Commit Graph

6779 Commits

Author SHA1 Message Date
Boaz Leskes 33e49a55f7 Increased timeout in ClusterServiceTests.testTimeoutUpdateTask to 100ms
The previous 2ms timeout was too small and caused a racing condition in timeout handling. This will be fixed but requires more work.
2013-12-16 10:02:24 +01:00
Simon Willnauer a668cd22d8 wait until cluster-state is ready after startup 2013-12-15 22:26:11 +01:00
Simon Willnauer d9e988aea9 Fix QuorumLocalGatewayTests#testChangeInitialShardsRecovery to start only one shard to not fulfill the quorum 2013-12-15 20:06:22 +01:00
Boaz Leskes 0f7c5dbc86 Update minimum_master_nodes comment to match documentation
Someone on the mailing list wrongly interpreted it to mean the current node is not included in the calculation.
2013-12-15 11:25:36 +01:00
Igor Motov d8ba92cfa8 Resolve potential deadlock state during EsThreadPoolExecutor shutdown
Fixes #4334

The deadlock occurs between monitor object of EsThreadPoolExecutor and mainLock of ThreadPoolExecutor. The shutdown method of EsThreadPoolExecutor obtains the lock on monitor first and waits for mainLock of ThreadPoolExecutor in ThreadPoolExecutor#shutdown for part of the processing, while EsThreadPoolExecutor#terminated is executed under mainLock and tries to obtain monitor to notify listeners.
2013-12-14 21:45:03 -05:00
Simon Willnauer 8f85d63b67 Enable incremental compilation using a workaround for the maven-compiler-plugin 3.1 bug 2013-12-14 21:56:01 +01:00
Shay Banon 80ab75e54e upgrade to latest jsr166 2013-12-14 01:27:18 +01:00
Shay Banon 59d85bc22a upgrade to Jackson 2.3.0 2013-12-14 00:53:52 +01:00
Luca Cavanna 9b121baafc More common AcknowledgedRestResponseActionListener
Introduced use of AcknowledgedRestResponseActionListener in put/delete index template
2013-12-13 19:39:46 +01:00
Simon Willnauer d684a2b8da Don't try to run check-index in MockDirectoryWrapper
if the MockDirWrapper checks the index on close it closes all
closeables that is has to crash the index if something is not flushed yet.
For us this is a problem since the input is still used. We need to fix this in
lucene first.
2013-12-13 17:44:10 +01:00
Adrien Grand 19c8d18b14 Relax a bit the accuracy tests on SloppyMaths.
Tests now ensure that the computed distance is correct within 1‰ instead of 1‱.
2013-12-13 15:42:39 +01:00
Adrien Grand 36bd9cc432 Aggregations: Ordinals-based string bucketing support.
When the ValuesSource has ordinals, terms ordinals are used as a cache key to
bucket ordinals. This can make terms aggregations on String terms significantly
faster.

Close #4350
2013-12-13 15:34:02 +01:00
Martijn van Groningen cc61348013 Fixed test bug 2013-12-13 14:54:21 +01:00
Martijn van Groningen 10e2528cce Added the `force_source` option to highlighting that enforces to use of the _source even if there are stored fields.
The percolator uses this option to deal with the fact that the MemoryIndex doesn't support stored fields,
this is possible b/c the _source of the document being percolated is always present.

Closes #4348
2013-12-13 13:39:53 +01:00
Simon Willnauer 59e4e58683 Add trace logging to AllocationDeciders if decision is NO() 2013-12-13 12:08:26 +01:00
Sebastian Geidies 6ed126deaf Skipping execution of remaining Deciders if one of them returns a Decision.NO.
Reordering execution of the different AllocationDeciders, so the cheapest get executed first.
2013-12-13 11:52:48 +01:00
Boaz Leskes 2acccf257d Add `queries` to XContent output of PercolateStats
The queries stats tracked but not exposed to the Rest API

Closes #4440
2013-12-13 11:36:51 +01:00
Boaz Leskes b65d07b495 A failure during the response construction could cause a Node Operation never to return 2013-12-13 10:36:39 +01:00
Diego Ongaro a4814c2f69 Print nice error in bin/elasticsearch if user needs to run maven
Before, people that cloned the repo and expected to be able to run
bin/elasticsearch would be met with an awful shell error: the shell
interprets Maven variables like ${project.build.finalName} as shell
variables yet can't handle names of the form ${x.y}. This commit
explicitly checks to make sure that Maven has done its substitutions
before continuing; if Maven hasn't been run, it gives a helpful error
message.

Fixes #2954.
2013-12-13 10:30:42 +01:00
Shay Banon 976bdd6a26 randomize transport tcp compression on nodes 2013-12-12 23:19:59 +01:00
Shay Banon 0469cca4be properly support global transport compression with new bytes transport request optimization 2013-12-12 22:29:54 +01:00
Shay Banon 0f4d81fcd4 refresh the mapping source directly into a compressed buffer
instead of building the mapping source, and then compressing it, we can generate it directly into a compressed buffer
2013-12-12 20:58:11 +01:00
Lee Hinman 77fcf71338 Add new `simple_query_string` query type
This adds support for Lucene's SimpleQueryParser by adding a new type
of query called the `simple_query_string`. The `simple_query_string`
query is designed to be able to parse human-entered queries without
throwing any exceptions.

Resolves #4159
2013-12-12 12:09:32 -07:00
Alexander Reelsen 348c3731c8 Fixing debian init script to work on ubuntu LTS
sysctl parameters must be separated like `-q -w` instead of `-qw`
2013-12-12 17:57:34 +01:00
Adrien Grand e2e19203d7 Remove IndexFieldData.getHighestNumberOfSeenUniqueValues().
Closes #4426
2013-12-12 15:44:05 +01:00
Shay Banon f0e1f9215b try and wait better for all mapping updates to be processes in test 2013-12-12 15:06:37 +01:00
Simon Willnauer 198d36dc15 Also wrap IndexInput if CFS is used 2013-12-12 12:14:47 +01:00
Simon Willnauer d95420cc67 Add more information to IndexInput failures related to abusing closed input 2013-12-12 10:44:01 +01:00
Adrien Grand 0554f8e875 Add missing mappers to MapperBuilders.
- completion
 - geo_point
 - geo_shape
 - parent
 - size
 - ttl

Close #4423
2013-12-12 09:44:55 +01:00
Clinton Gormley b8aba737f4 [TEST] Added skip clause to analyze API with text format.
No longer supported in 1.0.0.RC1
2013-12-11 22:43:18 +00:00
Martijn van Groningen b4d1d899d2 Replaced the synchronization method for enabling the realtime percolator from synchronized methods to use AtomicBoolean#compareAndSet instead.
Remove @TestLogging from percolator tests.
2013-12-11 22:54:35 +01:00
Luca Cavanna 2cbecd9cb3 Made sure that we never throw IndexMissingException in indices query and filter
It could happen although we internally use IgnoreIndices.MISSING, due to MetaData#concreteIndices contract, which throws IndexMissingException anyway if all requested indices are missing.
In case all the indices specified in the query/filter are missing, we just execute the no_match query/filter, no need to throw any error.

Closes #3428
2013-12-11 22:26:46 +01:00
Luca Cavanna 24434063b6 Introduced ok & acknowledged constant in AcknowledgedRestResponseActionListener 2013-12-11 21:28:54 +01:00
Luca Cavanna b1fa147968 More common AcknowledgedRestResponseActionListener
Introduced use of AcknowledgedRestResponseActionListener in put/delete alias & delete index
2013-12-11 21:28:15 +01:00
Luca Cavanna 037bf6a85f Unified default ack timeout to 30 seconds
Increased also default publish state timeout to 30 seconds (from 5 seconds) and introduced constant for it.
Introduced AcknowledgedRequest.DEFAULT_ACK_TIMEOUT constant.
Removed misleading default values coming from the REST layer.
Removed (in a bw compatible manner) the timeout support in put/delete index template as the timeout parameter was ignored.

Closes #4395
2013-12-11 20:54:27 +01:00
Clinton Gormley 13d9d17155 Update README with full command to generate spec 2013-12-11 18:33:45 +00:00
Martijn van Groningen f9ff733eb8 Renamed the methods in DocumentListenerType to actually reflect when these are invoked. 2013-12-11 17:51:09 +01:00
Martijn van Groningen 92c32dca9e Invoke the DocumentType listeners *before* the mappers are updated. This fixes an issue where if two or more concurrent percolate index requests are processed, the first request would the `.percolator` type mapping, but the real time percolator listener wouldn't be active, this would result in that the subsequent concurrent requests wouldn't be parsed and kept in memory and would never be included in any percolate api result. This issue any occurred when `.percolator` type is created on the fly.
Also made the call to PercolatorQueriesRegistry#enableRealTimePercolator and #disableRealTimePercolator synchronized, so that for the same shard the RealTimePercolatorOperationListener can't registered twice.
2013-12-11 17:41:10 +01:00
Shay Banon a3f1c428c2 bulk response has errors indication + status per item
closes #4002
2013-12-11 16:04:41 +01:00
Boaz Leskes ea4a908bdb Added field mapping API tests which use wild cards
See: https://github.com/elasticsearch/elasticsearch/issues/4367
2013-12-11 15:00:33 +01:00
Shay Banon 10cdb0ae22 Batch processing mapping updates can cause missed merged mappings when batching multiple types
when we bulk changes, we need to use the same index metadata builder across the tasks, otherwise we might remove mappings erroneously
 also, when we check if we can use a higher order mapping, we need to verify that its for the same mapping type
2013-12-11 14:34:59 +01:00
Martijn van Groningen a760f1f54a Make listeners a final field in ShardIndexingService, which fixed possible visibility issue. 2013-12-11 12:22:45 +01:00
Shay Banon aee388ec46 introduce an internal metadata uuid 2013-12-11 10:49:36 +01:00
Florian Schilling 937a4e9d9d SloppyMath
Added copy of SloppyMath.java from lucene 4.6+
and setup GeoDistance for new haversin method

closes #3862
2013-12-11 17:51:32 +09:00
Alexander Reelsen 81e13a870b Packaging: Ensure setting of sysctl vm.max_map_count
In order to be sure that memory mapped lucene directories are working
one can configure the kernel about how many memory mapped areas
a process may have. This setting ensure for the debian and redhat initscripts
as well as the systemd startup, that this setting is set high enough.

Closes #4397
2013-12-11 09:19:22 +01:00
Boaz Leskes 99b421925f Add wildcard support to field resolving in the Get Field Mapping API
Closes #4367
2013-12-10 23:46:37 +01:00
Shay Banon a9e259d438 Introduce BytesTransportRequest, allowing for downstream network optimization in buffers usage
When sending a request, mainly to multiple nodes, if we already have the "body" of the request in bytes, we can share it instead of copying it over to a new buffer. Also, it helps a lot when sending a relatively large body to multiple nodes, since it will use the same body buffer across all nodes
2013-12-10 22:49:31 +01:00
Alexander Reelsen c4f3da2b9d Preference only_node with unknown nodeId returns useful exception
When the search preference is set to only node, but this node is not a
data (or does not exist), we return a search exception, which indicates,
that this is actually a server problem.

However specifying a non-existing node id is a client problem
and should return a more useful error message than
{"error":"SearchPhaseExecutionException[Failed to execute phase [query_fetch], all shards failed]","status":503}
2013-12-10 20:15:14 +01:00
Martijn van Groningen 7d3b78c293 moved log statements 2013-12-10 19:54:27 +01:00
Richard Boulton 1037d071bf Fix bug in explain for function_score queries.
The explain output for function_score queries with score_mode=max or
score_mode=min was incorrect, returning instead the value of the last
function.  This change fixes this.
2013-12-10 17:59:00 +01:00