When a lucene 4.8+ file is transferred, Store returns a VerifyingIndexOutput
that verifies both the CRC32 integrity and the length of the file.
However, for older files, problems can make it to the lucene level. This is not great
since older lucene files aren't especially strong as far as detecting issues here.
For example, if a network transfer is closed on the remote side, we might write a
truncated file... which old lucene formats may or may not detect.
The idea here is to verify old files with their legacy Adler32 checksum, plus expected
length. If they don't have an Adler32 (segments_N, jurassic elasticsearch?, its optional
as far as the protocol goes), then at least check the length.
We could improve it for segments_N, its had an embedded CRC32 forever in lucene, but this
gets trickier. Long term, we should also try to also improve tests around here, especially
backwards compat testing, we should test that detected corruptions are handled properly.
Closes#8399
Conflicts:
src/main/java/org/elasticsearch/index/store/Store.java
src/test/java/org/elasticsearch/index/store/StoreTest.java
If a shard (e.g. replica) gets initialized after we indexed the document it gets refreshed internally and we find the doc and its term_vectors, thus the test fails
Fixes an issue where only absolute bytes were taken into account when
kicking off an automatic reroute due to disk usage. Also randomized the
tests to use either an absolute value or a percentage so this is tested.
Also adds logging for each node over the high and low watermark every
time a new cluster info usage is gathered (defaults to every 30
seconds).
Related to #8368Fixes#8367
Today we use busy waiting and sampling when we execute HealthReqeusts
on the master. This is tricky sicne we might sample a not yet fully applied
cluster state and make a decsions base on the partial cluster state. This can
lead to ugly problems since requests might be routed to nodes where shards are
already marked as relocated but on the actual cluster state they are still started.
Yet, this window is very small usually it can lead to ugly test failures.
This commit moves the health request over to a listener pattern that gets the actual
applied cluster state.
Closes#8350
When a node stops, we cancel any ongoing join process. With #8327, we improved this logic and wait for it to complete before shutting down the node. However, the joining thread is part of a thread pool and will not stop until the thread pool is shutdown.
Another issue raised by the unneeded wait is that when we shutdown, we may ping ourselves - which results in an ugly warn level log. We now log all remote exception during pings at a debug level.
Closes#8359
Previously the bulk of our shard recovery code was in a 300-line
anonymous class in `RecoverySource`. This made it difficult to find and
more difficult to read.
This factors out that code into a `ShardRecoveryHandler` class, adding
javadocs for each function and phase of the recovery, as well as
comments explaining some of the more esoteric functions performed during
recovery.
It's hoped that this will help more people understand Elasticsearch's
recovery procedure.
No *major* functionality has changed, only typo corrections, some minor
allocation improvements and logging clarification changes.
If the translog file is corrupted or truncated the stream is never closed
and the filehandle leaks. This commit closes the stream in the case of an
exception.
Today we use the File API for file deletion as well as recursive
directory deletions. This API returns a boolean if operations
are successful while hiding the actual reason why they failed.
The Path API throws and actual exception that might provide better
insights and debug information.
Closes#8366
When a node stops, we cancel any ongoing join process. With #8327, we improved this logic and wait for it to complete before shutting down the node. In our tests we typically shutdown an entire cluster at once, which makes it very likely for nodes to be joining while shutting down. This introduces a race condition where the joinThread.interrupt can happen before the thread starts waiting on pings which causes shutdown logic to be slow. This commits improves by repeatedly trying to stop the thread in smaller waits.
Another side effect of the change is that we are now more likely to ping ourselves while shutting down, we results in an ugly warn level log. We now log all remote exception during pings at a debug level.
Closes#8359
The tests were failing because there was a shard which didn't get any documents and the tests assumed all shards had documents. This commit fixes this assumption
This has a lot of improvements in lucene, particularly around memory usage, merging, safety, compressed bitsets, etc.
On the elasticsearch side, summary of the larger changes:
API changes: postings API became a "pull" rather than "push", collector API became per-segment, etc.
packaging changes: add lucene-backwards-codecs.jar as a dependency.
improvements to boolean filtering: especially ensuring it will not be slow for SparseBitSet.
use generic BitSet api in plumbing so that concrete bitset type is an implementation detail.
use generic BitDocIdSetFilter api for dedicated bitset cache, so there is type safety.
changes to support atomic commits
implement Accountable.getChildResources (detailed memory usage API) for fielddata, etc
change handling of IndexFormatTooOld/New, since they no longer extends CorruptIndexException
Closes#8347.
Squashed commit of the following:
commit d90d53f5f21b876efc1e09cbd6d63c538a16cd89
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Nov 5 21:35:28 2014 +0100
Make default codec/postings/docvalues format constants
commit cb66c22c71cd304a36e7371b199a8c279908ae37
Merge: d4e2f6d ad4ff43
Author: Robert Muir <rmuir@apache.org>
Date: Wed Nov 5 11:41:13 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit d4e2f6dfe767a5128c9b9ae9e75036378de08f47
Merge: 4e5445c 4111d93
Author: Robert Muir <rmuir@apache.org>
Date: Wed Nov 5 06:26:32 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit 4e5445c775f580730eb01360244e9330c0dc3958
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 16:19:19 2014 -0500
FixedBitSet -> BitSet
commit 9887ea73e8b857eeda7f851ef3722ef580c92acf
Merge: 1bf8894 fc84666
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 15:26:25 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit 1bf8894430de3e566d0dc5623b0cc28b0d674ebb
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 15:22:51 2014 -0500
remove nocommit
commit a9c2a2259ff79c69bae7806b64e92d5f472c18c8
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 13:48:43 2014 -0500
turn jenkins red again
commit 067baaaa4d52fce772c81654dcdb5051ea79139f
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 13:18:21 2014 -0500
unzip from stream
commit 82b6fba33d362aca2313cc0ca495f28f5ebb9260
Merge: b2214bb 6523cd9
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 13:10:59 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit b2214bb093ec2f759003c488c3c403c8931db914
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 13:09:53 2014 -0500
go back to my URL until we can figure out what is up with jenkins
commit e7d614172240175a51f580aeaefb6460d21cede9
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 10:52:54 2014 -0500
try this jenkins
commit 337a3c7704efa7c9809bf373152d711ee55f876c
Author: Simon Willnauer <simonw@apache.org>
Date: Tue Nov 4 16:17:49 2014 +0100
Rename temp-files under lock to prevent metadata reads while renaming
commit 77d5ba80d0a76efa549dd753b9f114b2f2d2d29c
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 10:07:11 2014 -0500
continue to treat too-old/too-new as corruption for now
commit 98d0fd2f4851bc50e505a94ca592a694d502c51c
Author: Robert Muir <rmuir@apache.org>
Date: Tue Nov 4 09:24:21 2014 -0500
fix last nocommit
commit 643fceed66c8caf22b97fc489d67b4a2a90a1a1c
Author: Simon Willnauer <simonw@apache.org>
Date: Tue Nov 4 14:46:17 2014 +0100
remove NoSuchDirectoryException
commit 2e43c4feba05cfaf451df70f946c0930cbcc4557
Merge: 93826e4 8163107
Author: Simon Willnauer <simonw@apache.org>
Date: Tue Nov 4 14:38:00 2014 +0100
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit 93826e4d56a6a97c2074669014af77ff519bde63
Merge: 7f10129 44e24d3
Author: Simon Willnauer <simonw@apache.org>
Date: Tue Nov 4 12:54:27 2014 +0100
Merge branch 'master' into enhancement/lucene_5_0_upgrade
Conflicts:
src/main/java/org/elasticsearch/index/store/DistributorDirectory.java
src/main/java/org/elasticsearch/index/store/Store.java
src/main/java/org/elasticsearch/indices/recovery/RecoveryStatus.java
src/test/java/org/elasticsearch/index/store/DistributorDirectoryTest.java
src/test/java/org/elasticsearch/index/store/StoreTest.java
src/test/java/org/elasticsearch/indices/recovery/RecoveryStatusTests.java
commit 7f10129364623620575c109df725cf54488b3abb
Author: Adrien Grand <jpountz@gmail.com>
Date: Tue Nov 4 11:32:24 2014 +0100
Fix TopHitsAggregator to not ignore the top-level/leaf collector split.
commit 042fadc8603b997bdfdc45ca44fec70dc86774a6
Author: Adrien Grand <jpountz@gmail.com>
Date: Tue Nov 4 11:31:20 2014 +0100
Remove MatchDocIdSet in favor of DocValuesDocIdSet.
commit 7d877581ff5db585a674c95ac391ac78a0282826
Author: Adrien Grand <jpountz@gmail.com>
Date: Tue Nov 4 11:10:08 2014 +0100
Make the and filter use the cost API.
Lucene 5 ensured that cost() can safely be used, and this will have the benefit
that the order in which filters are specified is not important anymore (only
for slow random-access filters in practice).
commit 78f1718aa2cd82184db7c3a8393e6215f43eb4a8
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 23:55:17 2014 -0500
fix previous eclipse import braindamage
commit 186c40e9258ce32f22a9a714ab442a310b6376e0
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 22:32:34 2014 -0500
allow child queries to exhaust iterators again
commit b0b1271305e1b6d0c4c4da51a3c54df1aa5c0605
Author: Ryan Ernst <ryan@iernst.net>
Date: Mon Nov 3 14:50:44 2014 -0800
Fix nocommit for mapping output. index_options will not be printed if
the field is not indexed.
commit ba223eb85e399c9620a347a983e29bf703953e7a
Author: Ryan Ernst <ryan@iernst.net>
Date: Mon Nov 3 14:07:26 2014 -0800
Remove no commit for chinese analyzer provider. We should have a
separate issue to address not using this provider on new indexes.
commit ca554b03c4471797682b2fb724f25205cf040c4a
Author: Ryan Ernst <ryan@iernst.net>
Date: Mon Nov 3 13:41:59 2014 -0800
Fix stop tests
commit de67c4653ec47dee9c671390536110749d2bb05f
Author: Ryan Ernst <ryan@iernst.net>
Date: Mon Nov 3 12:51:17 2014 -0800
Remove analysis nocommits, switching over to Lucene43*Filters for
backcompat
commit 50cae9bec72c25c33a1ab8a8931bccb3355171e2
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 15:32:25 2014 -0500
add ram accounting and TODO lazy-loading (its no worse than master, can be a followup improvement) for suggesters
commit 7a7f0122f138684b312d0f0b03dc2a9c16c15f9c
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 15:11:26 2014 -0500
bump lucene version
commit cd0cae5c35e7a9e049f49ae45431f658fb86676b
Merge: 446bc09 3c72073
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 14:49:05 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit 446bc09b4e8bf4602d3c252b53ddaa0da65cce2f
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 14:46:30 2014 -0500
remove hack
commit a19d85a968d82e6d00292b49630ef6ff2dbf2f32
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 12:53:11 2014 -0500
dont create exceptions with circular references on corruption (will open a PR for this)
commit 0beefb9e821d97c37e90ec556d81ac7b00369b8a
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 11:47:14 2014 -0500
temporarily add craptastic detector for this horrible bug
commit e9f2d298bff75f3d1591f8622441e459c3ce7ac3
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 10:56:01 2014 -0500
add nocommit
commit e97f1d50a91a7129650b8effc7a9ecf74ca0569a
Merge: c57a3c8 f1f50ac
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 10:12:12 2014 -0500
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit c57a3c8341ed61dca62eaf77fad6b8b48aeb6940
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 10:11:46 2014 -0500
fix nocommit
commit dd0e77e4ec07c7011ab5f6b60b2ead33dc2333d2
Author: Robert Muir <rmuir@apache.org>
Date: Mon Nov 3 09:54:09 2014 -0500
nocommit -> TODO, this is in much more places in the codebase, bigger issue
commit 3cc3bf56d72d642059f8fe220d6f2fed608363e9
Author: Ryan Ernst <ryan@iernst.net>
Date: Sat Nov 1 23:59:17 2014 -0700
Remove nocommit and awaitsfix for edge ngram filter test.
commit 89f115245155511c0fbc0d5ee62e63141c3700c1
Author: Ryan Ernst <ryan@iernst.net>
Date: Sat Nov 1 23:57:44 2014 -0700
Fix EdgeNGramTokenFilter logic for version <= 4.3, and fixed instanceof
checks in corresponding tests to correctly check for reverse filter when
applicable.
commit 112df869cd199e36aab0e1a7a288bb1fdb2ebf1c
Author: Robert Muir <rmuir@apache.org>
Date: Sun Nov 2 00:08:30 2014 -0400
execute geo disjoint query/filter as intersects
commit e5061273cc685f1252e9a3a9ae4877ec9bce7752
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 22:58:59 2014 -0400
remove chinese analyzer from docs
commit ea1af11b8978fcc551f198e24fe21d52806993ef
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 22:29:00 2014 -0400
fix ram accounting bug
commit 53c0a42c6aa81aa6bf81d3aa77b95efd513e0f81
Merge: e3bcd3c 6011a18
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 22:16:29 2014 -0400
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit e3bcd3cc07a4957e12c7b3affc462c31290a9186
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 22:15:01 2014 -0400
fix url-email back compat (thanks ryan)
commit 91d6b096a96c357755abee167098607223be1aad
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 22:11:26 2014 -0400
bump lucene version
commit d2bb9568df72b37ec7050d25940160b8517394bc
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 20:33:07 2014 -0400
remove nocommit
commit 1d049c471e19e5c457262c7399c5bad9e023b2e3
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 20:28:58 2014 -0400
fix eclipse to group org/com imports together: without this, its madness
commit 09d8c1585ee99b6e63be032732c04ef6fed84ed2
Author: Robert Muir <rmuir@apache.org>
Date: Sat Nov 1 14:27:41 2014 -0400
remove nocommit, if you dont liek it, print assembly and tell me how it can be better
commit 8a6a294313fdf33b50c7126ec20c07867ecd637c
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 20:01:55 2014 +0100
Remove deprecated usage of DocIdSets.newDocIDSet.
commit 601bee60543610558403298124a84b1b3bbd1045
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 14:13:18 2014 -0400
maybe one of these zillions of annotations will stop thread leaks
commit 9d3f69abc7267c5e455aefa26db95cb554b02d62
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 14:05:39 2014 -0400
fix some analysis nocommits
commit 312e3a29c77214b8142d21c33a6b2c2b151acf9a
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 18:28:45 2014 +0100
Remove XConstantScoreQuery/XFilteredQuery/ApplyAcceptedDocsFilter.
commit 5a0cb9f8e167215df7f1b1fad11eec6e6c74940f
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 17:06:45 2014 +0100
Fix misleading documentation of DocIdSets.toCacheable.
commit 8b4ef2b5b476fff4c79c0c2a0e4769ead26cf82b
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 17:05:59 2014 +0100
Fix CustomRandomAccessFilterStrategy to override the right method.
commit d7a9a407a615987cfffc651f724fbd8795c9c671
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 16:21:35 2014 +0100
Better handle the special case when there is a single SHOULD clause.
commit 648ad389f07e92dfc451f345549c9841ba5e4c9a
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 15:53:38 2014 +0100
Cut over XBooleanFilter to BitDocIdSet.Builder.
The idea is similar to what happened to Lucene's BooleanFilter.
Yet XBooleanFilter is a bit more sophisticated and I had to slightly
change the way it is implemented in order to make it work. The main difference
with before is that slow filters are now applied lazily, so eg. if you have 3
MUST clauses, two with a fast iterator and the third with a slow iterator, the
previous implementation used to apply the fast iterators first and then only
check the slow filter for bits which were set in the bit set. Now we are
computing a bit set based on the fast must clauses and then basically returning
a BitsFilteredDocIdSet.wrap(bitset, slowClause).
Other than that, BooleanFilter still uses the bitset optimizations when or-ing
and and-ind filters.
Another improvement is that BooleanFilter is now aware of the cost API.
commit b2dad312b4bc9f931dc3a25415dd81c0d9deee08
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 10:18:53 2014 -0400
clear nocommit
commit 4851d2091e744294336dfade33906c75fbe695cd
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 15:15:16 2014 +0100
cut over to RoaringDocIdSet
commit ca6aec24a901073e65ce4dd6b70964fd3612409e
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:57:30 2014 +0100
make nocommit more explicit
commit d0742ee2cb7a6c48b0bbb31580b7fbcebdb6ec40
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 09:55:24 2014 -0400
fix standardtokenizer nocommit
commit 7d6faccafff22a86af62af0384838391d46695ca
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:54:08 2014 +0100
fix compilation
commit a038a405c1ff6458ad294e6b5bc469e622f699d0
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:53:43 2014 +0100
fix compilation
commit 30c9e307b1f5d80e2deca3392c0298682241207f
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:52:35 2014 +0100
fix compilation
commit e5139bc5a0a9abd2bdc6ba0dfbcb7e3c2e7b8481
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 09:52:16 2014 -0400
clear nocommit here
commit 85dd2cedf7a7994bed871ac421cfda06aaf5c0a5
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:46:17 2014 +0100
fix CompletionPostingsFormatTest
commit c0f3781f616c9b0ee3b5c4d0998810f595868649
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 09:38:00 2014 -0400
add tests for these analyzers
commit 51f9999b4ad079c283ae762c862fd0e22d00445f
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 14:10:26 2014 +0100
remove nocommit - this is not an issue
commit fd1388fa03e622b0738601c8aeb2dbf7949a6dd2
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Fri Oct 31 14:07:01 2014 +0100
Remove redundant null check
commit 3d6dd51b0927337ba941a235446b22e8cd500dc3
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Fri Oct 31 14:01:37 2014 +0100
Removed the work around to prevent p/c error when invoking #iterator() twice, because the custom query filter wrapper now doesn't transform the result to a cache doc id set any more.
I think the transforming to a cachable doc id set in CustomQueryWrappingFilter isn't needed at all, because we use the DocIdSet only once and because of that is just slowed things down.
commit 821832a537e00cd1216064b379df3e01d2911d3a
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 13:54:33 2014 +0100
one more nocommit
commit 77eb9ea4c4ea50afb2680c29682ddcb3851a9d4f
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Fri Oct 31 13:52:29 2014 +0100
Remove cast
commit a400573c034ed602221f801b20a58a9186a06eae
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 13:49:24 2014 +0100
fix stop filter
commit 51746087cf8ec34c4d20aa05ba8dbff7b3b43eec
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 13:21:36 2014 +0100
fix changed semantics of FBS.nextSetBit to check for NO_MORE_DOCS
commit 8d0a4e2511310f1293860823fe3ba80ac771bbe3
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 08:13:44 2014 -0400
do the bogus cast differently
commit 46a5cc5732dea096c0c80ae5ce42911c9c51e44e
Author: Simon Willnauer <simonw@apache.org>
Date: Fri Oct 31 13:00:16 2014 +0100
I hate it but P/C now passes
commit 580c0c2f82bbeacf217e594f22312b11d1bdb839
Merge: a9d3c00 1645434
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 31 06:54:31 2014 -0400
fix nocommit/classcast
commit a9d3c004d62fe04989f49a897e6ff84973c06eb9
Author: Adrien Grand <jpountz@gmail.com>
Date: Fri Oct 31 08:49:31 2014 +0100
Update TODO.
commit aa75af0b407792aeef32017f03a6f442ed970baa
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 19:18:25 2014 -0400
clear obselete nocommits from lucene bump
commit d438534cf41fcbe2d88070e2f27c994625e082c2
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 18:53:20 2014 -0400
throw classcastexception when ES abuses regular filtercache for nested docs
commit 2c751f3a8feda43ec127c34769b069de21f3d16f
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 18:31:34 2014 -0400
bump lucene revision, fix tests
commit d6ef7f6304ae262bf6228a7d661b2a452df332be
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 22:37:58 2014 +0100
fix merge problems
commit de9d361f88a9ce6bb3fba85285de41f223c95767
Merge: 41f6aab f6b37a3
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 22:28:59 2014 +0100
Merge branch 'master' into enhancement/lucene_5_0_upgrade
Conflicts:
pom.xml
src/main/java/org/elasticsearch/Version.java
src/main/java/org/elasticsearch/gateway/local/state/meta/MetaDataStateFormat.java
commit 41f6aab388aa80c40b08a2facab2617576203a0d
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 17:48:46 2014 +0100
fix potiential NPE
commit c4428b12e1ae838b91e847df8b4a8be7f49e10f4
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 17:38:46 2014 +0100
don't advance iterator in a match(doc) method
commit 28ab948e99e3ea4497c9b1e468384806ba7e1790
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 17:34:58 2014 +0100
don't advance iterator in a match(doc) method
commit eb0f33f6634fadfcf4b2bf7327400e568f0427bb
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 16:55:54 2014 +0100
fix GeoUtilsTest
commit 7f711fe3eaf73b6c2268cf42d5a41132a61ad831
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 16:43:16 2014 +0100
Use a dedicated default index option if field type is not indexed by default
commit 78e3f37ab779e3e1b25b45a742cc86ab5f975149
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 10:56:14 2014 -0400
disable this test with AwaitsFix to reduce noise
commit 9a590f563c8e03a99ecf0505c92d12d7ab20d11d
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 09:38:49 2014 +0100
fix lucene version
commit abe3ca1d8bb6b5101b545198f59aec44bacfa741
Author: Simon Willnauer <simonw@apache.org>
Date: Thu Oct 30 09:35:05 2014 +0100
fix AnalyzingCompletionLookupProvider to wrok with new codec API
commit 464293b245852d60bde050c6d3feb5907dcfbf5f
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 00:26:00 2014 -0400
don't try to write stuff to tests class directory
commit 031cc6c19f4fe4423a034b515f77e5a0e282a124
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 00:12:36 2014 -0400
AwaitsFix these known issues to reduce noise
commit 4600d51891e35847f2d344247d6f915a0605c0d1
Author: Robert Muir <rmuir@apache.org>
Date: Thu Oct 30 00:06:53 2014 -0400
openbitset lives on
commit 8492bae056249e2555d24acd55f1046b66a667c4
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 23:42:54 2014 -0400
fixes for filter tests
commit 31f24ce4efeda31f97eafdb122346c7047a53bf2
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 23:12:38 2014 -0400
don't use fieldcache
commit 8480789942fdff14a6d2b2cd8134502fe62f20c8
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 23:04:29 2014 -0400
ancient index no longer supported
commit 02e78dc7ebdd827533009f542582e8db44309c57
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 23:37:02 2014 +0100
fix more tests
commit ff746c6df23c50b3f3ec24922413b962c8983080
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 23:08:19 2014 +0100
fix all mapper
commit e4fb84b517107b25cb064c66f83c9aa814a311b2
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 22:55:54 2014 +0100
fix distributor tests and cut over to FileStore API
commit 20c850e2cfe3210cd1fb9e232afed8d4ac045857
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 22:42:18 2014 +0100
use DOCS_ONLY if index=true and current options == null
commit 44169c108418413cfe51f5ce23ab82047463e4c2
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 22:33:36 2014 +0100
Fix index=yes|no settings in mappers
commit a3c5f77987461a18121156ed345d42ded301c566
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 21:51:41 2014 +0100
fix several field mappers conversion from setIndexed to indexOptions
commit df84d736908e88a031d710f98e222be68ae96af1
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 21:33:35 2014 +0100
fix SourceFieldMapper to be not indexed
commit b2bf01d12a8271a31fb2df601162d0e89924c8f5
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 21:23:08 2014 +0100
Cut over to .liv files in store and corruption tests
commit 619004df436f9ef05d24bef1b6a7f084c6b0ad75
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 17:05:52 2014 +0100
fix more tests
commit b7ed653a8b464de446e00456bce0a89e47627c38
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 16:19:08 2014 +0100
[STORE] Add dedicated method to write temporary files
Recovery writes temporary files which might not end up in the
right distributor directories today. This commit adds a dedicated
API that allows specifying the target file name in order to create the
tempoary file in the correct directory.
commit 7d574659f6ae04adc2b857146ad0d8d56ca66f12
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 10:28:49 2014 -0400
add some leniency to temporary bogus method
commit f97022ea7c2259f7a5cf97d924c59ed75ab65b32
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 10:24:17 2014 -0400
fix MultiCollector bug
commit b760533128c2b4eb10ad76e9689ef714293dd819
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:56:08 2014 +0100
CheckIndex is now closeable we need to close it
commit 9dae9fb6d63546a6c2427be2a2d5c8358f5b1934
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:45:11 2014 +0100
s/Lucene51/Lucene50
commit 7aea9b86856a8c1b06a08e7c312ede1168af1287
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:42:30 2014 +0100
fix BloomFilterPostingsFormat
commit 16fea6fe842e88665d59cc091e8224e8dc6ce08c
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:41:16 2014 +0100
fix some codec format issues
commit 3d77aa97dd2c4012b63befef3f2ba2525965e8a6
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:30:43 2014 +0100
fix CodecTests
commit 6ef823b1fde25657438ace1aabd9d552d6ae215e
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:26:47 2014 +0100
make it compile
commit 9991eee1fe99435118d4dd42b297ffc83fce5ec5
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 09:12:43 2014 -0400
add an ugly hack for TopHitsAggregator for now
commit 03e768a01fcae6b1f4cb50bcceec7d42977ac3e6
Author: Simon Willnauer <simonw@apache.org>
Date: Wed Oct 29 14:01:02 2014 +0100
cut over ES090PostingsFormat
commit 463d281faadb794fdde3b469326bdaada25af048
Merge: 0f8740a 8eac79c
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 08:30:36 2014 -0400
Merge branch 'master' into enhancement/lucene_5_0_upgrade
commit 0f8740a782455a63524a5a82169f6bbbfc613518
Author: Robert Muir <rmuir@apache.org>
Date: Wed Oct 29 01:00:15 2014 -0400
fix/hack remaining filter and analysis issues
commit df534488569da13b31d66e581456dfd4b55156b9
Author: Robert Muir <rmuir@apache.org>
Date: Tue Oct 28 23:11:47 2014 -0400
fix ngrams / openbitset usage
commit 11f5dc3b9887f4da80a0fa1818e1350b30599329
Author: Robert Muir <rmuir@apache.org>
Date: Tue Oct 28 22:42:44 2014 -0400
hack over sort comparators
commit 4ebdc754350f512596f6a02770d223e9f5f7975a
Author: Robert Muir <rmuir@apache.org>
Date: Tue Oct 28 21:27:07 2014 -0400
compiler errors < 100
commit 2d60c9e29de48ccb0347dd87f7201f47b67b83a0
Author: Robert Muir <rmuir@apache.org>
Date: Tue Oct 28 03:13:08 2014 -0400
clear some nocommits around ram usage
commit aaf47fe6c0aabcfb2581dd456fc50edf871da758
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 12:27:34 2014 -0400
migrate fieldinfo handling
commit ef6ed6d15d8def71cd880d97249678136cd29fe3
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 12:07:13 2014 -0400
more simple fixes
commit f475e1048ae697dd9da5bd9da445102b0b7bc5b3
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 11:58:21 2014 -0400
more fielddata ram accounting fixes
commit 16b4239eaa9b4262df258257df4f31d39f28a3a2
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 16:47:32 2014 +0100
add missing file
commit 5b542fa2a6da81e36a0c35b8e891a1d8bc58f663
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 16:43:29 2014 +0100
cut over completion posting formats - still some nocommits
commit ecdea49404c4ec4e1b78fb54575825f21b4e096e
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 11:21:09 2014 -0400
fielddata accountable fixes
commit d43da265718917e20c8264abd43342069198fe9c
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 16:19:53 2014 +0100
cut over BloomFilterPostings to new API
commit 29b192ba621c14820175775d01242162b88bd364
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 10:22:51 2014 -0400
fix more analyzers
commit 74b4a0c5283e323a7d02490df469497c722780d2
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 09:54:25 2014 -0400
fix tests
commit 554084ccb4779dd6b1c65fa7212ad1f64f3a6968
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 14:51:48 2014 +0100
maintain supressed exceptions on CorruptIndexException
commit cf882d9112c5e8ef1e9f2b0f800f7aa59001a4f2
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 14:47:17 2014 +0100
commitOnClose=false
commit ebb2a9189ab2f459b7c6c9985be610fd90dfe410
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 14:46:06 2014 +0100
cut over indexwriter closeing in InternalEngine
commit cd21b3d4706f0b562bd37792d077d60832aff65f
Author: Simon Willnauer <simonw@apache.org>
Date: Mon Oct 27 14:38:10 2014 +0100
fix constant
commit f93f900c4a1c90af3a21a4af5735a7536423fe28
Author: Robert Muir <rmuir@apache.org>
Date: Mon Oct 27 09:50:49 2014 -0400
fix test
commit a9a752940b1ab4699a6a08ba8b34afca82b843fe
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Mon Oct 27 09:26:18 2014 +0100
Be explicit about the index options
commit d9ee815babd030fa2ceaec9f467c105ee755bf6b
Author: Simon Willnauer <simonw@apache.org>
Date: Sun Oct 26 20:03:44 2014 +0100
cut over store and directory
commit b3f5c8e39039dd8f5caac0c4dd1fc3b1116e64ca
Author: Robert Muir <rmuir@apache.org>
Date: Sun Oct 26 13:08:39 2014 -0400
more test fixes
commit 8842f2684e3606aae0860c27f7a4c53e273d47fb
Author: Robert Muir <rmuir@apache.org>
Date: Sun Oct 26 12:14:52 2014 -0400
tests manual labor
commit c43de5aec337919a3fdc3638406dff17fc80bc98
Author: Robert Muir <rmuir@apache.org>
Date: Sun Oct 26 11:04:13 2014 -0400
BytesRef -> BytesRefBuilder
commit 020c0d087a2f37566a1db390b0e044ebab030138
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Sun Oct 26 15:53:37 2014 +0100
Moved over to BitSetFilter
commit 48dd1b909e6c52cef733961c9ecebfe4f67109fe
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Sun Oct 26 15:53:11 2014 +0100
Left over Collector api change in ScanContext
commit 6ec248ef63f262bcda400181b838fd9244752625
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Sun Oct 26 15:47:40 2014 +0100
Moved indexed() over to indexOptions != null or indexOptions == null
commit 9937aebfd8546ae4bb652cd976b3b43ac5ab7a63
Author: Martijn van Groningen <martijn.v.groningen@gmail.com>
Date: Sun Oct 26 13:26:31 2014 +0100
Fixed many compile errors. Mainly around the breaking Collector api change in 5.0.
commit fec32c4abc0e3309cf34260c8816305a6f820c9e
Author: Robert Muir <rmuir@apache.org>
Date: Sat Oct 25 11:22:17 2014 -0400
more easy fixes
commit dab22531d801800d17a65dc7c9464148ce8ebffd
Author: Robert Muir <rmuir@apache.org>
Date: Sat Oct 25 09:33:41 2014 -0400
more progress
commit 414767e9a955010076b0497cc4f6d0c1850b48d3
Author: Robert Muir <rmuir@apache.org>
Date: Sat Oct 25 06:33:17 2014 -0400
more progress
commit ad9d969fddf139a8830254d3eb36a908ba87cc12
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 24 14:28:01 2014 -0400
current state of fun
commit 464475eecb0be15d7d084135ed16051f76a7e521
Author: Robert Muir <rmuir@apache.org>
Date: Fri Oct 24 11:42:41 2014 -0400
bump to 5.0 snapshot
Currently MetaDataStateFormat loads the first available state file that has
the latest version. In case several files are available and some of them use
the new format while other ones use the legacy format, it should also prefer
the new format. This is typically useful when we upgrade the metadata when
recovering from the gateway: we might write the upgraded state with the new
format while the previous state used the legacy format, so we end up with
two files having the same version but using different formats.
Close#8343
This will help the exists/missing filters behave as expected in presence of
empty strings, as well as when using a default analyzer that would generate
tokens for an empty string (uncommon).
Close#8198
The transport client created within ExternalTestCluster needs a name that follows our naming convention otherwise the thread leak filter barfs when running tests against an external cluster. Used "transport_client_external_{n}" where n gets incremented every time a new external cluster gets created. Updated thread leak filters rules to ignore threads created by such transport client.
We currently use the djb2 hash function in order to compute the shard a
document should go to. Unfortunately this hash function is not very
sophisticated and you can sometimes hit adversarial cases, such as numeric ids
on 33 shards.
Murmur3 generates hashes with a better distribution, which should avoid the
adversarial cases.
Here are some examples of how 100000 incremental ids are distributed to shards
using either djb2 or murmur3.
5 shards:
Murmur3: [19933, 19964, 19940, 20030, 20133]
DJB: [20000, 20000, 20000, 20000, 20000]
3 shards:
Murmur3: [33185, 33347, 33468]
DJB: [30100, 30000, 39900]
33 shards:
Murmur3: [2999, 3096, 2930, 2986, 3070, 3093, 3023, 3052, 3112, 2940, 3036, 2985, 3031, 3048, 3127, 2961, 2901, 3105, 3041, 3130, 3013, 3035, 3031, 3019, 3008, 3022, 3111, 3086, 3016, 2996, 3075, 2945, 2977]
DJB: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 900, 900, 900, 900, 1000, 1000, 10000, 10000, 10000, 10000, 9100, 9100, 9100, 9100, 9000, 9000, 0, 0, 0, 0, 0, 0]
Even if djb2 looks ideal in some cases (5 shards), the fact that the
distribution of its hashes has some patterns can raise issues with some shard
counts (eg. 3, or even worse 33).
Some tests have been modified because they relied on implementation details of
the routing hash function.
Close#7954
Robert Muir
Chris Mattmann
Veres Lajos
Clinton Gormley
Michael McCandless
javanna
Lee Hinman
Simon Willnauer
Boaz Leskes
Adrien Grand
Kevin Kluge
Colin Goodheart-Smithe
Martijn van Groningen
Lukas Vlcek
Britta Weber