Commit Graph

2856 Commits

Author SHA1 Message Date
David Roberts da97325790 [ML] Speed up persistent task rechecks in ML failover tests (#43291)
The ML failover tests sometimes need to wait for jobs to be
assigned to new nodes following a node failure.  They wait
10 seconds for this to happen.  However, if the node that
failed was the master node and a new master was elected then
this 10 seconds might not be long enough as a refresh of the
memory stats will delay job assignment.  Once the memory
refresh completes the persistent task will be assigned when
the next cluster state update occurs or after the periodic
recheck interval, which defaults to 30 seconds.  Rather than
increase the length of the wait for assignment to 31 seconds,
this change decreases the periodic recheck interval to 1
second.

Fixes #43289
2019-06-18 09:19:20 +01:00
Nhat Nguyen 0c5086d2f3 Rebuild version map when opening internal engine (#43202)
With this change, we will rebuild the live version map and local
checkpoint using documents (including soft-deleted) from the safe commit
when opening an internal engine. This allows us to safely prune away _id
of all soft-deleted documents as the version map is always in-sync with
the Lucene index.

Relates #40741
Supersedes #42979
2019-06-17 18:08:09 -04:00
Benjamin Trent 365f87c622
[ML][Data Frame] only complete task after state persistence (#43230) (#43294)
* [ML][Data Frame] only complete task after state persistence

There is a race condition where the task could be completed, but there
is still a pending document write. This change moves
the task cancellation into the actionlistener of the state persistence.

intermediate commit

intermediate commit

* removing unused import

* removing unused const

* refreshing internal index after waiting for task to complete

* adjusting test data generation
2019-06-17 16:49:00 -05:00
Martijn Laarman 8b1b9f8ab9
Introduce stability description to the REST API specification (#38413) (#43278)
* introduce state to the REST API specification

* change state over to stability

* CCR is no GA updated to stable

* SQL is now GA so marked as stable

* Introduce `internal` as state for API's, marks stable in terms of lifetime but unstable in terms of guarantees on its output format since it exposes internal representations

* make setting a wrong stability value, or not setting it at all an error that causes the YAML test suite to fail

* update spec files to be explicit about their stability state

* Document the fact that stability needs to be defined

Otherwise the YAML test runner will fail (with a nice exception message)

* address check style violations

* update rest spec unit tests to include stability

* found one more test spec file not declaring stability, made sure stability appears after documentation everywhere

* cluster.state is stable, mark response in some way to denote its a key value format that can be changed during minors

* mark data frame API's as beta

* remove internal and private as states for an API

* removed the wrong enum values in the Stability Enum in the previous commit

(cherry picked from commit 61c34bbd92f8f7e5f22fa411c6b682b0ebd8a99d)
2019-06-17 16:57:13 +02:00
Lee Hinman 21da84edbc Make ILM force merging best effort (#43246)
It's possible for force merges kicked off by ILM to silently stop (due
to a node relocating for example). In which case, the segment count may
not reach what the user configured. In the subsequent `SegmentCountStep`
waiting for the expected segment count may wait indefinitely. Because of
this, this commit makes force merges "best effort" and then changes the
`SegmentCountStep` to simply report (at INFO level) if the merge was not
successful.

Relates to #42824
Resolves #43245
2019-06-17 08:45:22 -06:00
David Roberts 3effe264da [ML] Fix problem with lost shards in distributed failure test (#43153)
We were stopping a node in the cluster at a time when
the replica shards of the .ml-state index might not
have been created.  This change moves the wait for
green status to a point where the .ml-state index
exists.

Fixes #40546
Fixes #41742

Forward port of #43111
2019-06-17 09:28:56 +01:00
Przemysław Witek b2613a123d
[7.x] Report exponential_avg_bucket_processing_time which gives more weight to recent buckets (#43189) (#43263) 2019-06-17 08:58:26 +02:00
David Roberts 3928c624a3 [ML] Close sample stream in post_data endpoint (#43235)
A static code analysis revealed that we are not closing
the input stream in the post_data endpoint.  This
actually makes no difference in practice, as the
particular InputStream implementation in this case is
org.elasticsearch.common.bytes.BytesReferenceStreamInput
and its close() method is a no-op. However, it is good
practice to close the stream anyway.
2019-06-14 17:54:54 +01:00
Benjamin Trent 8c66149e2d
[ML][Data Frame] have sum map to a double to prevent overflows (#43213) (#43219) 2019-06-14 10:43:36 -05:00
Marios Trivyzas 9cd89c3453 SQL: Increase hard limit for sorting on aggregates (#43220)
To be consistent with the `search.max_buckets` default setting,
set the hard limit of the PriorityQueue used for in memory sorting,
when sorting on an aggregate function, to 10000.

Fixes: #43168

(cherry picked from commit 079e012fdea68ea0a7daae078359495047e9c407)
2019-06-14 13:51:38 +02:00
Alpar Torok cce5b0f018 Convert dataframes to use testclusters (#43032) 2019-06-14 11:02:39 +03:00
Przemysław Witek 65a584b6fb
[7.x] Report timing stats as part of the Job stats response (#42709) (#43193) 2019-06-14 09:03:14 +02:00
Marios Trivyzas 3c73602524 SQL: Fix wrong results when sorting on aggregate (#43154)
- Previously, when shorting on an aggregate function the bucket
processing ended early when the explicit (LIMIT XXX) or the impliciti
limit of 512 was reached. As a consequence, only a set of grouping
buckets was processed and the results returned didn't reflect the global
ordering.

- Previously, the priority queue shorting method had an inverse
comparison check and the final response from the priority queue was also
returned in the inversed order because of the calls to the `pop()`
method.

Fixes: #42851

(cherry picked from commit 19909edcfdf5792b38c1363b07379783ebd0e6c4)
2019-06-13 21:59:20 +02:00
Jason Tedor 5bc3b7f741
Enable node roles to be pluggable (#43175)
This commit introduces the possibility for a plugin to introduce
additional node roles.
2019-06-13 15:15:48 -04:00
Ryan Ernst c3ce3f6891 Add native code info to ML info api (#43172)
The machine learning feature of xpack has native binaries with a
different commit id than the rest of code. It is currently exposed in
the xpack info api. This commit adds that commit information to the ML
info api, so that it may be removed from the info api.
2019-06-13 11:38:58 -07:00
Alpar Torok 4ba94a5051 Testclusters: convert ccr tests (#42313) 2019-06-13 19:19:36 +03:00
David Roberts 43665183c2 [ML] Restrict detection of epoch timestamps in find_file_structure (#43188)
Previously 10 digit numbers were considered candidates to be
timestamps recorded as seconds since the epoch and 13 digit
numbers as timestamps recorded as milliseconds since the epoch.

However, this meant that we could detect these formats for
numbers that would represent times far in the future.  As an
example ISBN numbers starting with 9 were detected as milliseconds
since the epoch since they had 13 digits.

This change tweaks the logic for detecting such timestamps to
require that they begin with 1 or 2.  This means that numbers
that would represent times beyond about 2065 are no longer
detected as epoch timestamps.  (We can add 3 to the definition
as we get closer to the cutoff date.)
2019-06-13 13:15:41 +01:00
Alpar Torok 167e51335d Convert ILM tests to use testclusters (#43076)
Also improove the error message when bin scripts are not found
2019-06-13 12:24:48 +03:00
Alpar Torok eb7a8bb4a4 Testclusters: graph (#43033)
Convert x-pack  graph to use testClusters
2019-06-13 09:50:59 +03:00
Simon Willnauer f70141c862 Only load FST off heap if we are actually using mmaps for the term dictionary (#43158)
Given the significant performance impact that NIOFS has when term dicts are
loaded off-heap this change enforces FstLoadMode#AUTO that loads term dicts
off heap only if the underlying index input indicates a memory map.

Relates to #43150
2019-06-13 07:54:02 +02:00
Benjamin Trent ec50d4d281
[ML][Data Frame] write a warning audit on bulk index failures (#43106) (#43171)
* [ML][Data Frame] write a warning audit on bulk index failures

* adding failure message and moving to use volalitile
2019-06-12 14:50:17 -05:00
Benjamin Trent aff4795441
[ML][Data Frame] cleaning up tests since tasks are cancelled onfinish (#43136) (#43166)
* [ML][Data Frame] cleaning up usage test since tasks are cancelled onfinish

* Update DataFrameUsageIT.java

* Fixing additional test, waiting for task to complete

* removing unused import

* unmuting test
2019-06-12 14:39:38 -05:00
Benjamin Trent b110164bf4
[ML][Data Frame] add the src priv check for view_index_metadata (#43118) (#43161) 2019-06-12 13:22:46 -05:00
Benjamin Trent f13f55ede3
[ML][Data Frame] change failure count reset logic (#43064) (#43159) 2019-06-12 13:22:34 -05:00
David Kyle 597ae5c7b8 [ML DataFrame] Reject Data Frame Ids containing upper case characters (#43145) 2019-06-12 18:13:18 +01:00
Yannick Welsch 110f0c5b7e Mute testDataFrameTransformCrud
Relates to #43139
2019-06-12 14:12:01 +02:00
Dimitris Athanasiou b28e006f7c
[ML] Lock down extraction method when possible (#43104) (#43140) 2019-06-12 14:07:17 +03:00
Luca Cavanna afeda1a7b9 Split search in two when made against throttled and non throttled searches (#42510)
When a search on some indices takes a long time, it may cause problems to other indices that are being searched as part of the same search request and being written to as well, because their search context needs to stay open for a long time. This is especially a problem when searching against throttled and non-throttled indices as part of the same request. The problem can be generalized though: this may happen whenever read-only indices are searched together with indices that are being written to. Search contexts staying open for a long time is only an issue for indices that are being written to, in practice.

This commit splits the search in two sub-searches: one for read-only indices, and one for ordinary indices. This way the two don't interfere with each other. The split is done only when size is greater than 0, no scroll is provided and query_then_fetch is used as search type. Otherwise, the search executes like before. Note that the returned num_reduce_phases reflect the number of reduction phases that were run. If the search is split in two, there are three reductions: one non-final for each search, and a final one that merges the results of the previous two.

Closes #40900
2019-06-12 11:25:03 +02:00
Nhat Nguyen 5692be2161 Fix timing issue in CcrRetentionLeaseIT (#43054)
In these tests, we sleep for a small multiple of the renew interval,
then check that the retention leases are not changed. If a renewal
request takes longer than that interval because of GC or slow CI, then
the retention leases are not the same as before sleep. With this change,
we relax to assert that we eventually stop the renewable process.

Closes #39509
2019-06-11 18:03:16 -04:00
Benjamin Trent 7ff3d86cf0
[ML][Data Frame] adding dest.index and id validations (#43053) (#43109)
* [ML][Data Frame] adding dest.index and id validations

* adjusting message format

* Adjusting id validity pattern

* Update DataFrameStrings.java
2019-06-11 15:55:18 -05:00
Benjamin Trent e384bf0276
[ML-DataFrame] stop task at completion of data frame function (#42955) (#43114)
* stop data frame task after it finishes

* test auto stop

* adapt tests

* persist the state correctly and move stop into listener

* Calling `onStop` even if persistence fails, changing `stop` to rely on doSaveState
2019-06-11 15:55:02 -05:00
Ryan Ernst 172cd4dbfa Remove description from xpack feature sets (#43065)
The description field of xpack featuresets is optionally part of the
xpack info api, when using the verbose flag. However, this information
is unnecessary, as it is better left for documentation (and the existing
descriptions describe anything meaningful). This commit removes the
description field from feature sets.
2019-06-11 09:22:58 -07:00
David Roberts d3136f99e6 [ML] Fix race condition when closing time checker (#43098)
The tests for the ML TimeoutChecker rely on threads
not being interrupted after the TimeoutChecker is
closed.  This change ensures this by making the
close() and setTimeoutExceeded() methods synchronized
so that the code inside them cannot execute
simultaneously.

Fixes #43097
2019-06-11 16:39:17 +01:00
Nhat Nguyen 5d3849215b CCR should not replicate private/internal settings (#43067)
With this change, CCR will not replicate internal or private settings to
follower indices.

Closes #41268
2019-06-11 06:59:09 -04:00
Martijn Laarman cb7ce865b7
remove path from rest-api-spec (#41452) (#43084)
(cherry picked from commit f5fde1d0843d2f0f53d3b9a15b9cfc8b94471ab7)
2019-06-11 12:52:36 +02:00
Ioannis Kakavas 1776d6e055 Refresh remote JWKs on all errors (#42850)
It turns out that key rotation on the OP, can manifest as both
a BadJWSException and a BadJOSEException in nimbus-jose-jwt. As
such we cannot depend on matching only BadJWSExceptions to
determine if we should poll the remote JWKs for an update.

This has the side-effect that a remote JWKs source will be polled
exactly one additional time too for errors that have to do with
configuration, or for errors that might be caused by not synched
clocks, forged JWTs, etc. ( These will throw a BadJWTException
which extends BadJOSEException also )
2019-06-11 11:01:54 +03:00
Benjamin Trent 79052050bf
[ML] Adding support for geo_shape, geo_centroid, geo_point in datafeeds (#42969) (#43069)
* [ML] Adding support for geo_shape, geo_centroid, geo_point in datafeeds

* only supporting doc_values for geo_point fields

* moving validation into GeoPointField ctor
2019-06-10 21:52:53 -05:00
Benjamin Trent eadfe05587
[ML] Changes slice specification to auto. See #42996 (#43039) (#43070) 2019-06-10 21:52:22 -05:00
Nhat Nguyen 53eb630700 Fix NPE in CcrRetentionLeaseIT (#43059)
The retention leases stats is null if the processing shard copy is being
closed. In this the case, we should check against null then retry to
avoid failing a test.

Closes #41237
2019-06-10 17:58:37 -04:00
Nhat Nguyen f2e66e22eb Increase waiting time when check retention locks (#42994)
WriteActionsTests#testBulk and WriteActionsTests#testIndex sometimes
fail with a pending retention lock. We might leak retention locks when
switching to async recovery. However, it's more likely that ongoing
recoveries prevent the retention lock from releasing.

This change increases the waiting time when we check for no pending
retention lock and also ensures no ongoing recovery in
WriteActionsTests.

Closes #41054
2019-06-10 17:58:37 -04:00
Nhat Nguyen 4191df6e1d Unmute IndexFollowingIT#testFollowIndex
Fixed in #41987
2019-06-10 17:58:37 -04:00
Benjamin Trent 1ddc4c8fc6
[ML][Data Frame] Removes slice specification from DBQ. See #42996 (#43036) (#43055) 2019-06-10 13:40:55 -05:00
Dimitris Athanasiou 76a92b49a8
[ML] Get resources action should be lenient when sort field is unmapped (#42991) (#43046)
Get resources action sorts on the resource id. When there are no resources at
all, then it is possible the index does not contain a mapping for the resource
id field. In that case, the search api fails by default.

This commit adjusts the search request to ignore unmapped fields.

Closes elastic/kibana#37870
2019-06-10 19:50:19 +03:00
Alan Woodward 8e23e4518a Move construction of custom analyzers into AnalysisRegistry (#42940)
Both TransportAnalyzeAction and CategorizationAnalyzer have logic to build
custom analyzers for index-independent analysis. A lot of this code is duplicated,
and it requires the AnalysisRegistry to expose a number of internal provider
classes, as well as making some assumptions about when analysis components are
constructed.

This commit moves the build logic directly into AnalysisRegistry, reducing the
registry's API surface considerably.
2019-06-10 14:33:25 +01:00
David Turner 68339f90e9 Mute AutodetectMemoryLimitIT#testTooManyPartitions
Relates #43013
2019-06-10 09:20:36 +01:00
Andrei Stefan 036f9c4a55 SQL: cover the Integer type when extracting values from _source (#42859)
* Take into consideration a wider range of Numbers when extracting the
values from source, more specifically - BigInteger and BigDecimal.

(cherry picked from commit 561b8d73dd7b03c50242e4e3f0128b2142959176)
2019-06-10 09:25:56 +03:00
Jason Tedor 63bad28005
Do not allow modify aliases on followers (#43017)
Now that aliases are replicated by a follower from its leader, this
commit prevents directly modifying aliases on follower indices.
2019-06-09 22:53:54 -04:00
Jason Tedor 915d2f2daa
Refactor put mapping request validation for reuse (#43005)
This commit refactors put mapping request validation for reuse. The
concrete case that we are after here is the ability to apply effectively
the same framework to indices aliases requests. This commit refactors
the put mapping request validation framework to allow for that.
2019-06-09 10:19:04 -04:00
Benjamin Trent 553c73b22d
[ML][Data Frame] allow null values for aggs with sparse data (#42966) (#42998)
* [ML][Data Frame] allow null values for aggs with sparse data

* Making classes static, memory allocation optimization
2019-06-07 15:43:06 -05:00
Benjamin Trent 755ba72896
[ML][Data frame] make sure that fields exist when creating progress (#42943) (#42984) 2019-06-07 10:13:18 -05:00