e7fdf16402
add number of data nodes to cluster health
2010-10-28 00:39:53 +02:00
d487d809ea
Query DSL: Revert back and have range, terms, and prefix filter _cache set to true, closes #456 .
2010-10-27 20:34:24 +02:00
3a9dc8f452
change back to cached thread pool
2010-10-27 20:13:42 +02:00
185f5a9e18
Search: Change default operation threading to `thread_per_shard` from `single_thread`, closes #455 .
2010-10-27 19:51:49 +02:00
ca67c12de5
add and/not/or docid sets, not just docsets, and improve caching behavior
2010-10-27 18:16:15 +02:00
8fef3df16f
improve threading usage on discovery and fd
2010-10-26 23:49:09 +02:00
73e6aa72cd
for some reason the test only runs with @Test on the main class...
2010-10-26 20:20:58 +02:00
c08ad0cad6
Fix parsing of "geo_distance" filter wrt "distance" / "unit" parameters
...
The problem was that when "unit" was given,
the conversion to miles was happening too early,
which caused wrong computations. This change
postpones this computation when one really knows
which unit should be used.
2010-10-26 20:13:29 +02:00
9b206146e3
Fix MILES.toKilometers() and KILOMETERS.toMiles()
...
This mistake should have been caught by DistanceUnitTests.
But the problem is that the tests in this file does not
run during the execution of the test suite, and I don't have
a clue why this is so.
2010-10-26 20:13:29 +02:00
2d15cd8009
Another test for parsing "geo_distance" filter: 12mi and km
...
If an explicit unit is provided with "distance",
the "unit" can be safely ignored, as it works
as a fallback unit.
2010-10-26 20:13:29 +02:00
8c8b7bee3a
Three tests for parsing "geo_distance" filter: 12 miles in km
...
The added test files should be equivalent. Actually they
hit the same bug as change
Two tests for parsing "geo_distance" filter: distance/unit parameters
2010-10-26 20:13:29 +02:00
244a566e49
Two tests for parsing "geo_distance" filter: default unit is km
2010-10-26 20:13:29 +02:00
0660a2912e
Two tests for parsing "geo_distance" filter: distance/unit parameters
...
Those two are supposed to be equivalent:
distance: 12, unit: "mi"
vs
distance: "12", unit: "mi"
but they are not because of an underlying bug in the query parsing
code, providing non-equivalent behavior whether a number or a string
comes via JSON.
2010-10-26 20:13:29 +02:00
14703aa1be
doc
2010-10-26 20:05:57 +02:00
184276154c
add search stress test
2010-10-26 17:17:12 +02:00
5804e9132a
Query DSL: `term` filter to have `_cache` set to `true` by default, closes #453 ,
2010-10-26 13:40:14 +02:00
442f1d7677
add the request id to the timeout message
2010-10-26 11:22:02 +02:00
c59bfea43c
add the request id to the timeout message
2010-10-26 11:19:31 +02:00
529303c533
Query: match_all query (also in query_string: *:*) is very slow
2010-10-26 11:04:55 +02:00
5966d2005d
refacto package name from facets to facet (make more sense when facet will be easily pluggable)
2010-10-25 18:49:30 +02:00
0a0ac35792
bring back query parsing cache (weak)
2010-10-25 18:09:40 +02:00
2c65be2aad
better exception when a field is missing
2010-10-25 16:33:55 +02:00
cccefad091
upgrade to netty 3.2.3
2010-10-25 16:15:05 +02:00
bc4121c06b
Query DSL: `term`, `term`, `prefix`, and `range` filter are now weakly cached, for more strong caching, set `_cache` to true, closes #450 .
2010-10-25 15:59:15 +02:00
5c5b882b5d
clean code
2010-10-25 14:44:47 +02:00
517f3eb0fd
chance default query parser cache to none from weak
2010-10-25 14:43:06 +02:00
deada942e5
Query DSL: `constant_score` and `filtered` queries cache filters by default, remove it, closes #449 .
2010-10-25 13:52:15 +02:00
835a81c16c
always use deletion aware constant score
2010-10-25 13:35:43 +02:00
b7eacc3618
add comment / todo
2010-10-24 12:38:24 +02:00
9a80fedf52
add simple map stream capability
2010-10-23 23:19:08 +02:00
ca7a7467dc
Indices API: Add open and close index, closes #447 .
2010-10-23 21:52:09 +02:00
a76824e395
make index not recovered a bock used in both gateways (shared/blob and local)
2010-10-23 17:03:38 +02:00
3f8c03db95
add test in node allocation is there are no shards, no need to do local gateway allocation
2010-10-23 01:25:23 +02:00
f0f62ce00c
improve test to wait for 2 nodes
2010-10-23 01:24:53 +02:00
e11589f3c7
add doc
2010-10-23 01:07:37 +02:00
62526444fd
better caching of missing filter / query
2010-10-22 19:44:25 +02:00
b0a0604473
Query DSL: Query String _missing_ and _exists_ syntax, closes #446 .
2010-10-22 18:14:07 +02:00
8b0daf48f4
Query DSL: Query String _missing_ and _exists_ syntax, closes #446 .
2010-10-22 17:54:02 +02:00
9237dafef9
Query DSL: Query String _missing_ and _exists_ syntax, closes #446 .
2010-10-22 17:51:19 +02:00
f63ee3158a
Query DSL: Add `exists` and `missing` filters to filter documents where a field either has a value or not in them, closes #445 .
2010-10-22 15:14:05 +02:00
9a8e033424
improve fixing negative query with adding the optimized match_all query
2010-10-22 14:50:48 +02:00
5d1d927e09
Write Consistency Level for index/delete/delete_by_query/bulk with one/quorum/all. Defaults to quorum, closes #444 .
2010-10-22 02:51:00 +02:00
5649df572a
update search expiration to work in scheduled reaper mode
2010-10-21 16:38:44 +02:00
c37a0afbf0
improve update of mappings, and delete index process with the new optimization of when index is created
2010-10-20 12:49:21 +02:00
0b593bd6d9
upgrade to guava r07
2010-10-20 12:09:19 +02:00
ae5bc20959
Reduce multiple indices overhead, don't create Indices data on nodes that don't hold any shard for an index, closes #441 .
2010-10-20 02:57:49 +02:00
1235358848
remove unused indices service
2010-10-19 19:50:12 +02:00
a4eea0aeaa
change index creation / deletion logic not to wait for it to be created on other nodes since now, they might not be created...
2010-10-19 19:33:27 +02:00
69b8b0f437
don't use the index serivce to get the stored blobs in shared gateway, move it to upper node level
2010-10-19 18:03:48 +02:00
53a3df5d8e
move list of local files to be indices level, so no need for index created on a node in order to list it
2010-10-19 17:36:21 +02:00
5635260922
remove check on if indices has an index, cluster state is enough
2010-10-19 16:42:39 +02:00
4c8978237f
move operation routing to be cluster level and not index level
2010-10-19 16:23:39 +02:00
1eee7f381a
dont case node names and index names
2010-10-18 19:03:48 +02:00
5c760d1e8c
don't case the index name in cluster health, use string constants in xcontent generation
2010-10-18 18:55:38 +02:00
3a52c2f7f4
Improve automatic publish network address logic, closes #439 .
2010-10-18 15:19:24 +02:00
019c844dd1
Plugins: plugins should now be in extracted format under `ES_HOME/plugins`, closes #438 .
2010-10-18 11:14:16 +02:00
0a3d187e6a
Mapper: Allow to configure `date_formats` only on the root object mapper, closes #437 .
2010-10-17 19:56:58 +02:00
69ecf8b66b
rollback commit, allocation should only happen when index is not recovered
2010-10-17 19:43:30 +02:00
9a09261383
remove checking for client node on rivers, its automatically done with master and non river node anyhow
2010-10-17 19:20:37 +02:00
648da2e6ff
add cluster name to cluster health response
2010-10-17 18:58:59 +02:00
5d54e36964
Statistical Facet: Allow to compute statistical facets on more than one field, closes #436 .
2010-10-17 16:22:20 +02:00
2fc0022477
clean code
2010-10-17 02:33:12 +02:00
caa6495e2b
doc
2010-10-17 01:52:23 +02:00
4708d31040
add the option to control filter caching on each filter in the java filter builders
2010-10-16 16:12:38 +02:00
d9f966d83c
Query DSL: Numeric Range Filter - A filter that uses the field data cache to perform numeric tests, closes #435 .
2010-10-16 15:38:43 +02:00
cc3a1dbe90
move geo handling (filter) into the index package from common.lucene since it uses field data and not really only dependant on lucene constucts
2010-10-16 12:30:20 +02:00
1587a3815e
more options on stress rolling restart test, variable number of fields, variable length of text
2010-10-16 03:26:11 +02:00
b09675dcfa
add a stress test for rolling restart nodes
2010-10-16 01:05:25 +02:00
bbb05a57e0
add the ability to run check index before starting a shard, used for internal testing (as it has a large overhead)
2010-10-15 20:56:30 +02:00
6c9120a51b
Local Gateway: Don't block meta operations (delete index) on an index that is not recovered due to not all shards being available, closes #433 .
2010-10-15 11:04:54 +02:00
ae05ce0551
Scripting: Allow to define scripts within `config/scripts`, automatically compiled and can be referenced by name, closes #429 .
2010-10-14 16:14:23 +02:00
e0488fa87e
Plugins: Allow to place "extracted" plugins under the `plugins` directory, closes #428 .
2010-10-14 14:04:00 +02:00
8d533e8a99
dynamic tempaltes are now named (change to the mapping definition) to allow for simpler merging
2010-10-14 12:56:50 +02:00
ad01f19db8
move dynamic tempaltes to only allow to be defined on root object mapping (so easily reusable and simpler merging) #397 .
2010-10-14 12:25:32 +02:00
1578da404c
improve geo internal cache of lat/lon, stored two double arrays instead of an array of GeoPoints.
2010-10-13 22:03:48 +02:00
9077bb6528
Geo Overhaul (work with multiple locations), closes #414 . Change (again) how geo location is stored, not using geohash but explicit double conversion
2010-10-13 03:35:35 +02:00
8a8a4d648a
Query DSL: Fine grained filter caching control with sensible defaults, closes #424 .
2010-10-13 01:37:43 +02:00
3fa214653e
increase precision to 16 from 12, omit norms / tf on geohash
2010-10-12 22:30:56 +02:00
c3cb5a3e34
Use Base64 encoding for UUID generation (auto generated doc ids, and node ids), closes #422 .
2010-10-12 00:17:01 +02:00
4f9d9c1395
add test for merging two stats field facet using same facet name
2010-10-11 00:40:37 +02:00
18d8e9dcdd
add more script tests
2010-10-10 20:05:37 +02:00
a14b73bc13
Bulk API: Rename `index` to `_index`, `type` to `_type` and `id` to `_id`, closes #421 .
2010-10-10 19:31:56 +02:00
c15a612b8a
optimize get/delete/index response xcontent generation
2010-10-09 04:06:03 +02:00
087b4e6f23
optimize search response xcontent generation
2010-10-09 04:01:17 +02:00
f64e636415
add xcontent string that will allow to speed up json/smile serialization
2010-10-09 03:06:01 +02:00
bd962ba693
upgrade to jackson 1.6.1
2010-10-09 01:51:40 +02:00
6314c2460c
Geo Overhaul (work with multiple locations), closes #414 .
2010-10-08 22:31:21 +02:00
523a8b4c3e
rename xonctent mappers to just mappers
2010-10-08 15:05:07 +02:00
a61f3a939b
add option to reuse double[] when decoding geohash
2010-10-07 23:58:49 +02:00
0f6beeb263
reuse search lookup data for the two remaining cases (custom score and script filter) by having a "current" search context to access
2010-10-07 23:18:26 +02:00
34ed85a40f
internal refactoring of field data type
2010-10-07 18:41:17 +02:00
d0bf743ab4
Query DSL: query_string / field to use the optimized match_all query when using * (or *:*), closes #413 .
2010-10-07 16:48:24 +02:00
cf5ed1d177
Term Facet: Use a script that provides the facets to index, closes #410 .
2010-10-07 01:49:12 +02:00
4c34840627
fix dynamic tempaltes persistency and merging (so it will be there across restarts, and when a new mapping is merged)
2010-10-07 00:44:53 +02:00
4422a4c48a
doc
2010-10-06 14:07:34 +02:00
22ed71e2e6
allow to execute term facet on more than one field
2010-10-06 13:39:02 +02:00
31d94b19a1
Local Gateway: Possible failure to allocate shards to nodes when more than one index exists in the cluster (on full cluster restart), closes #409 .
2010-10-06 00:35:10 +02:00
1f49eb0b9d
Gateway: Add `gateway.expected_nodes` for state recovery, closes #404 .
2010-10-04 16:20:32 +02:00
c737b36df7
better handling of dynamic types in template mappings
2010-10-04 12:37:56 +02:00
425744e0db
add bzip2 compress lib
2010-10-03 16:24:10 +02:00
7180636df6
add close to script services
2010-10-03 14:31:36 +02:00
c9a47a126f
Scripts: Allow to access `_fields` providing access to stored fields, closes #402 .
2010-10-03 12:51:41 +02:00
5b8bc333bd
JavaScript Plugin: Allow to use javascript for scripts, closes #401 .
2010-10-03 02:20:37 +02:00
5a7c8fe2cb
allow script engines to have several types
2010-10-02 21:18:01 +02:00
a127499e3c
doc
2010-10-02 21:07:59 +02:00
31ebed11fd
script handling refactoring, better support for languages that perform better with some caching on a per search aspect (and execution per doc)
2010-10-02 17:32:29 +02:00
9e8ebd46e8
Groovy Plugin: Allow to run use groovy for scripts (where applicable), closes #398 .
2010-10-02 01:22:05 +02:00
aa116e5d40
add lang parameter to all script constructs, allowing for (later) custom script languages
2010-10-01 22:14:20 +02:00
99fcfde307
Mapper: Dynamic Template Support, closes #397 .
2010-10-01 16:56:04 +02:00
f66f0218ca
`/_cluster/nodes/stats` is broken in 0.11, closes #391 .
2010-09-30 08:35:11 +02:00
3b02bd3952
rename package
2010-09-30 08:17:33 +02:00
d3978383a5
when using keyword based analayzer on _all, an NPE is thrown since there is no current entry, ignore it (it does not make sense to have keywork analyzer on _all field...)
2010-09-28 14:53:37 +02:00
55ae6b2001
fix possible problem when trying to convert a binary format to a string
2010-09-28 12:32:40 +02:00
86e17eae2c
fix auto conversion from source content type to rest content type
2010-09-28 00:01:35 +02:00
1f3adc213f
add nodes attributes to cluster state
2010-09-27 10:46:11 +02:00
d9c2cc9d37
allow to sepcify stopwords as comma delimiated list of words
2010-09-26 22:50:25 +02:00
ce28882660
add refresh option to index/create/delete opereation, REST allows for refresh parameter (defaults to false)
2010-09-26 09:07:37 +02:00
2288c5d670
setHighlighterPostTags sets pre tags instead of post, closes #386 .
2010-09-25 00:01:14 +02:00
ab9aa15bb4
add auto sync of translog when using local gateway
2010-09-24 21:25:10 +02:00
1c2733ebee
add flush to gateway
2010-09-24 12:29:52 +02:00
485f904140
Default to not using compound file format for the index (increases number of open files, but faster indexing and recovery), closes #384 .
2010-09-24 11:46:41 +02:00
2f43af0bd7
use directly the directory and not the metaData API
2010-09-24 11:36:13 +02:00
9bb86ea865
imrpvoe boolean filter performance
2010-09-24 00:59:26 +02:00
31fd196687
doc the use of mapbuilder
2010-09-23 20:06:42 +02:00
097071a9d2
add smart name filter wrapping for geo filters to automatically filter by type if as well if provided
2010-09-23 19:34:41 +02:00
832914bf6d
create a mapbuilder when listing files and not ImmutableMap builder since it seems like listFiles might return duplicate files.
2010-09-23 19:13:48 +02:00
761639150d
doc
2010-09-23 19:03:48 +02:00
c96014bb90
default the _river to 1/1 for better OOB experiance (single node recovery with local gateway support)
2010-09-23 02:46:13 +02:00
ecaaeb5250
River CouchDB Plugin, closes #382 .
2010-09-23 02:26:12 +02:00
a59912db02
disable restart API
2010-09-23 00:02:12 +02:00
fd9266dbc6
fix wrong REST response generation of cluster state
2010-09-22 16:03:42 +02:00
91344170ee
doc
2010-09-22 15:39:30 +02:00
f301138e6f
Change default thread pool to scaling from cached, closes #381 .
2010-09-22 15:20:32 +02:00
6d0829f331
imrpove shard target streaming, only do it when needed
2010-09-22 14:36:29 +02:00
cfdf6b0103
add support for op type in also in bulk
2010-09-22 00:07:45 +02:00
3479f2a981
better exception message
2010-09-21 23:49:55 +02:00
2cbcc8dd2e
better handling of search context timeout
2010-09-21 23:17:46 +02:00
b9b91db5c8
building search response should handle cases where the sort is null valued
2010-09-21 22:13:27 +02:00
aa68667d63
rabbitmq river
2010-09-21 17:32:34 +02:00
32acd88b2e
default river index to have 1 shard with 2 replicas by default
2010-09-21 15:07:51 +02:00
8dfac512ba
move parsing of framed bulk binary data into the bulk request
2010-09-21 14:29:19 +02:00
4fe7cbccca
rename indexer to river
2010-09-21 13:35:09 +02:00
69fad80b01
add rest response for allocations
2010-09-21 13:05:10 +02:00
a1299a3d4f
refactoring in routing logic to allow adding allocation explanation that later can be shown
2010-09-21 11:39:22 +02:00
b5b36215ae
refactoring in routing logic to allow adding allocation explanation that later can be shown
2010-09-21 11:37:36 +02:00
dbb2ae69b9
upgrade to latest mvel (2.0.18)
2010-09-21 01:57:41 +02:00
a68c00cd0d
add special handling for primitive arrays as fields in xcontent builder
2010-09-21 01:41:35 +02:00
121e548d76
add to nodes stats the total store size of the indices shards allocated to that node
2010-09-21 01:11:43 +02:00
94a77b69d6
add a twitter indexer
2010-09-20 18:17:02 +02:00
eb4f4f99b3
add recover_after gateway tests
2010-09-20 09:44:57 +02:00
3ec95f4e84
javadoc
2010-09-20 08:58:33 +02:00
770bac252a
fix local discovery to handle master nodes specifically better
2010-09-19 23:49:15 +02:00
25246902cc
Gateway: Add `recover_after_data_nodes` and `recover_after_master_nodes`, closes #376 .
2010-09-19 22:46:13 +02:00
95bb02b330
simple specific master nodes test
2010-09-19 21:11:53 +02:00
8def6f59b3
simple specific master nodes test
2010-09-19 18:48:48 +02:00
20b6688b0a
make ping tests run faster using cluster state api
2010-09-19 18:13:03 +02:00
589dbce89b
move persistenting state and sharsd state in local gateway to another thread
2010-09-19 17:54:14 +02:00
f82ceb1e1e
do snapshots on more lower level interval
2010-09-18 22:18:21 +02:00
e3948115f4
doc
2010-09-18 22:00:18 +02:00
2372f481aa
Shared Storage Gateway (fs/s3): Wrong snapshotting of transaction log, closes #375 .
2010-09-18 21:33:38 +02:00
a2011e0151
no need to cache match all filter
2010-09-17 16:47:35 +02:00
cd7b84e808
improve match_all query to use a filter when possible
2010-09-17 16:38:27 +02:00
3c95d6a215
more work on inexers
2010-09-17 16:36:12 +02:00
ff081240eb
imrpove caching and fix delete mapping logic
2010-09-17 01:25:37 +02:00
08d7125cd5
more work on indexer
2010-09-17 00:22:42 +02:00
4b87f63bed
Terms Facet: Allow to provide `_index` as the field, resulting in facets on indices, closes #374 .
2010-09-16 15:11:30 +02:00
f49f3e169a
Mapping API: Delete Mapping (with data), closes #373 .
2010-09-16 14:35:07 +02:00
aef4704fed
more indexer work
2010-09-16 13:22:09 +02:00
0a6600818a
add count at the end of the stress test
2010-09-15 23:51:25 +02:00
09cdc1c379
don't wait for indices if there are none when using local gateway
2010-09-15 18:22:13 +02:00
e52daa9670
automatic index creation when using the bulk api
2010-09-15 18:18:48 +02:00
e93eb16deb
docs and wrap the rest bulk request in an array
2010-09-15 17:59:10 +02:00
5a5a892cc7
return the correct content type
2010-09-15 12:55:36 +02:00
8db2a217c0
Expose internal search source builder
2010-09-15 12:43:21 +02:00
3afe4da550
Bulk API, closes #371 .
2010-09-15 12:22:07 +02:00
f6fa6ea44e
more work on bulk, client API
2010-09-14 22:08:43 +02:00
d96ffe9153
internal bulk operaiton on the index shard
2010-09-14 01:46:44 +02:00
85160ae341
Mapper: Upgrading a simple mapping to multi_field mapping fails on merge conflicts, closes #369 .
2010-09-14 00:10:01 +02:00
686b59c33f
more work on indexer
2010-09-13 23:00:34 +02:00
0c8dfd967c
better raw generation of json
2010-09-13 02:43:39 +02:00
2d5883ccb5
improve normalization performance from one content type to the other
2010-09-13 02:09:15 +02:00
38aeba438d
XContent refactoring, closes #367 .
2010-09-12 23:20:15 +02:00
4822517d5d
add query parser cache to cache common query strings (+settings) with the parsed queries
2010-09-12 17:08:26 +02:00
77a7938891
move to use guice to register also the default query parsers and filters
2010-09-12 16:46:57 +02:00
2442e1fb16
improve query string parsing by reusing the same parser (bounded to the query parser context)
2010-09-12 15:39:07 +02:00
ed5ffd6591
Setting `path.work` does not affect the logs location, closes #366 .
2010-09-11 22:22:49 +03:00
bc42352f7e
change the local gateway metadata file from state- to metadata-
2010-09-11 21:45:45 +03:00
0718757612
Query DSL: `match_all` filter, closes #365 .
2010-09-11 16:01:30 +03:00
eccc7d5ef2
Search API: Allow to name filters and return per hit the filters it matched on, closes #364 .
2010-09-11 12:38:19 +03:00
ab2a655a59
parsing a query should result in ParsedQuery object, so we can piggyback on it more parsing state / data
2010-09-11 00:34:22 +03:00
a47638dd3c
imrove filter caching handling by not holding an implicit ref to the parent cache
2010-09-11 00:11:47 +03:00
de05a10ae5
use fastGet in docset
2010-09-10 16:16:45 +03:00
4e661c165a
Geo search across anti-meridian doesn't work, closes #363 .
2010-09-10 16:04:10 +03:00
b31b0e979c
Search failure when no results are found and sorting by score, closes #362 .
2010-09-09 15:47:55 +03:00
31362e4c60
upgrade to joda 1.6
2010-09-09 00:53:20 +03:00
a768016779
Allow to configure a common logger prefix using `es.logger.prefix` system prop, closes #360 .
2010-09-08 12:16:41 +03:00
e151b83372
add simple dummy indexer
2010-09-07 18:19:02 +03:00
ad5945f141
Mapper: `geo_point` to support passing array of [lat, lon], closes #361 .
2010-09-07 17:23:06 +03:00
c045b4d0ef
Mapper: `geo_point` type to allow to set `store` on the mapping, closes #359 .
2010-09-07 14:10:29 +03:00
7c48f27e60
throw an exception when no lat/lon or geohash are found
2010-09-07 11:36:12 +03:00
2cb45d4fe7
hack to work around complete text highlighting
2010-09-07 01:28:30 +03:00
bbb2033d3b
start working on indexers
2010-09-07 00:41:30 +03:00
f270fc00d2
earch API: Automatically identify "script" fields on the field elements in search, close #357 .
2010-09-06 11:51:31 +03:00
243b7455e8
fix javadoc
2010-09-06 01:10:46 +03:00
cc1eac147a
Allow to specify highlighter parameters on a per field level basis, closes #356 .
2010-09-05 20:47:38 +03:00
8efe5dc9ae
Highlighter enhancements.
2010-09-05 14:52:48 +03:00
c050bb58da
allow to set a query using a string as well
2010-09-04 21:30:02 +03:00
42b2e60af2
Create Index / Update Settings: Automatically prefix all settings with `index.`, closes #355 .
2010-09-04 14:04:51 +03:00
4f08801bb9
Thrift Plugin: Expose REST interface using thrift, closes #354 .
2010-09-04 02:45:37 +03:00
3d30b9c034
fix not called processModule on plugin for index and shard level modules
2010-09-03 19:06:11 +03:00
1079941397
javadoc
2010-09-03 18:40:30 +03:00
b40f0bc5a4
Local Gateway: Expose `gateway.local.initial_shards` to control when to recover and index, closes #352 .
2010-09-03 18:38:24 +03:00
575f5b406b
abstract away script execution
2010-09-03 01:49:06 +03:00
cd28afe950
Search API: Allow for script fields to extract parts of the stored _`source`, closes #349 .
2010-09-02 21:29:41 +03:00
1e75638b31
refactor field funtion to use internal docmap to be used in scripts
2010-09-02 17:19:50 +03:00
807c485a38
make the cached compile scripts a soft map
2010-09-02 16:29:41 +03:00
7bd08d638d
JMX: Disable creating by default an RMI JMX connector, closes #348 .
2010-09-02 14:17:53 +03:00
0250896a09
reduce shard search response size by not serilaizing the node id, index, and shard id
2010-09-02 11:19:25 +03:00
042af200e8
Gateway: Default to `local` gateway (replace `none` gateway), closes #347 .
2010-09-01 23:21:26 +03:00
0de24218ff
improve local gateway, store started shards and metadata in separated files, metadata stored on master nodes, and started shards on data nodes.
2010-09-01 21:32:07 +03:00
7709cd1bc2
improve GC logging
2010-09-01 15:02:56 +03:00
7d6aa6c027
local gateway: don't store the local state when not a data/master node
2010-09-01 00:07:10 +03:00
ba5cd19989
gateway test to make sure it recovers from the latest version
2010-08-31 21:13:42 +03:00
a3efa21d71
Mapper: Failure to handle null valued objects ("obj1" : null), closes #341 .
2010-08-31 13:33:21 +03:00
f3292f6778
upgrade netty version to 3.2.2
2010-08-30 23:57:37 +03:00
289cfdf25a
improve replica allocation, fix test to wait for recovery on node2
2010-08-30 23:37:09 +03:00
9503fca2ae
fix logging
2010-08-30 22:53:47 +03:00
61764c5b69
add another test for local gateway, fix race when loading the current state of a node
2010-08-30 19:29:00 +03:00
908fba44e7
fsync translog when closing and not deleting, allocate failed shard to another started shard
2010-08-30 16:55:57 +03:00
d9979f8dfe
more work on local gateway, start integration test it
2010-08-30 01:25:45 +03:00
8ed54c24bd
change nodeFile to nodeLocation
2010-08-29 23:25:23 +03:00
5b5d42445d
move the shared fs gateway base class to fs package
2010-08-29 23:22:55 +03:00
269616f35e
more work on getting local gateway to work, apply transaction log operations when primary "recovering"
2010-08-29 16:56:04 +03:00
4f4471483d
initial work on local gateway
2010-08-29 01:24:23 +03:00
163b7be639
refactor gateway service to allow for more custom gateway implemenations
2010-08-28 00:34:15 +03:00
a10ab7fe08
better exception thrown when failing to handle responses
2010-08-27 14:14:28 +03:00
e85c47db74
improve stack trace when failing to send a transaction request
2010-08-27 13:52:58 +03:00
e5b041c8ef
simplify gateway snapshot operation
2010-08-27 00:41:55 +03:00
19ddee5ec9
add a test for time interval
2010-08-26 18:12:51 +03:00
a49af6072b
add objects option to range filter builder
2010-08-26 16:39:39 +03:00
d56b4d266e
move query hint only to broadcast requests that use it
2010-08-26 14:31:54 +03:00
dfa24f6d03
improve caching of stored metadata fetched from nodes during allocation in order to reuse existing shards
2010-08-25 23:12:07 +03:00
2910b6ab7f
add callback for shards_started and shards_failed to better handle caching on node allocations level
2010-08-25 22:20:27 +03:00
3ebd03f69a
doc
2010-08-25 22:03:11 +03:00
9d615a4f0b
Query DSL: Improve `or` and `and` filters to accept just array of filters, closes #336 .
2010-08-25 18:06:28 +03:00
c7a36ed431
add an exception when on filtesr are provided to "and" and "not" filters
2010-08-25 17:56:50 +03:00
fcdc5f872a
improve (perf) getting last committed version of the index to create a new translog with
2010-08-25 15:12:26 +03:00
9d08e56d8a
cache nodes store metadata for on a shard basis when iteratring over the unassigned nodes
2010-08-25 15:03:25 +03:00
f36d80f66d
don't go and find local storage for a shard on nodes if its not going to be allocated
2010-08-25 15:01:09 +03:00
4bfd0a8c26
cache commit points during smart allocation based on local storage when recovering from gateway
2010-08-25 01:12:16 +03:00
244cd42298
allow to use just hosts when using unicast discover, will use the transport level ports
2010-08-25 00:44:04 +03:00
fc812a306b
change logging
2010-08-24 23:33:03 +03:00
ba5814a72f
change log level
2010-08-24 23:25:20 +03:00
31aed01220
increase timeout to 5 seconds
2010-08-24 23:18:58 +03:00
fdc82dd5a3
ignore parsing of wrong file format in the gateway (log warn), use index vesrion as the translog id
2010-08-24 23:12:18 +03:00
d0f5bc3403
add russian letter tokenizer
2010-08-24 23:11:22 +03:00
49b4659322
Indexing Buffer Size: Refine default setting to 10% (from 40%) and add more settings, closes #334 .
2010-08-24 18:12:45 +03:00
7ae8d4c669
thread pool type setting is wrong, fix blocking max setting to have a minimum of 10, use cached TP where needed
2010-08-24 15:28:54 +03:00
b8ab50828c
only list and delete files we created
2010-08-24 14:48:49 +03:00
df23d8f7ec
don't snapshot a translog that has no content
2010-08-24 03:58:21 +03:00
6c29142b91
throw recovery failure if failing to parse all commit points
2010-08-24 01:28:39 +03:00
b52d854711
smarter handling of throttling when preferring allocating to an existing work location
2010-08-23 21:39:54 +03:00
78b6879ecd
move rebalance only when active logic into node allocation, add canRebalance hook point
2010-08-23 18:28:35 +03:00
149962691e
move rebalance only when active logic into node allocation, add canRebalance hook point
2010-08-23 18:27:01 +03:00
4300a6ca18
move rebalance only when active logic into node allocation, add canRebalance hook point
2010-08-23 18:26:17 +03:00
0e47898902
add /_shutdown to the REST endpoints
2010-08-23 18:01:47 +03:00
5900e01a0d
list remote nodest storage information with a timeout
2010-08-23 18:00:37 +03:00
1461da5b49
move prefetch allocation to be a NodeAllocation
2010-08-23 17:22:44 +03:00
edbd586ee5
add load test for recovery from gateway
2010-08-23 16:33:33 +03:00
939d583550
initial chunk of work on refactoring the gateway, syntatic gateway files names, commit points that include translog information
2010-08-23 15:42:03 +03:00
b49d331c86
initial chunk of work on refactoring the gateway, syntatic gateway files names, commit points that include translog information
2010-08-23 15:28:12 +03:00
62cc4d554a
better logging message for reused vs. recovered from gateway
2010-08-23 15:25:06 +03:00
5bd37f6f47
initial chunk of work on refactoring the gateway, syntatic gateway files names, commit points that include translog information
2010-08-23 14:48:49 +03:00
1517fa3d28
initial chunk of work on refactoring the gateway, syntatic gateway files names, commit points that include translog information
2010-08-23 01:45:03 +03:00
36ad3b246a
addlow to add node allocation ot the module
2010-08-22 12:35:59 +03:00
664f28492d
improvements to the module system, allow for a module to spawn other modules, and pre process modules before injector creation
2010-08-22 10:42:25 +03:00
aa28b93610
refactor how throttling is done, instead of doing it after a shard is allocated to a node, and then wait till its allowed to recover, do it on the allocation level, and don't allocate a shard to a node that has N number of recoveries going on it
2010-08-22 02:47:34 +03:00
7592862646
add a callback for allocation as well
2010-08-22 01:00:37 +03:00
a3852766ff
only check for prefered allocation on data nodes
2010-08-22 00:32:43 +03:00
3f701365b0
move the allocate replica after primary active logic to a node allocation
2010-08-22 00:14:56 +03:00
5ded04c5cb
refactor shard allocation to be more pluggable for node allocation "deciders"
2010-08-21 21:59:31 +03:00
2865ceef85
more on sorted nodes when allocating shards above average number of shards per node
2010-08-21 19:55:38 +03:00
76e1a6b1bf
check also for negative values
2010-08-21 19:51:14 +03:00
5d2c005622
pass to canAllocate (though nothing is done there now) the actual routing nodes
2010-08-21 17:52:24 +03:00
1ddd969fa7
pass to canAllocate (though nothing is done there now) the actual routing nodes
2010-08-21 17:52:02 +03:00
11dbbebb7a
renamve var to better reflect its state (started, not active)
2010-08-21 17:50:08 +03:00
3117341f44
Shards Allocation: Only rebalance a shard if all its instances are already active, closes #331 .
2010-08-21 17:49:19 +03:00
38e6649a7e
Cluster Health API: Add `initializing_shards` and `unassigned_shards` to the response, closes #330 .
2010-08-21 16:03:29 +03:00
113ea1bb1b
don't throttle gateway recovery based on peer recoveries (gateway recoveries are important to do as fast as possible). still throttle peer recoveries based on both.
2010-08-21 15:37:12 +03:00
e12bdd9faf
rename RETRY stage in recovery to THROTTLE
2010-08-21 15:13:34 +03:00
0c72bb2125
better writing of local md5 caches
2010-08-21 02:16:36 +03:00
1ef1e9b25b
don't spawn on received cluster state
2010-08-20 19:36:39 +03:00
800e450e3a
handle local md5 cases, don't include them in the list of files the directory has
2010-08-20 19:30:26 +03:00
9acae60dbf
also have a minimum value for concurrent recoveries, at 3
2010-08-20 17:28:31 +03:00
f62f7b8ffe
check the progress in case the index size has not been set yet...
2010-08-20 17:17:32 +03:00
9f0d5f8ef1
better default concurrent recoveries, based on number of processors, tap it at 10
2010-08-20 17:06:56 +03:00
8c25be6dee
better exception message with primary not started exception on index/delete/delete_by_query apis
2010-08-20 16:36:25 +03:00
07285d81d1
add stream done when there is no file (should not happen)
2010-08-20 15:59:46 +03:00
b609162be3
don't ignore recovery on throttling unless the shard is closed
2010-08-20 15:58:24 +03:00
4e74001bde
move marking recovering before creating a status
2010-08-20 15:24:49 +03:00
kimchy
Adriano Ferreira
Lukáš Vlček