Commit Graph

1988 Commits

Author SHA1 Message Date
kimchy 7decb701d9 nicer bench of terms facet 2011-04-07 14:16:21 +03:00
kimchy 7231e08f7c also test terms facet with execution hint of map 2011-04-07 13:32:35 +03:00
kimchy c0879c2055 add bench on nano time 2011-04-06 22:50:36 +03:00
kimchy 08f594c3a0 Get Mapping API: when asking for a mapping of a single index and single type, don't wrap the mapping with the index name, closes #838. 2011-04-06 22:12:34 +03:00
kimchy 8c50a65699 Percolator doesn't work correctly after index recreation, closes #837. 2011-04-06 21:00:49 +03:00
kimchy 6317483fe3 add total count to terms stats and compute mean properly, also better use caching 2011-04-06 15:50:09 +03:00
kimchy 35be46df71 Bug: Sort on a column of type 'short' throws an exception, closes #835. 2011-04-06 02:25:56 +03:00
kimchy 4721f9aa69 fix cache recycling logic to make sure we don't override our own data in facets 2011-04-05 23:25:39 +03:00
kimchy f0de086479 for some reason the copyright header got reformatted 2011-04-05 15:42:12 +03:00
kimchy 6f827731d6 normalize to/from properly in bounded histogram 2011-04-05 15:10:55 +03:00
kimchy 1428abf2b7 Upgrade to jackson 1.7.5 2011-04-05 14:15:48 +03:00
kimchy 9b8eceb229 Histogram Facet: Add ability to define bounds (from/to) to both improve performance and provide additional bound filtering, closes #834. 2011-04-05 13:53:19 +03:00
kimchy 14def814fa fix typo 2011-04-05 12:07:46 +03:00
kimchy d845a306fa fill sentinal proeprly when its not in cache 2011-04-05 03:09:33 +03:00
kimchy a89aeb3a6d add JVM mem to node info API 2011-04-05 01:46:09 +03:00
kimchy 83b728a384 rename some histogram collectors 2011-04-05 01:32:36 +03:00
kimchy cea8c5fefa Geo Distance Facet: Fix wrong total computation with multi valued fields by introducing total_count, add min/max stats, closes #833. 2011-04-05 00:52:58 +03:00
kimchy 79939222e4 Range Facet: Fix wrong total computation with multi valued fields by introducing total_count, add min/max stats, closes #832. 2011-04-05 00:42:46 +03:00
kimchy 90a339ad5e Date Histogram Facet: Improve value field case performance, fix wrong total computation with multi valued fields by introducing total_count, add min/max stats, closes #831. 2011-04-05 00:17:07 +03:00
kimchy 8d1e9dbd3c Histogram Facet: Add min/max stats when providing value field / script, closes #830. 2011-04-04 23:48:54 +03:00
kimchy 46088b9f8a Histogram Facet: Improve value field case performance, fix wrong total computation with multi valued fields by introducing total_count, closes #829. 2011-04-04 21:30:46 +03:00
kimchy 58c606d6d9 break field data value iterator into two callback types, one with support for missing, and one without 2011-04-04 19:53:09 +03:00
kimchy c9838d5c75 use cache recycler for count histo and data histo facets 2011-04-04 19:44:46 +03:00
kimchy 105d60ac9c Geo Distance / Range Facets might count documents several times for a range entry if the field is multi valued, closes #824. 2011-04-04 17:44:38 +03:00
kimchy 5d6e84f206 Upgrade Lucene 3.1: Analyzers, closes #826. 2011-04-04 14:36:13 +03:00
kimchy 4e4495ff1d Upgrade to Lucene 3.1, closes #825. 2011-04-04 12:54:15 +03:00
kimchy db1dcaded3 Concurrent calls to refresh might result in "dangling" searchers, closes #823. 2011-04-01 22:43:48 +03:00
kimchy 1280512097 also support regex pattern matching with ordinals 2011-04-01 16:54:37 +03:00
kimchy e4cbdfa05b Terms Facet: Performance improvements, closes #822. 2011-04-01 16:30:45 +03:00
kimchy fb8a389380 more bench work 2011-03-31 18:25:23 +02:00
kimchy 44d4474f95 anoter small bench 2011-03-31 17:48:57 +02:00
kimchy a0cddf0745 small improvement to stats facet computing min/max 2011-03-31 17:41:21 +02:00
kimchy 3ff35d42b5 Indexing Buffer: Automatically inactivate unindexed into shards and lower their indexing buffer size, closes #821. 2011-03-31 17:03:37 +02:00
kimchy 09d3b882f0 more small bench 2011-03-31 02:30:17 +02:00
kimchy 74838fe1aa Translog: Reduce the number of translog ops to flush after from 20000 to 5000, closes #820. 2011-03-31 01:09:03 +02:00
kimchy 14d98a7319 Shard Allocation: Add node_initial_primaries_recoveries setting, closes #819. 2011-03-31 01:04:09 +02:00
kimchy f90b4e6fee some more micro bench support 2011-03-30 23:32:33 +02:00
kimchy ae0eed937b update to the new lzf compression (0.7) 2011-03-30 20:43:20 +02:00
kimchy 2dd5094d37 Shard Allocation: Rename cluster.routing.allocation.concurrent_recoveries to cluster.routing.allocation.node_concurrent_recoveries (old one still works), closes #817. 2011-03-30 15:43:56 +02:00
kimchy 7ce7fb33e5 Shard Allocation: Allow to control how many cluster wide concurrent rebalance (relocation) are allowed, default to 3, closes #816. 2011-03-30 15:40:11 +02:00
kimchy 8a5dd90885 HTTP: Support compression (gzip, deflate) when using Accept-Encoding header, closes #815. 2011-03-30 14:46:39 +02:00
kimchy eafe686013 warm up with compression bench 2011-03-30 13:52:52 +02:00
kimchy fdbccf28b0 Shard Allocation: Add a setting to control when rebalancing will happen based on the cluster wide active shards state, closes #814. 2011-03-30 12:56:53 +02:00
kimchy 9bb0bcf4e3 add logging when trying to obtain node lock 2011-03-29 21:34:15 +02:00
kimchy 0069d5a75d add adler to checksum, faster and is good enough with length check for our use case 2011-03-29 21:27:18 +02:00
kimchy a1be2bbf20 add adler to checksum tests 2011-03-29 18:17:09 +02:00
kimchy 3138269573 Indices Status API: Add refresh stats, closes #811. 2011-03-29 17:54:00 +02:00
kimchy 2a032f8718 add remote many indices creation test 2011-03-29 17:17:54 +02:00
kimchy 0e6bbf3f85 call shard refresh API and not directly the engine one 2011-03-29 17:17:39 +02:00
kimchy ff8dc2673f Get Settings API: Allow to retrieve (just) a specific index / indices settings, closes #810. 2011-03-29 13:50:25 +02:00
kimchy 508d1d40fb Indices Status API: Remove settings/aliases section, and add `recovery`/`snapshot` flags, closes #809. 2011-03-29 13:05:36 +02:00
kimchy 95e36a073a move index/delete refresh to happen outside of the engine 2011-03-29 01:28:37 +02:00
kimchy 53935f078a optimize array list creation with exact size 2011-03-28 21:50:39 +02:00
kimchy 49f79f54d2 log when search failure happen when scanning 2011-03-28 18:17:36 +02:00
kimchy f00664621f more scan tests to try and recreate the context missing failure, no luck yet... 2011-03-28 17:21:59 +02:00
kimchy e0d8094f3d clear listeners only if not null (can cause failure to properly close a shard), also, do the CLOSE check if to reschedule within the sync block 2011-03-28 17:20:07 +02:00
kimchy 953a99c75c fix a bug in new checksum mechanism that caused for replicas not to retain the _checksums file. Also, now that checksums are widely used, consider files without checksums as ones that need to be recovered 2011-03-28 16:08:01 +02:00
kimchy 64a89eb0e4 add the ability to control the number of shards and replicas in the restart stress test, also add a step to verify version numbers 2011-03-28 14:21:49 +02:00
kimchy af419cc636 upgrade to guice 3.0 2011-03-27 19:31:51 +02:00
kimchy b28461dd0e compress cluster state when publishing it 2011-03-24 01:12:36 +02:00
kimchy 992becf013 add value_script as an option to terms stats 2011-03-24 01:03:17 +02:00
kimchy f20c6ec9cf check if index service is null, and don't load percolators in this case 2011-03-24 00:40:21 +02:00
kimchy 36edcef640 compress metadata stored in gateway 2011-03-24 00:38:16 +02:00
kimchy 4f27fa3f45 add the version to the jvm info, and, warn if running using the client vm 2011-03-23 18:06:29 +02:00
kimchy 20593fb966 Mapping: Add _size field mapping, indexing the original source size, closes #804. 2011-03-23 17:37:40 +02:00
kimchy b2d7cd78b7 don't clean up indices if there is a block disabling persistence 2011-03-22 18:41:29 +02:00
kimchy 8c9000c54c Extend stats API to include filter's cache count and eviction count, closes #802. 2011-03-22 13:30:14 +02:00
kimchy e759b4c971 Add option to expire filter's cache based on access time, closes #803. 2011-03-22 13:14:31 +02:00
kimchy 0d150e6918 Date Field Mapper: Allow to use timestamp value (milliseconds since epoch, UTC), closes #801. 2011-03-22 01:29:41 +02:00
kimchy 9f44c93a25 make sure to log failures in search when reducing results 2011-03-22 01:05:54 +02:00
kimchy 61034f5c86 Mapping: Failure to update _source mapping - compress and compress_threshold, closes #800. 2011-03-21 18:10:13 +02:00
kimchy 8f8fb89c4e Update Settings: Allow to control `index.auto_expand_replicas`, closes #799. 2011-03-21 16:18:03 +02:00
kimchy 2f1aba962c add the exception as well to the logging 2011-03-21 14:36:56 +02:00
kimchy 6c241f3de8 protected against a wrong format in the index.auto_expand_replicas setting 2011-03-21 14:36:28 +02:00
kimchy d2e61af9ee maintain total hits across scan scroll requests 2011-03-20 00:19:26 +02:00
kimchy 52c750fc42 internal support for custom attributes in scroll id 2011-03-19 21:51:18 +02:00
kimchy f3160b71bb can't check on token value, since it might be valid, so, handle the null token case better 2011-03-19 21:34:31 +02:00
kimchy c19a93da89 countDown the latch if no need to delete percolator, introduced when added delete to percolated queries when deleting an index 2011-03-19 21:28:17 +02:00
kimchy fb7fbc8c83 NPE when the JSON to index doesn't match the mapping's expectations, closes #795. 2011-03-19 20:42:50 +02:00
kimchy 70fc8d9af0 Percolator: When deleting an index, make sure to delete all its percolated queries from the _percolator index, closes #793. 2011-03-19 00:24:20 +02:00
kimchy bd3f490d60 Delete IndexAPI: Allow to delete more than one index or _all indices, closes #791. 2011-03-18 20:44:59 +02:00
kimchy 09fbc919b8 "exclude" functionality missing for terms facet on numeric fields, closes #789. 2011-03-18 11:08:11 +02:00
kimchy 923fcf239c Percolator failure when deleting and creating an index, also causes problem with percolated queries isolation between different indices, closes #790. 2011-03-18 10:49:28 +02:00
kimchy 616b3dcb18 shard the default analyzer (standard) on the node level 2011-03-18 01:35:29 +02:00
kimchy 25124b084b Query: Provide an option to analyze wildcard/prefix in query_string / field queries, closes #787. 2011-03-17 22:01:22 +02:00
kimchy 96d06d6dc8 Cluster Health API: Asking for health on an index that does not exists should return RED status, closes #781. 2011-03-15 21:22:46 +02:00
kimchy 0291ab1d8d sync version when getting the state from master 2011-03-15 14:15:16 +02:00
kimchy 1bcd3b67ee Search / Get: Allow to specify a preference on which shards (or order) it will be executed, closes #769. 2011-03-14 11:30:01 +02:00
kimchy 3192654e80 add iterator based on specific index 2011-03-13 19:51:21 +02:00
kimchy 96dfdcf97c Bulk Request: Streamline parameters names to be the same as the single REST request, closes #778. 2011-03-13 11:43:30 +02:00
kimchy 700a2a9577 Sort: Support "missing" specific handling, include _last, _first, and custom value (for numeric values), closes #772. 2011-03-13 02:29:05 +02:00
kimchy 7c04ef6cbc use byte value 2011-03-13 00:51:12 +02:00
kimchy 09006f17c3 Support true|false for the 'store' parameter in mappings, closes #775. 2011-03-13 00:41:58 +02:00
kimchy 6d108e77f1 handle whitespace text on search fields query string parameter to be treated as no fields 2011-03-13 00:37:25 +02:00
kimchy aff8a4407f better exception message 2011-03-13 00:31:21 +02:00
kimchy bdb0fd23e9 Versioning: Add a new version_type parameter, with a new type - `external`, closes #776. 2011-03-12 20:41:55 +02:00
kimchy 7bab5270fc add a comment on max shard buffer size 2011-03-12 17:17:02 +02:00
kimchy 73dec5d63f Add `max_shard_index_buffer_size` to control the upper limit of a shard indexing buffer size (defaults to `512mb`), closes #773. 2011-03-12 17:03:04 +02:00
kimchy 19052a3538 Add another flag to gateway if it even requires snapshot, so we don't try and snapshot on shutdown for none/local gateway (even though its a no op) 2011-03-11 15:20:16 +02:00
kimchy 3e624bf9e0 better failure when trying to obtain a native lock on the node data location, and reduce the tries from 100 to 50. 2011-03-11 15:05:22 +02:00
kimchy e1c86b86e5 add simplified ways to create XContentBuilder that will output to OutputStream using XContentFactory 2011-03-10 16:00:19 +02:00
kimchy 0df3dfd72b add the ability to set source and extra source on the SearchRequestBuilder 2011-03-09 21:55:46 +02:00
kimchy 80a797fc4f Update Settings: Allow to dynamically update `index.translog` settings, closes #765. 2011-03-09 14:17:34 +02:00
kimchy c2a0e0b767 Mapper: Using `dynamic_template` can result in warning of parsed and original source difference (resulting in excessive mapping parsing), closes #764. 2011-03-09 11:36:51 +02:00
kimchy f694b9dfd8 Percolator: Failures when using docs where fields repeat, such as arrays, closes #763. 2011-03-09 11:31:05 +02:00
kimchy ad0d681b6d Update Settings: Allow to dynamically change `index.term_index_interval` and `index.term_index_divisor, closes #762. 2011-03-09 07:36:18 +02:00
kimchy 353d2cb21f Mvel Script: add more random options, and optimize random, closes #759. 2011-03-08 22:24:22 +02:00
kimchy 016e5bce04 Update Settings: Allow to dynamically change refresh_interval and merge policy settings, closes #758. 2011-03-08 21:53:33 +02:00
kimchy 77cf3e1924 small refactor in index settings, allowing for future dynamic changes to it 2011-03-08 11:12:40 +02:00
kimchy e709a0bde4 Logging: Add merge level logging, closes #755. 2011-03-08 08:20:48 +02:00
kimchy 9e2143f343 add a warning for types not to include dots (can't really throw a failure because of it since it will break backward) 2011-03-07 19:17:39 +02:00
kimchy 4bdae621f9 Scripts: Allow to register native scripts (Java) for better script execution performance, closes #752. 2011-03-07 13:09:07 +02:00
kimchy 1242cf59f6 upgrade to guice 3 rc3 2011-03-07 06:52:55 +02:00
kimchy cb912de255 ignore engine closed exception as well 2011-03-05 10:46:00 +02:00
kimchy f0825414d3 upgrade to jackson 1.7.4 2011-03-05 10:45:20 +02:00
kimchy bc882182ab change scripts to start the ElasticSearch main class (a wrapper around Bootstrap) just so the process name will look nicely on jps 2011-03-04 20:14:22 +02:00
kimchy c097735196 improve async merge process, don't spawn a thread unless a merge is really needed, and add an optimized "maybeMerge" operation 2011-03-04 20:05:52 +02:00
kimchy 20ed540fe7 don't block node master node startup if recover_after_time is set and nodes settings are not met (we already handle it with the not recovered state) 2011-03-04 05:11:28 +02:00
kimchy 4b92928c77 Internal: Improve checksum process by bulk writing them into a single file instead of checksum file per index file, closes #747. 2011-03-04 04:22:47 +02:00
kimchy b629d36d8b add a merge thread pool that only does async merges (optimize), so it can be controlled by itself 2011-03-04 01:04:59 +02:00
kimchy 06ad13f373 no need to call maybe merge as part of flush, we do it async anyhow 2011-03-04 00:56:34 +02:00
kimchy 607ada75a6 optimize the index status response 2011-03-04 00:55:39 +02:00
kimchy 41ddcdca93 Node Stats API: Change the structure of the response (more structured), closes #746. 2011-03-04 00:26:47 +02:00
kimchy f901429aaa Node Stats API: Change the structure of the response (more structured), closes #746. 2011-03-04 00:20:05 +02:00
kimchy 682ad7e2fc Indices Status / Node Stats: Add (Lucene) index merge stats, closes #745. 2011-03-03 23:53:28 +02:00
kimchy 2909060af8 Query DSL: flt_field does not allow to set min_similarity and prefix_length, closes #744. 2011-03-03 04:17:02 +02:00
kimchy 78c6f951a3 All Field: All field cane fail to analyze input data (rare cases), closes #743. 2011-03-03 02:59:27 +02:00
kimchy 68c1dd771c upgrade to mvel 2.1 beta2 2011-03-02 22:37:36 +02:00
kimchy 4631df9d01 Delete API: Allow to set _parent on it (will simply set the routing value), closes #742. 2011-03-02 22:14:49 +02:00
kimchy 655cbb9440 improve failure when disabling dynamic creating of types, and trying to index into a non existing type 2011-03-02 20:42:43 +02:00
kimchy faefc772a4 Searching while an index is being allocated and no active shards exists within a "shard replication group" can cause search "misses", closes #736. 2011-03-02 05:11:02 +02:00
kimchy bb1668c2b7 count date histo using the wrong counts aggregator to return results when using time based setting interval 2011-03-02 03:13:25 +02:00
kimchy 3361a9365b since we now take all#enabled into account when checking for include_in_all, there is no need to set it on every mapper 2011-03-02 03:01:01 +02:00
kimchy 185cd68045 really fix time format interval for date_histogram now 2011-03-02 00:39:18 +02:00
kimchy 1d240aaff6 prefix inside a dis_max / bool query broken, closes #732. 2011-03-02 00:36:30 +02:00
kimchy ce862da8f4 add tests for boosted wildcard and prefix queries 2011-03-02 00:26:49 +02:00
kimchy 9560451896 optimize query facet to extract filter from constant_score queries automatically 2011-03-02 00:18:26 +02:00
kimchy 60b423b741 Don't "double" cache a facet filter / query facet, closes #734. 2011-03-01 22:14:36 +02:00
kimchy 0e30ad90f5 support 'w' as suffix for time format 2011-03-01 20:16:18 +02:00
kimchy 0183516221 better failure message 2011-03-01 05:34:21 +02:00
kimchy e639ffbc93 Search: Allow to use has_child filter in facets, closes #730. 2011-03-01 04:45:25 +02:00
kimchy 6097365738 eer Recovery: Batch translog based operation recovery based on size and not just number of operations, closes #729. 2011-03-01 00:49:41 +02:00
kimchy 9d81d01c93 only flush when the index shard is in STARTED mode (nothing will break flushing on non started index, just remove a possible warn logging) 2011-03-01 00:26:03 +02:00
kimchy 0662f65fda log warning when failing to list started shards on other nodes when doing local gateway allocation 2011-03-01 00:25:03 +02:00
kimchy 906ec57f20 Ids with # in them will cause search failures, also, fail when # is used in a type name, closes #728. 2011-02-28 22:43:48 +02:00
kimchy 3bac33e69e don't use AfterTest / BeforeTest in any tests 2011-02-28 22:28:55 +02:00
kimchy 24199a3394 add simple unicast disco test on hte node level 2011-02-28 21:43:31 +02:00
kimchy 47a6065805 Date Histogram Facet: NPE if using "1w" interval, closes #727. 2011-02-28 20:31:02 +02:00
kimchy d395c0a844 better failure message when failing to find a type mapping for a document loaded from the index 2011-02-28 20:21:47 +02:00
kimchy 3b967040da Percolate: Failure to percolate with specific query (when creating the _percolate index before the actual index), closes #725. 2011-02-27 01:06:50 +02:00