Commit Graph

4121 Commits

Author SHA1 Message Date
Martijn van Groningen 8003171a0c
Move more token filters to analysis-common module
The following token filters were moved: arabic_normalization, german_normalization, hindi_normalization, indic_normalization, persian_normalization, scandinavian_normalization, serbian_normalization, sorani_normalization, cjk_width and cjk_width

Relates to #23658
2017-07-17 08:29:44 +02:00
Ryan Ernst 072402463b Scripting: Remove search template actions (#25717)
The dedicated search template put/get/delete actions are deprecated in
5.6. This commit removes them from 6.0.
2017-07-14 23:12:05 -07:00
Jim Ferenczi 13da3eb53e Refactor QueryStringQuery for 6.0 (#25646)
This change refactors the query_string query to analyze the query text around logical operators of the query string the same way than a match_query/multi_match_query.
It also adds a type parameter that can be used to change the way multi fields query are built the same way than a multi_match query does.

Now that these queries share the same behavior regarding text analysis, some parameters are obsolete and have been deprecated:

split_on_whitespace: This setting is now ignored with a deprecation notice
if it is used explicitely. With this PR The query_string always splits on logical operator.
It simplifies the understanding of the other parameters that can have different meanings
depending on the value of split_on_whitespace.

auto_generate_phrase_queries: This setting is now ignored with a deprecation notice
if it is used explicitely. This setting only makes sense when the parser splits on whitespace.

use_dismax: This setting is now ignored with a deprecation notice
if it is used explicitely. The tie_breaker parameter is sufficient to handle best_fields/most_fields.

Fixes #25574
2017-07-13 15:32:17 +02:00
Luca Cavanna ec66d655b5 Rename client artifacts (#25693)
It was brought up that our current client artifacts have generic names like 'rest' that may cause conflicts with other artifacts.

This commit renames:

- rest -> elasticsearch-rest-client
- sniffer -> elasticsearch-rest-client-sniffer
- rest-high-level -> elasticsearch-rest-high-level-client

A couple of small changes are also preparing the high level client for its first release.

Closes #20248
2017-07-13 09:44:25 +02:00
Ryan Ernst 70b2897bdf Scripting: Deprecate stored search template apis (#25437)
This commit deprecates the PUT, GET and DELETE search template apis.
Instead, the stored script api should be used.

closes #24596
2017-07-12 16:07:28 -07:00
Simon Willnauer e81804cfa4 Add a shard filter search phase to pre-filter shards based on query rewriting (#25658)
Today if we search across a large amount of shards we hit every shard. Yet, it's quite
common to search across an index pattern for time based indices but filtering will exclude
all results outside a certain time range ie. `now-3d`. While the search can potentially hit
hundreds of shards the majority of the shards might yield 0 results since there is not document
that is within this date range. Kibana for instance does this regularly but used `_field_stats`
to optimize the indexes they need to query. Now with the deprecation of `_field_stats` and it's upcoming removal a single dashboard in kibana can potentially turn into searches hitting hundreds or thousands of shards and that can easily cause search rejections even though the most of the requests are very likely super cheap and only need a query rewriting to early terminate with 0 results.

This change adds a pre-filter phase for searches that can, if the number of shards are higher than a the `pre_filter_shard_size` threshold (defaults to 128 shards), fan out to the shards
and check if the query can potentially match any documents at all. While false positives are possible, a negative response means that no matches are possible. These requests are not subject to rejection and can greatly reduce the number of shards a request needs to hit. The approach here is preferable to the kibana approach with field stats since it correctly handles aliases and uses the correct threadpools to execute these requests. Further it's completely transparent to the user and improves scalability of elasticsearch in general on large clusters.
2017-07-12 22:19:20 +02:00
Jack Conradson d2b4f7ac5a Disallow lang to be used with Stored Scripts (#25610)
Requests that execute a stored script will no longer be allowed to specify the lang of the script. This information is stored in the cluster state making only an id necessary to execute against. Putting a stored script will still require a lang.
2017-07-12 07:55:57 -07:00
Tim Brooks a3ade99fcf Fix BytesReferenceStreamInput#skip with offset (#25634)
There is a bug when a call to `BytesReferenceStreamInput` skip is made
on a `BytesReference` that has an initial offset. The offset for the
current slice is added to the current index and then subtracted from the
length. This introduces the possibility of a negative number of bytes to
skip. This happens inside a loop, which leads to an infinte loop.

This commit correctly subtracts the current slice index from the
slice.length. Additionally, the `BytesArrayTests` are modified to test
instances that include an offset.
2017-07-11 09:54:29 -05:00
Adrien Grand 481d5d09b2 Upgrade to lucene-7.0.0-snapshot-00142c9. (#25641)
Lucene 7.0 is feature-frozen now, so there should not be many changes until GA.
2017-07-11 13:58:55 +02:00
Tim Brooks b22bbf94da Avoid blocking on channel close on network thread (#25521)
Currently when we close a channel in Netty4Utils.closeChannels we
block until the closing is complete. This introduces the possibility
that a network selector thread will block while waiting until a
separate network selector thread closes a channel.

For instance: T1 closes channel 1 (which is assigned to a T1 selector).
Channel 1's close listener executes the closing of the node. That
means that T1 now tries to close channel 2. However, channel 2 is
assigned to a selector that is running on T2. T1 now must wait until T2
closes that channel at some point in the future.

This commit addresses this by adding a boolean to closeChannels
indicating if we should block on close. We only set this boolean to true
if we are closing down the server channels at shutdown. This call is
never made from a network thread. When we call the closeChannels method
with that boolean set to false, we do not block on close.
2017-07-10 10:50:51 -05:00
Jason Tedor c084542731 Bump version to 6.0.0-beta1
This commit does two things:
 - bumps the version from 6.0.0-alpha3 to 6.0.0-beta1
 - renames the 6.0.0-alpha3 version constant to 6.0.0-beta1

Relates #25621
2017-07-09 18:12:50 -04:00
Adrien Grand 40bb1663ee Index ids in binary form. (#25352)
Indexing ids in binary form should help with indexing speed since we would
have to compare fewer bytes upon sorting, should help with memory usage of
the live version map since keys will be shorter, and might help with disk
usage depending on how efficient the terms dictionary is at compressing
terms.

Since we can only expect base64 ids in the auto-generated case, this PR tries
to use an encoding that makes the binary id equal to the base64-decoded id in
the majority of cases (253 out of 256). It also specializes numeric ids, since
this seems to be common when content that is stored in Elasticsearch comes
from another database that uses eg. auto-increment ids.

Another option could be to require base64 ids all the time. It would make things
simpler but I'm not sure users would welcome this requirement.

This PR should bring some benefits, but I expect it to be mostly useful when
coupled with something like #24615.

Closes #18154
2017-07-07 14:22:47 +02:00
Martijn van Groningen 6db708ef75
Move more token filters to analysis-common module
The following token filters were moved: common grams, limit token, pattern capture and pattern raplace.

Relates to #23658
2017-07-07 10:02:52 +02:00
Simon Willnauer 1f67d079b1 Validate `transport.profiles.*` settings (#25508)
Transport profiles unfortunately have never been validated. Yet, it's very
easy to make a mistake when configuring profiles which will most likely stay
undetected since we don't validate the settings but allow almost everything
based on the wildcard in `transport.profiles.*`. This change removes the
settings subset based parsing of profiles but rather uses concrete affix settings
for the profiles which makes it easier to fall back to higher level settings since
the fallback settings are present when the profile setting is parsed. Previously, it was
unclear in the code which setting is used ie. if the profiles settings (with removed
prefixes) or the global node setting. There is no distinction anymore since we don't pull
prefix based settings.
2017-07-07 09:40:59 +02:00
Jason Tedor c96257ca73 Upgrade to Netty 4.1.13.Final
This commit upgrades the Netty dependency from version 4.1.11.Final to
4.1.13.Final.

Relates #25581
2017-07-06 15:37:00 -04:00
Martijn van Groningen d0f9f425bd
parent/child: Removed ParentJoinFieldSubFetchPhase 2017-07-06 13:15:02 +02:00
Martijn van Groningen 407273f81d
parent/child: Support parent id being specified as number in the _source 2017-07-06 11:48:57 +02:00
Jun Ohtani 6894ef6057 [Analysis] Support normalizer in request param (#24767)
* [Analysis] Support normalizer in request param

Support normalizer param
Support custom normalizer with char_filter/filter param

Closes #23347
2017-07-04 19:16:56 +09:00
Colin Goodheart-Smithe 43efcffcc2 Adds check for negative search request size (#25397)
* Adds check for negative search request size

This change adds a check to `SearchSourceBuilder` to throw and exception if the size set on it is set to a negative value.

Closes #22530

* fix error in reindex

* update re-index tests

* Addresses review comment

* Fixed tests

* Added random negative size test

* Fixes test
2017-07-04 10:51:38 +01:00
Christoph Büscher f576c987ce Remove QueryParseContext (#25486)
QueryParseContext is currently only used as a wrapper for an XContentParser, so
this change removes it entirely and changes the appropriate APIs that use it so
far to only accept a parser instead.
2017-07-03 17:30:40 +02:00
Simon Willnauer 5a7c8bb04e Cleanup network / transport related settings (#25489)
This commit makes the use of the global network settings explicit instead
of implicit within NetworkService. It cleans up several places where we fall
back to the global settings while we should have used tcp or http ones.

In addition this change also removes unnecessary settings classes
2017-07-02 10:16:50 +02:00
Simon Willnauer 6f131a63d3 Remove unregistered `transport.netty.*` settings (#25476)
These settings have not be working for a full major version since they
are not registered. Given that they are simply duplicates we can just remove
them.
2017-06-29 20:56:18 +02:00
Christoph Büscher 927111c91d Remove QueryParseContext from parsing QueryBuilders (#25448)
Currently QueryParseContext is only a thin wrapper around an XContentParser that
adds little functionality of its own. I provides helpers for long deprecated
field names which can be removed and two helper methods that can be made static
and moved to other classes. This is a first step in helping to remove
QueryParseContext entirely.
2017-06-29 17:10:20 +02:00
olcbean 3518e313b8 Unify the result interfaces from get and search in Java client (#25361)
As GetField and SearchHitField have the same members, they have been unified into
DocumentField.

Closes #16440
2017-06-29 11:35:28 +02:00
Martijn van Groningen c85ac402b0
test: Make many percolator integration tests real integration tests 2017-06-27 17:44:30 +02:00
Simon Willnauer d338a09812 Remove `mapping.single_type` from parent join test (#25391)
This removes the remaining usage of `mapping.single_type` from the parent join
module and moves it's bwc test to the mixed cluster tests

Relates to #24961
Relates to #20257
2017-06-26 17:33:07 +02:00
Martijn van Groningen a34f5fa812
Move more token filters to analysis-common module
The following token filters were moved: stemmer, stemmer_override, kstem, dictionary_decompounder, hyphenation_decompounder, reverse, elision and truncate.

Relates to #23658
2017-06-26 09:02:16 +02:00
Nik Everett da0b991331 Remove `index.mapping.single_type=false` from reindex tests (#25365)
* Remove the setting from the yml tests and replace with tests using
`join` field. We can't use the setting in yml tests without lots of
backflips but we have `ReindexParentChildTests` for the coverage.
There weren't tests for `join` field with reindex before this. Adding
these tests discovered #25363.
* Remove the setting from `ReindexParentChildTests` and replace with
`index.version.created=V_5_6_0`. This test can be entirely removed
when legacy parent/child support is dropped from core.
* Port the yml tests that set _parent into integ tests so they
can set the index created version. These tests can be removed
when we drop support for _parent in core.
* Port a delete-by-query test for filtering based on type to an
`ESIntegTestCase` so it can use `index.version.created=5.6.0` to
setup documents of multiple types. This whole feature can be dropped
when we no longer support multiple types per index.

Relates to #24961
2017-06-23 17:14:59 -04:00
Simon Willnauer 4ae426a552 Remove remaining `index.mapping.single_type=false` (#25369)
This change cleans up remaining tests  to not use index.mapping.single_type=false
but instead where applicable use a single type or markt the index as created
with a pre 6.x version.

Yet, there is still on leftover in the client tests that needs special attention.
See `org.elasticsearch.client.SearchIT`

Relates to #24961
2017-06-23 10:26:06 +02:00
Tal Levy 1ac7818201 fix sort and string processor tests around targetField (#25358)
Tests were randomly assigning `targetField` to an existing field that was an array,
causing path resolution issues. This PR fixes those tests

Closes #25346 & #25348
2017-06-22 13:14:18 -07:00
Jack Conradson 96b62409a8 Update Painless to Allow Augmentation from Any Class (#25360)
Custom whitelists in Painless will need to allow classes to be augmented beyond the currently hard-coded Augmentation class tied to Painless directly. This change allows any class to specify an augmentation on a Painless struct using an appropriate static method. Changes to loading the whitelist have also been created to allow for this specification of a different class for augmentation.
2017-06-22 12:16:46 -07:00
Martijn van Groningen 343e7571b9
test: single type defaults to true since alpha1 and not alpha3
Closes #25354
2017-06-22 16:31:15 +02:00
Adrien Grand 44e9c0b947 Upgrade to lucene-7.0.0-snapshot-ad2cb77. (#25349)
Most notable changes:
 - better update concurrency: LUCENE-7868
 - TopDocs.totalHits is now a long: LUCENE-7872
 - QueryBuilder does not remove the boolean query around multi-term synonyms:
   LUCENE-7878
 - removal of Fields: LUCENE-7500

For the `TopDocs.totalHits` change, this PR relies on the fact that the encoding
of vInts and vLongs are compatible: you can write and read with any of them as
long as the value can be represented by a positive int.
2017-06-22 12:35:33 +02:00
Martijn van Groningen a977569085
percolator: Deprecate `document_type` parameter.
The `document_type` parameter is no longer required to be specified,
because by default from 6.0 only a single type is allowed. (`index.mapping.single_type` defaults to `true`)
2017-06-22 09:55:06 +02:00
Nik Everett 8d9a08e239 Fix reindex test when log level is debug
When log level is debug we'd dereference null because the test
was being cute and cutting corners.

Relates to #25256
2017-06-20 16:06:58 -04:00
Jun Ohtani 62d1969595 Parse synonyms with the same analysis chain (#8049)
* [Analysis] Parse synonyms with the same analysis chain

Synonym Token Filter / Synonym Graph Filter tokenize synonyms with whatever tokenizer and token filters appear before it in the chain.

Close #7199
2017-06-20 21:50:33 +09:00
Nik Everett 3261586cac Tweak reindex cancel logic and add many debug logs (#25256)
I'm still trying to hunt down rare failures in the cancelation tests
for reindex and friends. Here is the latest:
https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+5.x+multijob-unix-compatibility/os=ubuntu/876/console

It doesn't show much, other than that one of the tasks didn't kill
itself when asked to cancel.

So I'm going a bit crazy with debug logging so that the next time this
comes up I can trace exactly what happened.

Additionally, this tweaks the logic around how rethrottles were
performed around cancel. Previously we set the `requestsPerSecond`
to `0` when we cancelled the task. That was the "old way" to set them
to inifity which was the intent. This switches that from `0` to
`Float.MAX_VALUE` which is the "new way" to set the `requestsPerSecond`
to infinity. I don't know that this is much better, but it feels better.
2017-06-19 18:46:42 -04:00
Andy Bristol 4c5bd57619 Rename simple pattern tokenizers (#25300)
Changed names to be snake case for consistency

Related to #25159, original issue #23363
2017-06-19 13:48:43 -07:00
Simon Willnauer a8d5a58801 Replace deprecated API usage in Netty4HttpChannel 2017-06-17 14:04:23 +02:00
Christoph Büscher e99ced06cc [Tests] Check that parsing aggregations works in a forward compatible way (#25219)
This change adds tests for the aggregation parsing that try to simulate that we
can parse existing aggregations in a forward compatible way in the future,
ignoring potential newly added fields or substructures to the xContent response.
2017-06-17 13:06:31 +02:00
Simon Willnauer f18b0d293c Move TransportStats accounting into TcpTransport (#25251)
Today TcpTransport is the de-facto base-class for transport implementations.
The need for all the callbacks we have in TransportServiceAdaptor are not necessary
anymore since we can simply have the logic inside the base class itself. This change
moves the stats metrics directly into TcpTransport removing the need for low level
bytes send / received callbacks.
2017-06-16 22:34:11 +02:00
Nik Everett ecc87f613f Move pre-configured "keyword" tokenizer to the analysis-common module (#24863)
Moves the keyword tokenizer to the analysis-common module. The keyword tokenizer is special because it is used by CustomNormalizerProvider so I pulled it out into its own PR. To get the move to work I've reworked the lookup from static to one using the AnalysisRegistry. This seems safe enough.

Part of #23658.
2017-06-16 11:48:15 -04:00
Jack Conradson 50db8cb351 Add needs methods for specific variables to Painless script context factories. (#25267) 2017-06-15 17:00:33 -07:00
Martijn van Groningen 428e70758a
Moved more token filters to analysis-common module.
The following token filters were moved: `edge_ngram`, `ngram`, `uppercase`, `lowercase`, `length`, `flatten_graph` and `unique`.

Relates to #23658
2017-06-15 18:28:31 +02:00
Jim Ferenczi 5e64cd08bc [Test] restore BWC for parent-join now that the new mapping format is in 5.x 2017-06-15 15:15:48 +02:00
Jim Ferenczi 9ca33e2450 Add a section named "relations" in the ParentJoinFieldMapper (#25248)
* Add a section named "relation" in the ParentJoinFieldMapper

This commit puts the parent/child definition in an inner section named "relation".
Mapping for the parent-join will look like this:

```
"join_field": {
  "type": "join"
  "relations":
    "parent": "child"
  }
}
```
2017-06-15 14:56:20 +02:00
Tal Levy 2cd771a230 fix: Sort Processor does not have proper behavior with targetField (#25237)
to specify a `targetField`. This results in some interesting behavior that was missed in the review.
This processor sorts in-place, so there is a side-effect in both the original field and the target field.
Another bug was that the targetField was not being set if the list being sorted was fewer than two elements.

The new behavior works like this: If targetField and fieldName are not the same, we copy the list.
2017-06-15 05:28:54 -07:00
Boaz Leskes 648b4717a4 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.
2017-06-15 13:24:07 +02:00
Tanguy Leroux 27f1206999 Use SPI in High Level Rest Client to load XContent parsers (#25098)
This commit adds a NamedXContentProvider interface that can 
be implemented by plugins or modules using Java's SPI feature 
in order to provide additional NamedXContent parsers to external
applications like the Java High Level Rest Client.
2017-06-15 12:50:02 +02:00
Adrien Grand 0c117145f6 Upgrade to lucene-7.0.0-snapshot-92b1783. (#25222)
This snapshot has faster range queries on range fields (LUCENE-7828), more
accurate norms (LUCENE-7730) and the ability to use fake term frequencies
(LUCENE-7854).
2017-06-15 09:52:07 +02:00