Permalink
b1d4557 Jan 19, 2017
192 lines (160 sloc) 10 KB
2.4.2-5 - 2017-01-18
* Add support for uuid and timeuuid Cassandra Types #74
2.4.2-4 - 2017-01-16
* Upgrade to Cassandra 3.0.10
* Support CQL TRUNCATE
2.4.2-3 - 2017-01-04
* Fix search and indexing issue when cassandra replcation factor > 1
2.4.2-2 - 2016-12-25
* Fix useless compactions when creating an index.
* Support for ES Object fields w/ enabled settings
* UDT by name support #64
2.4.2-1 - 2016-12-23
* Based on Elasticsearch 2.4.2 + Cassandra 2.2.7
* Add support for Elasticsearch unit tests (2790 unit tests OK).
* Manage indexes and documents types containing dash or dot caracters (works with Kibana without any modification).
* New index settings *drop_on_delete_index* to delete Cassandra keyspace/table when deleting an Elasticsearch index.
* New index settings *synchrnous_refresh* to refresh Elasticsearch index on each document update.
* Improve search performance by removing some lock contention.
* [WARNING]: RPM & Debian packages not ready to use (please install tarballs).
2.1.1-18.1 - 2016-11-02
* Fix #59 delete-by-query plugin does not work
* Fix #57 Static members cause java.lang.ArrayIndexOutOfBoundsException
2.1.1-18 - 2016-10-24
* Add support for multi-threaded nodetool rebuild_index to improve re-indexing performances.
* Add a new random search strategy to distribute search request to a minimum of nodes.
* Add support for multi-fields and geo_shape.
* Add support for ES index update when deleting by partition key (but deleting with a range of clustering keys not yet implemented).
* Fix issue #50 Elastic Index ignores documents created using Cassandra UPDATE
* Fix issue when using distinct RPC and INTERNAL ip addresses for Cassandra.
* Don't close stdout when running in foreground.
* Creates Elasticsearch indices only for keyspaces using the NetworkReplicationStrategy.
2.1.1-17.1 - 2016-09-04
* Temporary revert CASSANDRA-11427 because of CASSANDRA-12351 (CASSANDRA-12143).
* [breaking change] Replace DatacenterReplicationStrategy by NetworkReplicationStrategy to avoid boostraping issues (Replace all DatacenterReplicationStrategy by
NetworkReplicationStrategy before upgrading).
* Add support for completion type.
* Add support for percolator.
* Add support for Spark/Hadoop with the elasticsearch-hadoop spark connector.
* Add a new metafield _node when indexing document and a token_range aggregation allowing to check cassandra consistency of an index.
* Fix a lat_lon order issue.
2.1.1-16 - 2016-08-12
* Upgrade to cassandra 2.2.7
* Provide a JSON-REST access to cassandra rows (get/index/delete without indexing in elasticsearch).
* Fix a bug when using cassandra index (issue #41)
* Implement the delete document in elaticsearch indices (issue #36), and nodetool cleanup effectively remove documents.
* [breaking change] Change the DatacenterReplicationStrategy options to be compatible with the standard apache cassandra 2.2.7.
* Add documentation about backup and restore.
2.1.1-15 - 2016-07-10
* Add a per-index logging level.
* Fix a parent-child mapping issue.
* Update documentation (integration part and numerous typos #32)
2.1.1-14 - 2016-07-01
* Fix an erroneous number_of_replica in a multidatacenter configuration.
* Fix a bug when partially updating a row indexed in an index different from the keyspace name.
* Fix a bug on CQL schema update.
2.1.1-13 - 2016-06-29
* Limit CQL schema update from the coordinator node only.
* Fix a bug in routing table with vnode and one or more unavailable nodes.
2.1.1-12 - 2016-06-21
* Add a configurable cross-datacenters replication.
* Add support for index partitioning (see index.partition_function)
* Optimize search operation with _routing (send the search request to a node hosting the corresponding token, issue #23)
* Avoid useless reads of metadata with consistency quorum for better performances.
* Remove translog for better performances (durability ensured by cassandra commitlogs)
* Disable the TTService (see cassandra TTL and index partitioning)
* Add caching of CQL query on fetch phase.
* Add a native CQL function toJsonArray() to cassandra.
* Change Elasticserach JNA to rely on cassandra JNA class.
* Map the cassandra TTL and WRITETIME to _ttl and _timestamp meta attributes.
* Improve parent-child relationship management with a new mapping parmeter pk_columns.
* Fix bug when SERIAL consistency not met on elassandra restart.
* Fix a bug when searching with no type filter.
* Fix index open/close operation
* Fix a bug during shard initialization and a bug when a keyspace is unavailable.
* [breaking change] Rename the mapping attribute "columns_regexp" by "discover".
* [breaking change] Rename the index setting "keyspace_name" by "keyspace".
* [breaking change] Rename the field mapping parameter "cql_partial_update" by "cql_mandatory".
* [breaking change] Rename the ThreadLocalOptimizedElasticSecondary class by ExtendedElasticSecondaryIndex.
2.1.1-11 - 2016-05-20
* Improve secondary index implementation.
* Add a cluster setting cluster.default_secondary_index_class.
* Change the default memtable_allocation_type: offheap_objects in conf/cassandra.yml
* Change the default bootstrap.mlockall: true in conf/elasticsearch.yml
* Include the cassandra driver to support cassandra user defined functions.
* Merge the JNA management of cassandra and elasticsearch (and disable seccomp).
* Correctly delete cassandra secondary indices when removing an elasticsearch index.
* Fix potential duplicate result when searching without any _type filter.
2.1.1-10 - 2016-05-11
* Fix a bug when setting number_of_replicas on a create index request.
2.1.1-9 - 2016-05-10
* Upgrade to cassandra 2.2.5
* Change the default elasticsearch data directory to ${cassandra_storagedir}/data/elasticsearch.data.
* Provides various cassandra secondary index implementations with better performances by removing json generation and parsing.
* Change the default to index=not_analyzed for text columns.
* Improve cluster metadata management (fix a bug when deleting an index).
* Fix a mapping bug when indexing a cassandra map as a nested document.
* Fix issue #16 with geo_point.
* Fix issue #17 with float fields.
2.1.1-8 - 2016-04-18
* Add mapping attributes cql_partition_key:boolean and cql_primary_key_order:integer to build cassandra table with composite primary key.
* Add support for static column with the mapping attribute cql_static_column:boolean. This give the ability to index static columns in Elasticsearch.
* Fix various mapping issues.
2.1.1-7 - 2016-04-02
* Fix a ClassCastException when indexing a document with single partition key other than a string.
* Add support for mapping update of nested object (update cassandra UDT).
2.1.1-6 - 2016-03-20
* Fix an indexing bug (regression from 2.1.1-5)
* Disable cassandra timestamp update when Elasticsearch _timestamp is disabled (improve insert performance, regression from 2.1.1-4)
* Fix a bug on flush and refresh operations causing a performance issue with kibana.
2.1.1-5 - 2016-03-15
* Add missing classes from org.elasticsearch.common.logging.
* Add index setting number_of_replica in cluster state.
* Change elastic_admin keyspace replication strategy to 'DatacenterReplicationStrategy' (RF=Datacenter number of nodes to improve resilancy).
* Manage token_range in SearchRequest has an optional field to remain compatible with the Elasticsearch binary protocol.
* Set Elasticsearch version to 2.1.1 to remain compatible with some plugins.
2.1.1-4 - 2016-02-21
* Throw a DocumentAlreadyExsistsException when op_type=create failed.
* Fix a bug on cassandra timestamp when indexing or deleting documents.
2.1.1-3 - 2016-02-21
* Add a start script for Linux Debian provided by Chris Feldmeier.
* Add sigar binaries in the tarball package.
* Fix the build_hash in the cluster information.
2.1.1-2 - 2016-02-16
* Elassandra now build as a standalone project.
* Add a modified bin/cassandra.in.sh without empty element in CLASSPATH (cause a JarHell error)
2.1.1-1 - 2016-02-03
* Add support for parent-child relation-ship
* Bug fix when indexing inet type (Internet Address).
2.1.1 - 2016-01-22
* Based on Elasticsearch 2.1.1 and Apache Cassandra 2.2.4
* Manage the _token field as an internal metadata field. This is the murmur3 hash of the partition key for cassandra row (an Elasticsearch document).
* Index settings number_of_shards and number_of_replicas now dynamically reflects the number of nodes in the ES cluster (in the cassandra datacenter) and the replication factor for each index (= keyspace).
0.5 - 2015-12-21
* Based on cassandra 2.2.4
* Improve mapping and cassandra secondary index management :
* to map a keyspace to many Elasticserach indices allowing mapping update with zero-downtime.
* to create cassandra secondary indices when all shards are started, thus indexing existing data on all nodes.
0.4.2 - 2015-11-23
* Improve automatic mapping from an existing CQL3 table.
0.4.1 - 2015-11-17
* Bug fix when indexing inet type (Internet Address).
0.4 - 2015-11-11 (Based on cassandra-2.1.8 + elasticsearch 1.5.1)
* Fix various bugs
* New tarball package.
0.3.2 - 2015-11-05 (Based on cassandra-2.1.8 + elasticsearch 1.5.1)
* Bug fixes
* Fix multiget on a multiple nodes cluster (Allow to save kibana configuration on multi-node cluster).
* Fix a bug in cluster state management when restarting.
* Fix CASSANDRA-9912
* New features
* Add support for compound cassandra keys.
* Add support for partial CQL update of cassandra collection types.
* Enhance Elasticsearch put mapping to automatically build mapping from an existing CQL3 table.
* Replace the single_value mapping attribute by cql_collection, cql_struct and cql_partial_update. This allow to map nested document to a User Defined Types or to a cassandra map<text,?>.
0.2 - 2015-10-07
* Add support for partial document update
* Add support for Kibana
* Fix User Defined Type creation
* Change cassandra StorageService hooks to start Elasticsearch (Fix commitlogs recover)
0.1 - Initial release