Commit Graph

19723 Commits

Author SHA1 Message Date
Isabel Drost-Fromm 4c9e1361fc Merge pull request #16498 from MaineC/test-fix/sort-messy
Move sorting tests w/o scripting back to core
2016-02-08 15:29:20 +01:00
Isabel Drost-Fromm 0b3cab6da3 Move sorting tests w/o scripting back to core
In preparation of #16127 this moves sorting tests back to core
that don't actually need scripting or groovy to work.
2016-02-08 14:45:33 +01:00
Boaz Leskes 4decc72da6 Fix recovery translog stats totals when recovering from store
Recovery from store fails to correctly set the translog recovery stats. This fixes it and tightens up the logic bringing it all to IndexShard (previously it was set by the recovery logic).

Closes #15974
Closes #16493
2016-02-08 11:36:59 +01:00
Simon Willnauer 7bfb1dfcc1 [TEST] Don't assert on null value. It's fine to not always see an exception in this part of the test. 2016-02-07 19:58:03 +01:00
Jason Tedor 16fe948a15 Objects#requireNonNull guard in MessageDigests 2016-02-07 11:20:53 -05:00
Jason Tedor d0c03b26f6 Fix trace logging statement in ZenDiscovery 2016-02-06 09:17:49 -05:00
Jason Tedor 542373a779 Avoid cloning MessageDigest instances
This commit modifies the MessageDigests message digest provider to
return a thread local instance of MessageDigest instances instead of
using clone since some providers do not support clone.

Closes #16479
2016-02-06 09:10:31 -05:00
Jack Conradson f6f2d40fd5 Minor clean up.
* Minor clean up of Writer constants.
* Removed synthetic attribute from the generated constructor and method.
* Added a safeguard for maximum script length.

Closes #16457
2016-02-05 12:00:18 -08:00
Simon Willnauer f1d1c3c79f Merge pull request #16470 from s1monw/make_indices_warmer_private
Make IndicesWarmer a private class of IndexService
2016-02-05 15:53:05 +01:00
Simon Willnauer b185209b63 Make IndicesWarmer a private class of IndexService
There is no need for IndicesWarmer to be a global accessible class. All it needs
access to is inside IndexService. It also doesn't need to be mutable once it's not a per node
instance. This commit move IndicesWarmer to IndexWarmer and makes the default impls like field data and
norms warming an impl detail. Also the IndexShard doesn't depend on this class anymore, instead it accepts
an Engine.Warmer as a ctor argument which delegates to the actual warmer from the index.
2016-02-05 15:24:17 +01:00
Clinton Gormley 0c41a68690 Fixed asciidoc issues 2016-02-05 14:47:44 +01:00
Simon Willnauer 778ec84a7c Merge pull request #16469 from s1monw/simplify_indices_field_data_cache
Simplify IndicesFieldDataCache and detach from guice
2016-02-05 14:24:19 +01:00
Simon Willnauer 9b2f40f627 Merge pull request #16471 from s1monw/no_sys_props_leaked
Apply system properties after all arguemnts are parsed in BootstrapCLIParser
2016-02-05 14:23:00 +01:00
Christoph Büscher 97172aa2fd Merge pull request #16465 from cbuescher/issue-13884-test
Tests: Add test for minimum_should_match for SimpleQueryStringQuery
2016-02-05 13:12:53 +01:00
Christoph Büscher d885fe62ea Add test for minimum_should_match, one term and multiple fields
This adds a test case similar to the issue in #13884 which was
fixed in #16155.
2016-02-05 12:33:13 +01:00
Jason Tedor 21e22a2ddc Reword cat API breaking changes 2016-02-05 06:28:39 -05:00
Andrej Kazakov 7f2b369dfd Use Accept header field in cat API
The cat API previously used the Content-Type header field for
determining the media type of the response. This is in opposition to the
HTTP spec which specifies the Accept header field for this purpose. This
commit replaces the use of the Content-Type header field with the Accept
header field in the cat API.

Closes #14421
2016-02-05 06:28:39 -05:00
Jim Ferenczi 5d23d9f034 Merge pull request #16454 from jimferenczi/rename_plugin_script
Rename bin/plugin in bin/elasticsearch-plugin
2016-02-05 11:25:00 +01:00
Simon Willnauer 1e18c9a7b6 Apply system properties after all arguemnts are parsed in BootstrapCLIParser
One of our tests leaked a system property here since we failed after appling some
system properties in BootstrapCLIParser. This is not a huge deal in production since
we exit the JVM if we fail on that. Yet for correctnes we should only apply them if
we manage to parse them all.
This also caused a test failure lately on CI but on an unrelated test:
  https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+master+periodic/314/console
2016-02-05 11:21:00 +01:00
Simon Willnauer 8685a003d9 fix another test 2016-02-05 11:17:53 +01:00
Simon Willnauer b5ef48ce57 fix more tests 2016-02-05 10:36:02 +01:00
Yannick Welsch 89a5eadfea Change path separator for Checkstyle suppressions to be Windows compatible 2016-02-05 10:18:18 +01:00
Jim Ferenczi 7d0181b5d4 Rename bin/plugin in bin/elasticsearch-plugin 2016-02-05 10:09:14 +01:00
Simon Willnauer 9d07501c96 fix test 2016-02-05 09:54:55 +01:00
Simon Willnauer 00be9e58f2 Simplify IndicesFieldDataCache and detach from guice
Indices level field data cacheing belongs into IndicesService and doesn't need to be
wired by guice. This commit also moves the async cache refresh out of the class into
IndicesService such that threadpool dependencies are removed and testing / creation becomes
simpler.
2016-02-05 09:43:33 +01:00
Ryan Ernst 608f4f8b6e Merge pull request #16467 from rjernst/ide_build_dirs
Make IDEs use separate build dirs
2016-02-04 18:09:50 -08:00
Ryan Ernst b9342b305b Also use separte buildSrc build dir for intellij 2016-02-04 15:58:05 -08:00
Robert Muir dcec4791c4 fix eclipse to compile at all. why does the build not fail? 2016-02-04 18:57:53 -05:00
Ryan Ernst 2fef5dee96 Build: use separate build directories for eclipse and intellij
This fixes the problem where doing a gradle clean from the command line
completely breaks the IDE.
2016-02-04 15:44:48 -08:00
Martijn van Groningen 7a6adfd93a ingest: Added foreach processor.
This processor is useful when all elements of a json array need to be processed in the same way.
This avoids that a processor needs to be defined for each element in an array.
Also it is very likely that it is unknown how many elements are inside an json array.
2016-02-04 23:44:01 +01:00
Jason Tedor 79ebe0682c Fix typo in Zen Discovery docs 2016-02-04 17:12:51 -05:00
Lee Hinman 7bf5c45159 Merge pull request #16420 from dakrone/update-dir-locals
Update .dir-locals.el for enforced line length
2016-02-04 14:54:54 -07:00
Nik Everett 1f5cfd0a76 Merge pull request #16413 from nik9000/line_length_suppress
Fail build for lines longer than 140 characters
2016-02-04 16:45:54 -05:00
Ryan Ernst 4b7885969a Merge pull request #16466 from rjernst/disable_module_publishing
Do not publish modules
2016-02-04 13:34:57 -08:00
Ryan Ernst fe8c9d394d Build: Do not publish modules
Modules are an internal implementation detail of our build, and there is
no need to publish them with gradle. This change disables publishing of
all modules.
2016-02-04 13:03:03 -08:00
Nik Everett 7a2947a4ee Suppress LineLength in more files
These are files that grew long lines between when the last line length
suppression patch was proposed about a day ago.
2016-02-04 15:39:25 -05:00
Nik Everett 63aedb8720 Fail build for lines longer than 140 characters
140 characters is our official line length from CONTRIBUTING.md. It is a
little longer than fits well in github but you can use a userstyle to fix
that. It is perfect for Eclipse and unified diffs but too wide for side by
side diffs. Regardless, its the official limit we've had for years. We just
haven't enforced it automatically and we haven't enforced it using github
because line limits are hard to notice there.

This only hits about 2/3 of the java files - those that didn't have failures
already. If they did have a failure it suppresses them. We should pick files
off that list as time goes on.

For posterity I generated the suppressions by running checkstyle with
ignoreFailures = true, piping the output to a file, and then running it
through this:
perl -ne 'print if s{.*\[ant:checkstyle\] /.+/elasticsearch/}{  <suppress files="} && s{\.java.+}{\.java" checks="LineLength" />}' | uniq
2016-02-04 15:02:39 -05:00
Simon Willnauer d87625c9f0 Merge pull request #16447 from s1monw/detach_shard_from_term_vectors
Remove dependency from IndexShard on TermVectorService
2016-02-04 17:42:21 +01:00
Simon Willnauer 59a20015d6 Make TermVectorService static 2016-02-04 17:17:16 +01:00
Simon Willnauer 75802515ea Merge branch 'master' into detach_shard_from_term_vectors 2016-02-04 17:12:09 +01:00
Simon Willnauer f346d51723 Merge pull request #16452 from s1monw/drop_dist_tv_df
Remove DFS support from TermVector API
2016-02-04 16:47:42 +01:00
Simon Willnauer 3de0004ea3 add migration notice 2016-02-04 16:23:58 +01:00
Simon Willnauer 450ee70038 Remove DFS support from TermVector API
Retrieving distributed DF for TermVectors is beside it's esotheric justification
a very slow process and can cause serious load on the cluster. We also don't have nearly
enough testing for this stuff and given the complexity we should remove it rather than carrying it
around.
2016-02-04 16:20:24 +01:00
Yannick Welsch 15507580bb Merge pull request #16243 from ywelsch/fix/remove-shardrouting-version
Remove obsolete version in ShardRouting
2016-02-04 15:52:00 +01:00
Yannick Welsch 4937531a17 Remove obsolete version in ShardRouting
Closes #16243
2016-02-04 15:50:25 +01:00
Boaz Leskes 4cdab159f1 Remove plugability of the gateway
In the spirit of simplification, remove the ability to have a custom gateway. We haven't seen any valid use for it so far.

closes #16448
2016-02-04 15:45:05 +01:00
Boaz Leskes 924e880480 Remove `gateway.initial_meta` and always rely on min master nodes
During initial cluster forming, when  a master is elected, it reaches out to all other masters nodes and ask the last cluster state they persisted. To make sure we select the right state, we must successfully read from a `min_master_nodes` nodes. The gateway currently have specific settings to override this behavior, but I don't think they are ever used. We can drop them and reach out to the discovery layer, the single source of truth for the min master nodes settings.

Closes #16446
2016-02-04 13:42:28 +01:00
Simon Willnauer 30f569f958 Remove dependency from IndexShard on TermVectorService
This dependency is just syntactic sugar and complicates IndexShard creation.
This commit move it out where it's actually used and needed.
2016-02-04 12:55:45 +01:00
Simon Willnauer e69350f2ba Fix method visibility in SettingsModule.java 2016-02-04 11:53:06 +01:00
Simon Willnauer cb811a85da Add javadocs for SettingsModule 2016-02-04 11:49:59 +01:00