repository to 2.8.1. | 30 June 2017, 22:02:19 UTC |
7f161ef | Jason Tedor | 30 June 2017, 16:59:02 UTC | Add doc note regarding explicit publish host This commit adds a note to the docs regarding explicilty setting a publish host if the network.host setting results in multiple bind addresses. Relates #25496 | 30 June 2017, 16:59:41 UTC |
59096e0 | Martijn van Groningen | 29 June 2017, 21:25:05 UTC | WrapperQueryBuilder should also rewrite the parsed query. Failing to do so can cause other errors later on during query execution. For example if `WrapperQueryBuilder` wraps a `GeoShapeQueryBuilder` that fetches the shape from an index then it will skip the shape fetching and fail later with the error that no shapes have been fetched. | 30 June 2017, 11:56:27 UTC |
fcba345 | Russ Cam | 30 June 2017, 06:24:53 UTC | Update defaults in documentation (#25483) Following internal discussion Relates: https://github.com/elastic/windows-installers/pull/49 (cherry picked from commit 099ccd36eee6903805c2f885d9218f22fb04c9d9) | 30 June 2017, 06:28:44 UTC |
3479574 | Lisa Cawley | 29 June 2017, 23:31:54 UTC | [DOCS] Add docs-dir to Painless (#25482) | 29 June 2017, 23:54:23 UTC |
e43669c | Lisa Cawley | 29 June 2017, 20:48:58 UTC | [DOCS] Update shared attributes for Elasticsearch (#25479) * [DOCS] Update shared attributes for Elasticsearch * [DOCS] Moved shared attributes to Versions.asciidoc * [DOCS] More more book URLs to shared attributes | 29 June 2017, 23:53:43 UTC |
0d396ff | Clinton Gormley | 29 June 2017, 17:00:32 UTC | Updated 5.5.0 release notes | 29 June 2017, 17:00:43 UTC |
841f72c | Clinton Gormley | 29 June 2017, 16:51:27 UTC | Added release notes for 5.5.0 | 29 June 2017, 16:51:27 UTC |
e65ca02 | Martijn van Groningen | 29 June 2017, 13:31:15 UTC | test: protect against empty collection | 29 June 2017, 13:31:50 UTC |
1884e34 | Martijn van Groningen | 29 June 2017, 10:37:47 UTC | docs: fix painless getting started to work with ES 5.5 Relates #25466 | 29 June 2017, 10:37:47 UTC |
935fec9 | Jason Tedor | 29 June 2017, 02:18:46 UTC | Emit settings deprecation logging at most once When a setting is deprecated, if that setting is used repeatedly we currently emit a deprecation warning every time the setting is used. In cases like hitting settings endpoints over and over against a node with a lot of deprecated settings, this can lead to excessive deprecation warnings which can crush a node. This commit ensures that a given setting only sees deprecation logging at most once. Relates #25457 | 29 June 2017, 02:19:56 UTC |
8165ed0 | DeDe Morton | 28 June 2017, 23:17:53 UTC | [Docs] Add link to grok debugger docs (#25460) | 28 June 2017, 23:17:53 UTC |
3afbafa | Ali Beyad | 28 June 2017, 22:37:56 UTC | Output all empty snapshot info fields if in verbose mode (#25455) In #24477, a less verbose option was added to retrieve snapshot info via GET /_snapshot/{repo}/{snapshots}. The point of adding this less verbose option was so that if the repository is a cloud based one, and there are many snapshots for which the snapshot info needed to be retrieved, then each snapshot would require reading a separate snapshot metadata file to pull out the necessary information. This can be costly (performance and cost) on cloud based repositories, so a less verbose option was added that only retrieves very basic information about each snapshot that is all available in the index-N blob - requiring only one read! In order to display this less verbose snapshot info appropriately, logic was added to not display those fields which could not be populated. However, this broke integrators (e.g. ECE) that required these fields to be present, even if empty. This commit is to return these fields in the response, even if empty, if the verbose option is set. | 28 June 2017, 22:38:42 UTC |
415b81e | Jay Modi | 28 June 2017, 21:50:24 UTC | Fix race condition in RemoteClusterConnection node supplier (#25432) This commit fixes a race condition in the node supplier used by the RemoteClusterConnection. The node supplier stores an iterator over a set backed by a ConcurrentHashMap, but the get operation of the supplier uses multiple methods of the iterator and is suceptible to a race between the calls to hasNext() and next(). The test in this commit fails under the old implementation with a NoSuchElementException. This commit adds a wrapper object over a set and a iterator, with all methods being synchronized to avoid races. Modifications to the set result in the iterator being set to null and the next retrieval creates a new iterator. | 28 June 2017, 22:00:34 UTC |
4ca1567 | Deb Adair | 28 June 2017, 20:33:04 UTC | [DOCS] Adding Painless doc to the version branches. | 28 June 2017, 20:33:04 UTC |
2c58ef5 | Deb Adair | 28 June 2017, 20:07:03 UTC | [DOCS] Bumped major version and branch to 5.5 | 28 June 2017, 20:07:27 UTC |
1a43e4d | Jay Modi | 28 June 2017, 18:41:37 UTC | Do not search locally if remote index pattern resolves to no indices (#25436) This commit changes how we determine if there were any remote indices that a search should have been executed against. Previously, we used the list of remote shard iterators but if the remote index pattern resolved to no indices there would be no remote shard iterators even though the request specified remote indices. The map of remote cluster names to the original indices is used instead so that we can determine if there were remote indices even when there are no remote shard iterators. Closes #25426 | 28 June 2017, 19:15:12 UTC |
d71eada | Ryan Ernst | 28 June 2017, 19:09:36 UTC | Build: Remove duplicate verifyVersions (#25433) The verifyVersions in core was removed in master, in favor of that in the root build which only runs with the branchConsistency task. This commit removes the legacy task. | 28 June 2017, 19:10:05 UTC |
e08526a | Martijn van Groningen | 28 June 2017, 15:10:30 UTC | docs: added percolator script query limitation | 28 June 2017, 15:11:23 UTC |
b71f3c0 | Chris Earle | 28 June 2017, 14:08:45 UTC | _nodes/stats should not fail due to concurrent AlreadyClosedException (#25016) This catches `AlreadyClosedException` during `stats` calls to avoid failing a `_nodes/stats` request because of the ignorable, concurrent index closure. | 28 June 2017, 14:29:10 UTC |
8947cb3 | Robin Clarke | 28 June 2017, 12:31:03 UTC | Docs: Fix typo for request cache (#25444) | 28 June 2017, 12:33:15 UTC |
4648326 | Marcus Wittig | 28 June 2017, 10:46:06 UTC | [DOCS] reworded to prevent code span rendering glitch (#25442) Changed `rescore`s to `rescore` requests as an backtick followed by the s character appears to be interpreted as an apostrophe which then leads to an unbalanced backtick for the next code span in the remainder of the paragraph Closes #25443 | 28 June 2017, 10:48:10 UTC |
cd2d641 | Daniel Mitterdorfer | 28 June 2017, 05:53:16 UTC | Add backwards compatibility indices for 5.4.3 | 28 June 2017, 06:03:22 UTC |
5e98644 | Daniel Mitterdorfer | 28 June 2017, 06:03:11 UTC | Add version 5.4.3 and 5.4.4 (unreleased) after release | 28 June 2017, 06:03:11 UTC |
00424a2 | Russ Cam | 27 June 2017, 23:04:23 UTC | Update MSI installer images (#25414) Slight updating to styling (cherry picked from commit 2804fc4c29170829d621f2760206160d10fddfa5) | 27 June 2017, 23:08:15 UTC |
d0c9dbe | Deb Adair | 27 June 2017, 00:48:56 UTC | [DOCS] Fixed broken cross doc links to security settings. | 27 June 2017, 17:18:41 UTC |
f9b4661 | Clinton Gormley | 27 June 2017, 10:33:46 UTC | Added release notes for 5.4.3 | 27 June 2017, 15:52:43 UTC |
e432e22 | Lisa Cawley | 27 June 2017, 15:33:28 UTC | [DOCS] Update docs to use shared attribute file (#25403) * [DOCS] Update docs to use shared attribute file * [DOCS] Add shared attributes to Versions.asciidoc | 27 June 2017, 15:43:05 UTC |
998875b | dkimdon | 23 June 2017, 08:38:04 UTC | Update percolate-query.asciidoc (#25364) | 23 June 2017, 08:39:20 UTC |
47f0df7 | Jason Tedor | 23 June 2017, 01:53:16 UTC | Fix use of spaces on Windows if JAVA_HOME not set When JAVA_HOME is not set we try to detect the location of Java. If its location contains a space, due to a lack of quoting we will be unsuccessful in invoking Java. This commit adds the necessary quoting to handle this case. Relates #23822 | 23 June 2017, 01:56:32 UTC |
507418b | Boaz Leskes | 22 June 2017, 09:50:19 UTC | testMockFailToSendNoConnectRule should wait for node disconnect event to be processed Those are performed in an async fashion Closes #25347 | 22 June 2017, 09:50:19 UTC |
64dc637 | Deb Adair | 21 June 2017, 21:27:30 UTC | [DOCS] Fixed typo. | 21 June 2017, 21:35:27 UTC |
f1228db | Spencer | 21 June 2017, 20:44:27 UTC | [rest-api-spec/indices.refresh] Remove old params Fixes #25234 (cherry picked from commit c5b79cd46021075e3696eee3b6bae45d71af31fa) | 21 June 2017, 20:45:04 UTC |
280014a | joachimdraeger | 21 June 2017, 19:41:17 UTC | Remove redundant and broken MD5 checksum from repository-s3 (#25270) Remove redundant and not resettable (fails on retries) check-summing. Checksums are calculated and compared by the S3 client already. Closes #25269 | 21 June 2017, 20:16:59 UTC |
c651c89 | Boaz Leskes | 21 June 2017, 08:27:57 UTC | Update MockTransportService to the age of Transport.Connection (#25320) MockTransportServices allows us to simulate network disruptions in our testing infra. Sadly it wasn't updated to the state of the art in Transport land. This PR brings it up to speed. Specifically: 1) Opening a connection is now also blocked (before only node connections were blocked) 2) Simplifies things using the latest connection based notification between TcpTransport and TransportService for when a disconnect happens. 3) By 2, it fixes a race condition where we may fail to respond to a sent request when it is sent concurrently with the closing of a connection. The old code relied on a node based bridge between tcp transport and transport service. Sadly, the following doesn't work any more: ``` if (transport.nodeConnected(node)) { // this a connected node, disconnecting from it will be up the exception transport.disconnectFromNode(node); <-- this may now be a noop and it doesn't mean that the transport service was notified of the disconnect between the nodeConnected check and here. } else { throw new ConnectTransportException(node, reason, e); } ``` | 21 June 2017, 13:09:42 UTC |
2468f18 | Tanguy Leroux | 21 June 2017, 12:37:18 UTC | ... and the missing Version.fromId() | 21 June 2017, 12:37:18 UTC |
50534b5 | Tanguy Leroux | 21 June 2017, 12:31:34 UTC | Add 5.4.3 unreleased version constant | 21 June 2017, 12:31:34 UTC |
1694a7f | Tanguy Leroux | 21 June 2017, 09:42:09 UTC | Add backward compatibility indices for 5.4.2 | 21 June 2017, 09:42:09 UTC |
6d5cdc5 | Tanguy Leroux | 21 June 2017, 09:41:11 UTC | Add version v5.4.2 after release | 21 June 2017, 09:41:11 UTC |
ffc519a | Nathan Taylor | 21 June 2017, 08:47:19 UTC | Docs: Removed duplicated line in mapping docs | 21 June 2017, 08:55:21 UTC |
b1c1877 | Clinton Gormley | 20 June 2017, 13:04:47 UTC | Added release notes for 5.4.2 | 20 June 2017, 13:04:56 UTC |
622ff76 | Russ Cam | 19 June 2017, 04:52:47 UTC | Add MSI installation to documentation (#25213) * Add MSI installation to documentation Move installation documentation for Windows with the .zip archive into the zip and tar installation documentation, and clearly indicate any differences for installing on macOS/Linux and Windows. * Separate out installation with .zip on Windows (cherry picked from commit 9eca380a76a92c6ede955e863d1a9eed895fd930) | 20 June 2017, 05:55:31 UTC |
b58895a | debadair | 15 June 2017, 15:54:10 UTC | [DOCS] Defined es-test-dir and plugins-examples-dir in index.asciidoc. (#25232) Use these attributes when specifying the location of included tests. | 20 June 2017, 00:36:36 UTC |
bc73745 | lcawley | 19 June 2017, 23:17:00 UTC | [DOCS]Re-add painless, omitted by mistake | 19 June 2017, 23:17:00 UTC |
ca4df95 | Lisa Cawley | 19 June 2017, 22:14:53 UTC | [DOCS] Split index-shared.asciidoc into multiple smaller files (#25302) | 19 June 2017, 22:27:18 UTC |
40df9b6 | Jay Modi | 19 June 2017, 18:43:26 UTC | Keystore CLI should use the AddFileKeyStoreCommand for files (#25298) This commit fixes a typo in the KeyStoreCli class. The add-file command was incorrectly set to use the AddStringKeyStoreCommand instead of the AddFileKeyStoreCommand. | 19 June 2017, 20:08:53 UTC |
8ec148d | Nik Everett | 19 June 2017, 17:55:32 UTC | Fix artifact location Fix the location of the rpm and deb used for the packaging upgrade tests when upgrading from -SNAPSHOT version. | 19 June 2017, 18:22:00 UTC |
85ada96 | David Kyle | 19 June 2017, 15:37:21 UTC | Initialise empty lists in BaseTaskResponse constructor (#25290) * Initialise empty lists in BaseTaskResponse constructor * Remove little used default constructor which leaves uninitialised members | 19 June 2017, 16:06:32 UTC |
53fbd70 | Jason Tedor | 17 June 2017, 01:14:34 UTC | Fix queries in cross-cluster search docs This commit fixes two queries in the cross-cluster search docs; they were missing the query object wrapping the actual query. Relates #25282 | 17 June 2017, 01:16:45 UTC |
500a9f6 | Ali Beyad | 16 June 2017, 20:57:01 UTC | Adds AwaitsFix on snapshot test failing due to #25281 | 16 June 2017, 20:58:20 UTC |
662008c | Ali Beyad | 15 June 2017, 23:43:19 UTC | Improves snapshot logging and snapshoth deletion error handling (#25264) This commit does two things: 1. Adds logging at the DEBUG level for when the index-N blob is updated. 2. When attempting to delete a snapshot, if the snapshot was not found in the repository data, an exception is now thrown instead of silently ignoring the lack of presence of the snapshot in the repository data. | 15 June 2017, 23:51:26 UTC |
ca5ad31 | Guillaume Le Floch | 15 June 2017, 16:41:05 UTC | Deprecate tribe service This commit deprecates the tribe service so that deprecation log messages are delivered if a tribe node is configured. Relates #24598 | 15 June 2017, 17:17:17 UTC |
b8fe621 | Boaz Leskes | 15 June 2017, 11:24:07 UTC | move assertBusy to use CheckException (#25246) We use assertBusy in many places where the underlying code throw exceptions. Currently we need to wrap those exceptions in a RuntimeException which is ugly. | 15 June 2017, 13:23:49 UTC |
a3869f9 | Adrien Grand | 14 June 2017, 14:31:16 UTC | Make sure range queries are correctly profiled. (#25108) We introduced a new API for ranges in order to be able to decide whether points or doc values would be more appropriate to execute a query, but since `ProfileWeight` does not implement this API, the optimization is disabled when profiling is enabled. | 14 June 2017, 14:38:52 UTC |
9212cea | Jay Modi | 14 June 2017, 14:21:56 UTC | Test: allow setting socket timeout for rest client (#25221) In #25201, a setting was added to allow setting the retry timeout for the rest client under the impression that this would allow requests to go longer than 30s. However, there is also a socket timeout that needs to be set to greater than 30s, which this change adds a setting for. | 14 June 2017, 14:27:35 UTC |
ae5b19e | Nicholas Knize | 13 June 2017, 21:49:00 UTC | Fix GeoPoint FieldStats ternary logic bug The ternary logic in prefixCodedToGeoPoint was reversed such that numericEncoded GeoPoints were using the decoding logic for GeoCoded points and vice versa. This commit fixes that boneheaded bug. | 14 June 2017, 12:36:19 UTC |
5bd427d | Russ Cam | 14 June 2017, 07:37:20 UTC | Remove mention of params and lang on update action payload (#21364) Both `params` and `lang` are properties on the object assigned to the `script` property on the payload, not top level properties themselves. Remove the mention of them for clarity. (cherry picked from commit d5a225d6504f1d1f5a821bc4fdef27b8b2e16575) (cherry picked from commit 84428da39221087bddd9100c0feb3d75e9412ed9) | 14 June 2017, 09:40:32 UTC |
cbbd6b0 | António Ribeiro | 24 May 2017, 09:43:54 UTC | Fix link to perl docs (#24842) * Fixes Elasticsearch issue #24606. * Fixes Elasticsearch issue #24606. * Fixes Elasticsearch issue #24606. * Fixes Elasticsearch issue #24606. * Issue #24606 - Changed the link text to Search::Elasticsearch::Client::5_0::Bulk and Search::Elasticsearch::Client::5_0::Scroll. | 14 June 2017, 08:36:14 UTC |
5bae1cb | David Roberts | 14 June 2017, 08:23:41 UTC | When stopping via systemd only kill the JVM, not its control group (#25195) This prevents possible race conditions between the Elasticsearch JVM and plugin native controller processes that can cause the Elasticsearch shutdown to hang. The problem can happen when the JVM and the controller process receive a SIGTERM at almost the same time. (There's an assumption here that Elasticsearch will continue to use other mechanisms to kill native controller processes.) | 14 June 2017, 08:25:08 UTC |
aa1aad9 | Ryan Ernst | 14 June 2017, 04:04:16 UTC | Docs: Add note about which secure settings are valid (#25212) This commit adds a note to the docs to clarify that only some settings can be used with the keystore. | 14 June 2017, 04:05:03 UTC |
9d29b52 | Boaz Leskes | 13 June 2017, 21:42:55 UTC | Indices.rollover/10_basic should refresh to make the doc visible in lucene stats | 13 June 2017, 21:47:22 UTC |
6e6ac0d | Lisa Cawley | 13 June 2017, 21:03:42 UTC | [DOCS] Add ML node to node.asciidoc (#24495) * [DOCS] Add ML node to node.asciidoc * [DOCS] Clarify ML node in node.asciidoc * [DOCS] Add X-Pack icon for admonition blocks * [DOCS] Formatting X-Pack blocks in node.asciidoc * [DOCS] Add xpack icon images to node.asciidoc * [DOCS] Add final xpack role attributes * [DOCS] Remove unnecssary xpackicon image * [DOCS] Add link to X-Pack node settings * [DOCS] Fix path to X-Pack repository * [DOCS] Add links to X-Pack node settings * [DOCS] Fixed text for links to X-Pack node settings * [DOCS] Change standalone node to dedicated node | 13 June 2017, 21:16:17 UTC |
f876a61 | Boaz Leskes | 13 June 2017, 12:46:15 UTC | Adapt skip version in rest-api-spec/test/indices.rollover/20_max_doc_condition.yml The relevant change was backported. | 13 June 2017, 21:09:07 UTC |
f9f950e | Jay Modi | 13 June 2017, 18:19:17 UTC | Test: add setting to change request timeout for rest client (#25201) This commit adds a setting to change the request timeout for the rest client. This is useful as the default timeout is 30s, which is also the same default for calls like cluster health. If both are the same then the response from the cluster health api will not be received as the client usually times out first making test failures harder to debug. Relates #25185 | 13 June 2017, 18:48:43 UTC |
23d4b54 | Jason Tedor | 13 June 2017, 18:43:38 UTC | Set JDK version in docs to latest This commit bumps the JDK version referenced in the docs to the latest version as of the time of this writing. | 13 June 2017, 18:44:57 UTC |
365da1e | Jason Tedor | 13 June 2017, 17:26:48 UTC | Fix secure repository-hdfs tests on JDK 9 The secure repository-hdfs tests fail on JDK 9 because some Hadoop code reaches into sun.security.krb5. This commit adds the necessary flags to open the java.security.jgss module. Note that these flags are actually needed at runtime as well when using secure repository-hdfs. For now we will punt on how best to help users obtain this when running on JDK 9 with this plugin. Relates #25205 | 13 June 2017, 17:36:03 UTC |
d54c667 | Sergey Galkin | 13 June 2017, 12:30:46 UTC | Rollover max docs should only count primaries (#24977) max_doc condition for index rollover should use document count only from primary shards Fixes #24217 | 13 June 2017, 12:39:32 UTC |
a727193 | Simon Willnauer | 13 June 2017, 07:37:05 UTC | Ensure pending transport handlers are invoked for all channel failures (#25150) Today if a channel gets closed due to a disconnect we notify the response handler that the connection is closed and the node is disconnected. Unfortunately this is not a complete solution since it only works for published connections. Connections that are unpublished ie. for discovery can indefinitely hang since we never invoke their handers when we get a failure while a user is waiting for the response. This change adds connection tracking to TcpTransport that ensures we are notifying the corresponding connection if there is a failure on a channel. | 13 June 2017, 08:10:53 UTC |
1a9de17 | Jason Tedor | 13 June 2017, 01:35:56 UTC | Use synchronized Wildfly shutdown We need to use the variant of shutdown that blocks until the connection to Wildfly is closed or we can get spurious build failures. | 13 June 2017, 01:40:16 UTC |
856a02f | Russ Cam | 13 June 2017, 00:03:18 UTC | Supported Azure Storage account types (#25167) * Supported Azure Storage account types Add important note for Azure Storage account types Relates #20844 (cherry picked from commit a0f50e8aa41fcdc7676aea95d19ca09bce040438) | 13 June 2017, 00:08:19 UTC |
cc56b54 | Russ Cam | 12 June 2017, 23:45:43 UTC | Add half_float and scaled float (#22988) to numeric datatypes (cherry picked from commit 67ea06145a80d5ec52ba55d1f2e1e8287e1882b1) | 12 June 2017, 23:53:49 UTC |
02e9613 | Lisa Cawley | 12 June 2017, 20:43:06 UTC | [DOC] Add X-Pack links to Elasticsearch Reference (#25164) * [DOC] Add X-Pack links to Elasticsearch Reference * [DOCS] Address alignment of attributes in Versions.asciidoc | 12 June 2017, 21:06:33 UTC |
5510c43 | spalger | 12 June 2017, 19:39:23 UTC | [rest-api-spec] fix remote.info doc link | 12 June 2017, 19:39:23 UTC |
1ddbdbe | spalger | 12 June 2017, 19:39:15 UTC | [docs] add remote-info.asciidoc to cluster index | 12 June 2017, 19:39:15 UTC |
35fa256 | Spencer | 12 June 2017, 19:34:16 UTC | [5.5] [rest-api-spec] point links to 5.5 docs (#25182) | 12 June 2017, 19:34:16 UTC |
8da7cc8 | Jason Tedor | 12 June 2017, 16:55:19 UTC | Explicitly reject duplicate data paths Duplicate data paths already fail to work because we would attempt to take out a node lock on the directory a second time which will fail after the first lock attempt succeeds. However, how this failure manifests is not apparent at all and is quite difficult to debug. Instead, we should explicitly reject duplicate data paths to make the failure cause more obvious. Relates #25178 | 12 June 2017, 17:00:46 UTC |
3bcf8ab | Jason Tedor | 12 June 2017, 15:52:05 UTC | Fix compilation in NodeEnvironment This commit fixes compilation after backport removed a local variable that was no longer needed in the source branch but is still needed in the target branches of the backport. | 12 June 2017, 15:53:45 UTC |
2dbf98a | Jason Tedor | 12 June 2017, 15:42:45 UTC | Do not swallow node lock failed exception When attempting to obtain the node lock, if an exception is thrown it is not logged. This makes debugging difficult. This commit causes such an exception to be logged. Relates #25176 | 12 June 2017, 15:44:28 UTC |
d32f391 | Jason Tedor | 11 June 2017, 21:05:26 UTC | Add migration note on get alias and get mapping APIs These APIs now return 404 if any requested alias or type, respectively, does not exist. This commit adds a note to the migration docs regarding this change. | 11 June 2017, 21:05:26 UTC |
0226f82 | Jason Tedor | 11 June 2017, 21:02:11 UTC | Fix label on breaking packaging changes The label here should be 5.5 instead of 5.4 since this is the version the change was introduced in. | 11 June 2017, 21:02:11 UTC |
16e4573 | Jason Tedor | 11 June 2017, 18:58:56 UTC | Fix get mappings HEAD requests Get mappings HEAD requests incorrectly return a content-length header of 0. This commit addresses this by removing the special handling for get mappings HEAD requests, and just relying on the general mechanism that exists for handling HEAD requests in the REST layer. Relates #23192 | 11 June 2017, 19:06:25 UTC |
d528002 | Jason Tedor | 11 June 2017, 03:44:18 UTC | Fix handling of exceptions thrown on HEAD requests Today when an exception is thrown handling a HEAD request, the body is swallowed before the channel has a chance to see it. Yet, the channel is where we compute the content length that would be returned as a header in the response. This is a violation of the HTTP specification. This commit addresses the issue. To address this issue, we remove the special handling in bytes rest response for HEAD requests when an exception is thrown. Instead, we let the upstream channel handle the special case, as we already do today for the non-exceptional case. Relates #25172 | 11 June 2017, 03:52:02 UTC |
b198127 | Jason Tedor | 10 June 2017, 17:25:44 UTC | Fix comment formatting in EvilLoggerTests The comments here were formatted inconsistently so this commit fixes them. | 10 June 2017, 17:26:35 UTC |
2cb5a94 | Jason Tedor | 10 June 2017, 17:20:45 UTC | Remove unneeded weak reference from prefix logger We have a custom logger implementation known as a prefix logger that is used to write every message by the logger with a given prefix. This is useful for node-level, index-level, and shard-level messages where we want to log the node name, index name, and shard ID, respectively, if possible. The mechanism that we employ is that of a marker. Log4j has a built-in facility for managing these markers, but its effectively a memory leak because these markers are held in a map and can never be released. This is problematic for us since indices and shards do not necessarily have infinite life spans and so on a node where there are many indices being creted and destroyed, this infinite lifespan can be a problem indeed. To solve this, we use our own cache of markers. This is necessary to prevent too many instances of the marker for the same prefix from being created (just think of all the shard-level components that exist in the system), and to workaround the effective leak in Log4j. These markers are stored as weak references in a weak hash map. It is these weak references that are unneeded. When a key is removed from a weak hash map, the corresponding entry is placed on a reference queue that is eventually cleared. This commit simplifies prefix logger by removing this unnecessary weak reference wrapper. Relates #22460 | 10 June 2017, 17:21:55 UTC |
f380975 | Jason Tedor | 10 June 2017, 11:50:03 UTC | Skip verifying version constants on pre-5.0.0 This commit causes the verification of the Lucene version constants to be skipped on pre-5.0.0 versions. This is because we do not have the infrastructure in place to easily start pre-5.0.0 nodes. This is okay anyway because version constants for pre-5.0.0 are not in a state of being changed. | 10 June 2017, 11:53:03 UTC |
ef97384 | Boaz Leskes | 10 June 2017, 11:43:57 UTC | a harsher await fix for `:qa:verify-version-constants` https://github.com/elastic/elasticsearch/issues/25170 | 10 June 2017, 11:47:52 UTC |
a51a7be | Boaz Leskes | 10 June 2017, 11:31:15 UTC | await fix `:qa:verify-version-constants` https://github.com/elastic/elasticsearch/issues/25170 | 10 June 2017, 11:38:08 UTC |
2d36a97 | Jason Tedor | 10 June 2017, 02:06:51 UTC | Log checkout so SHA is known This commit changes the task type of the checkoutBwcBranch task to Exec from LoggedExec so that the output of the checkout command is shown. This enables us to see the SHA used for the checkout which can be useful when debugging a BWC break. Relates #25166 | 10 June 2017, 02:07:44 UTC |
0907e56 | Chris Earle | 09 June 2017, 18:52:42 UTC | "shard started" should show index and shard ID (#25157) When the cluster state is updated with Shard Started entries, it simply adds "shard-started" as the source of the change. This adds the index name and shard ID so that we can see who/what is spamming the changes when the index creation step has already left the cluster state. | 09 June 2017, 19:13:33 UTC |
c292fdf | Adrien Grand | 09 June 2017, 12:56:23 UTC | `type` and `id` are lost upon serialization of `Translog.Delete`. (#24586) This was introduced in #24460: the constructor of `Translog.Delete` that takes a `StreamInput` does not set the type and id. To make it a bit more robust, I made fields final so that forgetting to set them would make the compiler complain. | 09 June 2017, 14:03:08 UTC |
f99485b | Martijn van Groningen | 09 June 2017, 10:48:41 UTC | nested: In case of a single type the _id field should be added to the nested document instead of _uid field. When `index.mapping.single_type` is `true` the `_uid` field is not used and instead `_id` field is used. Prior to this change nested documents would in this case still use the `_uid` field to mark to what root document they belong to. In case of deleting documents this could lead to only the root Lucene document to be deleted and not the nested Lucene documents. This broke the docid block ordering the block join relies on in order to work correctly and thus causing the `nested` query, `nested` aggregation, nested sorting and nested inner hits to either fail or yield incorrect results. This bug only manifests in 6.0.0-ALPHA2 release and snaphots (5.5.0-SNAPSHOT, 5.6.0-SNAPSHOT, 6.0.0-SNAPSHOT). | 09 June 2017, 13:04:57 UTC |
cc1fa4e | Pandiyan Murugan | 09 June 2017, 10:43:44 UTC | Fix typo in shards.asciidoc (#25143) | 09 June 2017, 10:44:12 UTC |
8c1f7ce | Sanne Grinovero | 09 June 2017, 10:01:08 UTC | List Hibernate Search (#25145) among the community framework integrations | 09 June 2017, 10:06:05 UTC |
21bab87 | javanna | 09 June 2017, 09:24:05 UTC | [DOCS] update maxRetryTimeout in java REST client usage page | 09 June 2017, 09:25:25 UTC |
a6f4a44 | Andrey Groshev | 09 June 2017, 07:50:43 UTC | Made the same length of opening and closing lines (#23583) | 09 June 2017, 07:51:41 UTC |
95972ef | Jim Ferenczi | 09 June 2017, 07:22:48 UTC | update docs that claim that classic is the default similarity | 09 June 2017, 07:24:14 UTC |
e8a40c8 | Jason Tedor | 08 June 2017, 20:09:34 UTC | Mark Log4j API dependency as non-optional The Log4j dependency is separated into two artifacts, the API and the core implementation. This is to enable replacing Log4j on the backend through the SLF4J bridge with another logging implementation. For this reason, the dependencies are marked as optional. This causes confusion amongst users as to use the bridge, the API should be non-optional since it is needed for the bridge to function correctly. While they could pull it into their application directly, it would be clearer if we simply marked this depdendency as non-optional. Note that this does not mean that users have to use Log4j for logging in their application, so we are not marking core as required, it only clarifies what they need to be able to plug in a different logging implementation. Relates #25136 | 08 June 2017, 20:12:12 UTC |
ebc6027 | Jay Modi | 08 June 2017, 16:54:33 UTC | Fix branch logic for bwc tests in the same major version (#25134) When testing against the previous 5.x release, the bwc project incorrectly would checkout the 5.x branch instead of the 5.5 branch as it still had the logic that applies for major versions bwc. This change adds a check to compare the major version when making the decision on the branch to use. | 08 June 2017, 17:43:39 UTC |
c58179d | Martijn van Groningen | 08 June 2017, 08:57:36 UTC | test: backported test from #25129 | 08 June 2017, 12:25:59 UTC |