Commit Graph

2252 Commits

Author SHA1 Message Date
kimchy 0b59c00e0d Upgrade to Lucene 3.3. closes #1083. 2011-07-01 18:30:06 +03:00
kimchy bac6240d17 improve clearing caches in guice 2011-07-01 13:32:23 +03:00
kimchy a8969cd672 Without pre-creating index, heavy write rates OOM with too many threads, closes #1081. 2011-06-30 21:22:21 +03:00
kimchy 4c913693a9 bite the bullet and add guice source, we don't need much from it, and will allow to optimize for our use case as well as remove the AOP support from it 2011-06-30 20:24:16 +03:00
kimchy e78e665273 Zen Discovery: Add `minimum_master_nodes` setting helping with split brains, closes #1079. 2011-06-30 17:02:29 +03:00
kimchy c31ee7d68d add xcontent builder settings option (though just calling builder#string() was always possible) to create index 2011-06-29 15:12:53 +03:00
kimchy 0bf1ed964e fix wrong routing table versioning when deleting an index 2011-06-29 14:51:32 +03:00
kimchy 4139bb5180 Misbehaving "missing" field in facet results, closes #1070. 2011-06-28 12:38:21 +03:00
kimchy f6ebee3785 allow to provide fields for multi get 2011-06-28 11:54:52 +03:00
kimchy 2e83a2f045 clean unused code 2011-06-28 11:37:52 +03:00
kimchy 73898067b8 format 2011-06-28 00:29:06 +03:00
Nicolas Lalevée 759d7d693d Support for highlighting of phrase prefix queries 2011-06-28 00:26:06 +03:00
kimchy ce6fbc687d upgrade to mvel 2.1 Beta5 2011-06-28 00:25:37 +03:00
kimchy 9bf686ef62 Multi GET API, closes #1068. 2011-06-27 22:23:49 +03:00
kimchy 475564449f properly take offset and length into account (introduced with realtime get) 2011-06-27 21:51:14 +03:00
kimchy f851279db1 delegate to the proper source method 2011-06-27 21:38:43 +03:00
kimchy cb0d89700c get shards should use preference as well when routing (fix wrong previous commit) 2011-06-26 23:07:40 +03:00
kimchy 0bbf71188c make broadcast action more extendable by refactoring out type/id 2011-06-26 16:33:44 +03:00
kimchy b670a7f2a3 Indices Segments API: Internal segments info of shard level Lucene indices, closes #1064. 2011-06-26 00:20:50 +03:00
kimchy 7acdda74f4 Transport Client: Adding more nodes causes more scheduled reconnect tasks, closes #1062. 2011-06-24 21:33:24 +03:00
kimchy e373bf09f2 upgrade to joda 1.6.2 2011-06-24 18:42:53 +03:00
kimchy 6521143a38 upgrade to mvel 2.1 Beta4 2011-06-24 18:27:09 +03:00
kimchy c884adf603 more get tests 2011-06-24 15:52:45 +03:00
kimchy 7ed5e9e79a Get API: Make type optional, closes #1061. 2011-06-24 13:43:37 +03:00
kimchy 89a46d1c87 filter out crappy gc reporting 2011-06-24 13:18:16 +03:00
Lukáš Vlček a9a80958ce commented out unused 2011-06-24 10:51:41 +03:00
Lukáš Vlček 2f9a1c7a6a Analysis: expose Lucene StemmerOverrideFilter. Closes #1053 2011-06-24 10:51:41 +03:00
Lukáš Vlček 5913e74286 Analysis: expose Lucene KeywordMarkerFilter. Closes #1052 2011-06-24 10:51:41 +03:00
Igor Motov bae67ca808 Fix NullPointerException when search request partially fails on one or more shards 2011-06-24 09:41:10 +03:00
kimchy 72ee0aaee7 Realtime GET, closes #1060. 2011-06-24 09:39:37 +03:00
kimchy 4547bc3ef4 upgrade to trove 3.0.0rc2 2011-06-20 00:11:28 +03:00
kimchy f3af1fd1d4 upgrade to jackson 1.8.2 (though with not real changes to us) 2011-06-19 20:58:56 +03:00
kimchy 35fa6d93ad Update Settings: Allow to change non dynamic settings on a closed index, closes #1048. 2011-06-19 17:15:05 +03:00
Igor Motov 8470e79aed Fix non-varargs call of varargs method with inexact argument type for last parameter exception warning 2011-06-19 10:10:35 +03:00
Igor Motov 4794345efb Remove unnecessary Thread.sleep calls after alias creation 2011-06-19 10:10:34 +03:00
Igor Motov ae5b963e97 Specify a unique cluster name for all Integration tests 2011-06-19 10:10:34 +03:00
kimchy 5540fbf7ab Delete API: Using refresh parameter might not refresh delete operation on replica, closes #1047. 2011-06-19 09:15:08 +03:00
kimchy 3e9dff8b7e add kstem also to the static shared token filters 2011-06-17 19:23:08 +03:00
Lukáš Vlček aefc3ab3e5 Closes #1041 2011-06-17 16:51:24 +02:00
Lukáš Vlček 44af2cda41 Adding Czech default stopwords into named stopwords map 2011-06-17 13:22:22 +02:00
kimchy dfed1b947e fix bench 2011-06-17 01:19:12 +03:00
kimchy 2e64edd4e7 Parent Child: complex parent child mapping (>2 hierarchy) might fail, closes #1039. 2011-06-17 01:03:34 +03:00
kimchy f5fbe0d66c River might not start properly after cluster shutdown (timing), closes #1038. 2011-06-16 14:04:50 +03:00
Paul_Loy e1514aa8e0 reverted adding IOException to NetworkService, return null if cannot resolve name and log 2011-06-15 22:03:16 +03:00
Paul_Loy 004f91dfc0 add throws IOException to CustomNameResolver interface 2011-06-15 22:03:15 +03:00
Paul_Loy 6a60bbba1d adding ability for gradle to create eclipse project files and classpaths 2011-06-15 22:03:15 +03:00
Igor Motov 1d8b545854 Wait for alias operations to be propagated to all nodes before returning from the call. 2011-06-15 17:16:18 +03:00
kimchy db6f5a7146 Search: When a search request failed completely (all shards fail) return a proper HTTP status code, closes #1035. 2011-06-15 16:28:53 +03:00
kimchy bbd73d5afa Search: When all shards fail on second / third phase, make sure to return a response with proper shard failures, closes #1033. 2011-06-15 15:28:31 +03:00
kimchy b7d492af4b fix synonym filter problem due to refactoring done in analysis components regestrations 2011-06-15 15:13:09 +03:00
kimchy cc39eb76d0 add a narrowing search scan test 2011-06-15 14:11:15 +03:00
kimchy 64bf849cb1 Analyzers for FLT / MLT, closes #1031. 2011-06-15 13:43:10 +03:00
kimchy 11f19ff6da Query DSL: Span Term Query wrongly parses when boost is provided, causing using it in span or to fail, closes #1032. 2011-06-15 13:11:22 +03:00
kimchy 551b98f2f8 fix error message 2011-06-15 12:32:28 +03:00
kimchy fee1747947 also support direct discovery.zen.ping_timeout setting even when it gets extended (like in the case of ec2 disco) 2011-06-14 20:50:05 +03:00
kimchy 8bb448473d more optimizations for sharing same analyzers across different indices on the same node 2011-06-14 18:51:34 +03:00
kimchy bdc0b88c49 Unicast discovery: Improve initial ping connection to nodes to be more lightweight, closes #1028. 2011-06-13 22:19:42 +03:00
kimchy fc17ba1652 allow to nullify routing values when providing then through REST API 2011-06-13 01:23:14 +03:00
kimchy adbe0565f1 validate when setting aliases that index routing does not resolve into several routing values 2011-06-13 01:21:03 +03:00
kimchy fbdf11d3f1 verify that index routing is not, by mistake, composed of several routing values 2011-06-13 01:00:34 +03:00
kimchy 7c67f11e7c when providing explicit routing in index/delete operation, and working against an alias with routing, fail the index operation 2011-06-13 00:41:32 +03:00
kimchy 00e010a52a on dynamic index creation when indexing, we might raise failure on routing (for example), but then not notify the listener 2011-06-12 23:22:12 +03:00
Igor Motov b35dfd3aa7 Add ability to specify routing information for aliases 2011-06-12 22:51:01 +03:00
kimchy 35b4cabed8 try and batch started shards events (causing rerouting) by draining a shared queue (not time based) 2011-06-12 21:35:22 +03:00
kimchy 110c4d0625 add version on a shard replication group level, optimizing both the initial allocation of large number of shards and persistency in case of local gateway 2011-06-12 21:16:58 +03:00
kimchy ffc7f3e12b make sure to initialize the routing table version with the cluster state version as well 2011-06-12 14:39:44 +03:00
kimchy 0eaccd483f Indices exists API, closes #1022. 2011-06-12 12:09:57 +03:00
kimchy 2dbe890022 Indices exists API, closes #1022. 2011-06-12 12:01:29 +03:00
kimchy 5ee6bbfae9 Numeric Types: By default, omit_norms (boosting) is set to true, automatically set it to false if explicit field mapping boosting is provided, closes #1021. 2011-06-12 11:13:06 +03:00
kimchy d4d80525de Numeric values always omit norms (boosting), regardless of the configuration set, closes #1020. 2011-06-12 11:03:43 +03:00
kimchy fd80fcbca0 add separator to several validation failures 2011-06-12 10:34:42 +03:00
kimchy eca244054f Merging complex objects / types (geo, objects) can cause failure to lookup field names (without type prefix), closes #1016. 2011-06-10 20:00:59 +03:00
kimchy 3e1e58e394 add debug logging to print when sigar is not loaded 2011-06-10 17:59:24 +03:00
kimchy 6382ddf43c some improvements to the many indices stess test 2011-06-10 05:12:44 +03:00
kimchy 03dc146928 optimize translog sync to not happen if there is no need for it 2011-06-10 04:58:41 +03:00
kimchy 3b2c186503 improve shutodwn with many indices / shards by bounding the concurrency of closing indices / shards 2011-06-10 03:48:43 +03:00
Erick Tryzelaar b143400bea Fix some typos. 2011-06-10 03:10:09 +03:00
Erick Tryzelaar 112935fcdb Allow fuzzy IP matches to accept a long min_similarity. 2011-06-10 03:10:09 +03:00
kimchy 4b6e2ddd7d add metadata level version, reducing the persistency requirements fo rit 2011-06-10 01:37:20 +03:00
kimchy f87b9e3656 add a version to routing table, so it can be reused when not changed after publishing 2011-06-09 22:51:40 +03:00
kimchy 1d9942847e fix check for dynamic update of a setting in tiered merge policy 2011-06-09 22:25:51 +03:00
kimchy a07030ccf3 improve local shard allocation to utilize same version shards allocation within the same replication group 2011-06-09 10:51:45 +03:00
kimchy 1a0ee00fbb change logging statement to indicate ping_timeout, not initial_ping_timeout 2011-06-08 19:00:02 +03:00
kimchy fff78d6a38 Transport: Improve concurrency when connecting to several nodes, closes #1007. 2011-06-08 18:47:37 +03:00
kimchy 597f3b7a8f throw explicit failure when trying to change the number of shards on an index 2011-06-08 13:49:22 +03:00
kimchy c9d4ce59ce add mime types and a todo 2011-06-06 11:18:12 +03:00
kimchy 4890e43980 Analysis: Add `reverse` and `skip` to `path_hierarchy` tokenizer, closes #999. 2011-06-05 15:39:19 +03:00
kimchy 9652ad0d08 New default merge policy - `tiered`, closes #998. 2011-06-05 14:23:51 +03:00
kimchy 6788c6c375 Upgrade to Lucene 3.2, closes #997. 2011-06-05 01:57:10 +03:00
kimchy c872be75ae make sure to reroute whenever required on the action itself (most already did, added index state case), and then, no need to do additional reroute (which probably does not do anything, and returns the same cluster state) on the routing update handler 2011-06-04 21:43:21 +03:00
kimchy abfc7f0db4 internal refactoring simplifying the type mapping system 2011-06-04 03:56:01 +03:00
kimchy dcd2bbd9a9 ',' can be used in type name, closes #996. 2011-06-04 01:26:41 +03:00
kimchy 5fbb4f7043 simplify plugin registration of custom queries and filters 2011-06-03 06:14:14 +03:00
kimchy f80d1c8fd9 share common query parsers and filter parsers among different indices 2011-06-03 06:02:18 +03:00
kimchy 0a4e582404 Query Parser Module Refactor, closes #994. 2011-06-03 04:32:30 +03:00
kimchy 58330d9c23 Aliases: Validate alias filter before adding it, closes #993. 2011-06-03 02:23:21 +03:00
kimchy 0f1655cf22 Query DSL: Simplify not filter to accept the filter directly, without a wrapping `filter` element, closes #992. 2011-06-03 01:42:57 +03:00
kimchy ec1f0b16f4 remove pretty from parameter when doing update settings 2011-06-02 22:01:51 +03:00
kimchy ee5bfba52f Get Mapping: Better error response when asking for specific index type, closes #991. 2011-06-02 18:30:06 +03:00
kimchy b72ba7d874 remove dead code 2011-06-02 18:23:30 +03:00