CHANGES.txt

6.8.4.16 - 2022-01-03
 * Upgrade log4j 2.17.1 to fix CVE-2021-44832

6.8.4.15 - 2021-12-21
 * Upgrade log4j 2.17.0 to fix CVE-2021-45105

6.8.4.14 - 2021-12-14
  * Upgrade log4j 2.16.0 to fix CVE-2021-44228

6.8.4.13 - 2021-01-05
  * Fix a possible StackOverflowError
  * Rename the finalize method to avoid GC issue

6.8.4.12 - 2020-12-07
  * Fix a mapping update issue when having a datacenter group
  * Update doc about reindexing from a remote cluster

6.8.4.11 - 2020-11-16
  * Upgrade to cassandra 3.11.9.1
  * Fix nested types matching error
  * Fixup multiple delete of the same document

6.8.4.10 - 2020-08-10
  * Upgrade jackson 2.8.11 to fix a security issue (see elasticsearch #45225)
  * Upgrade to cassandra 3.11.7.1
  * Add support for histogram aggregation over CQL

6.8.4.9 - 2020-07-28
  * Fixup decommissioning a node

6.8.4.8 - 2020-07-26
  * Fix flush, refresh and close distributed actions
  * Fix open index from a node that is not the last master node.

6.8.4.7 - 2020-07-23
  * Fixup node update for removed nodes
  * Fixup routing table update

6.8.4.6 - 2020-06-11
  * Fix #346 Nested objects with inner hits query returns "extracted source isn't an object or an array"
  * Fix NPE when network config is wrong

6.8.4.5 - 2020-05-01
  * Fixup NPE on cluster state update
  * Fix kibana _timestamp field mapper not found
  * Fix open/close index
  * Add a new MessageFormatTimeUUIDPartitionFunction
  * Add support for datacenter tags

6.8.4.4 - 2020-04-13
  * Upgrade to cassandra-3.11.6
  * Fix RandomSearchStrategy
  * Add a new RackAwareSearchStrategy
  * Improve integration tests

6.8.4.3 - 2020-02-29
  * Switch to cassandra 3.11.5.3
  * Improve gradle publication

6.8.4.2 - 2020-02-17
  * Fix TimeUUID & Base64 Ingest processor declaration
  * Compression of X1 application state to increase the number of created indices

6.8.4.1 - 2020-01-23
  * Fix #323 Elassandra Mapping coercing issue with dynamic template
  * Add documentation for TimeUUIDPartitionFunction
  * Improvement for Gossip X1 publication

6.8.4.0 - 2020-01-02
  * Upgrade to Elasticsearch 6.8.4
  * Upgrade to Cassandra 3.11.5.1
  * Fix #315 missing subfield when indexing a doc with 2 or more new subfields
  * Known issue on geoshapes

6.2.3.24 - 2020-01-21
  * Fix #323 Elassandra Mapping coercing issue with dynamic template
  * Add documentation for TimeUUIDPartitionFunction
  * Improvement for Gossip X1 publication

6.2.3.23 - 2019-12-28
  * Fix #315 missing subfield when indexing a doc with 2 or more new subfields
  * Add support for index name longer than 48 characters

6.2.3.22 - 2019-11-28
  * Update Apache Cassandra version from 3.11.4 to 3.11.5
  * Fix DecodeSmile command issue
  * Fix possible gossip update race condition on bootstrap
  * Add automated integration tests based on ECM (CCM for Elassandra)
  * Improve unit testing

6.2.3.21 - 2018-10-16
  * Remove elastic_admin RF adjustment

6.2.3.20 - 2018-10-10
  * Fix possible mapping loss when having multiple indices on the same table
  * Add support for aggregation on opaque_map keys

6.2.3.19 - 2018-09-29
  * Update the cassandra.yaml to follow the default cassandra one.
  * Update documentation
  * Add logback DuplicateMessageFilter in the default logback configuration
  * Add elassandra-cli to decode Elasticsearch mapping from the CQL schema
  * Enable mapping compression in CQL schema
  * Improve mapping update concurrency management
  * Improve upsert retry with mapping update
  * Add logback DuplicateMessageFilter in the default configuration
  * Add support for virtual index
  * Add support for opaque_map with no mapping update
  * Fix index setting update issue #306
  * Fix drop column on a deleted index #302

6.2.3.18 - 2018-07-18
  * Improve routing table update with batched cluster state updates
  * Fix blocking local cluster state update
  * Add support for Elasticsearch range fields #201

6.2.3.17 - 2018-06-26
  * Move some JVM props to jvm.options
  * Fix #298 build_hash and build_date as unknown
  * Fix #295 dots in object mapping ignored
  * Fix #296 null_value on an integer field causes an exception
  * Fix wrong assertion when deleteing index

6.2.3.16 - 2018-05-31
  * Add doc_values support for _routing
  * Add #139 support for Cassandra decimal as Elasticsearch text
  * Fix #292 index.mapping.ignore_malformed not working on Multifield
  * Fix #294 adding template from many nodes

6.2.3.15 - 2018-05-09
  * Fix mapping update while index is building
  * Fix #286 index deletion when having multiple index on the same table
  * Fix number_of_replica in a multi-dc configuration #283
  * Add support cql for attributes for geo_shape (#282)
  * Upgrade to cassandra 3.11.4.1

6.2.3.14 - 2018-04-29
  * Upgrade to cassandra 3.11.4
  * Add support for Elasticsearch copyTo
  * Fix create indexes in case the table has materialized views #274
  * Fix _source including IP with bad format (#279)
  * Fix missing docs when searching on an opaque storage index

6.2.3.13 - 2018-03-25
  * Fix schema update race condition
  * Fix dead node detection issue
  * Improve mapping check when ussing Cassandra maps.
  * #271 Fix assertion error on update and add unit test

6.2.3.12 - 2018-03-11
  * Allow indexing through the ES API with consitency level QUORUM, LOCAL_QUORUM and EACH_QUORUM.
  * Fix issue when migrating 6.2.3.8+ index with removed keyspaces
  * Improve paging and aggregation support in the ElasticQueryHandler
  * Silently ignore null fields with no mapping

6.2.3.11 - 2018-02-07
  * Add support for Elasticsearch query over CQL
  * Ship cassandra-stress jar with elassandra #19 #261
  * Manage elassandra artifacts with a 4 digits version number
  * Do not publish javadoc to avoid issue with java 9
  * Fix log corruption edge case when TOKENS= epState is printed to console #262
  * Fix token mapper compatibility issue for existing indices
  * Fix migration for indices created before 6.2.3.8

6.2.3.10 - 2018-01-10
  * Fix indexing a static UDT in insert only mode
  * Fix partition key ordering issue when building CQL schema from the ES mapping
  * Add indexing support for map<text, UDT>
  * Add a new Base64 processor
  * Optimize _routing field storage

6.2.3.9 - 2018-12-27
  * Fix index rebuild when bootstraping

6.2.3.8 - 2018-12-17
  * Fix cluster settings
  * Fix date field in primary key #222
  * Refactor mapping management to ensure atomic and sequential CQL schema updates.
  * Add a boolean cql_clustering_key_desc mapping attribute
  * Add a index.table_options to properly configure cassandra tables
  * Add support for opaque storage of document _source
  * Improve index_insert_only mode by using cassandra UNSET_BYTE_BUFFER
  * Fix cqlsh help pages result in 404 errors #241
  * Add a Timeuuid processor
  * Add support for custom cql_type in mapping

6.2.3.7 - 2018-10-15
  * Improve license notice
  * Index document with empty mapping
  * Add support for ingest feature
  * Move docker image build to docker-ci.sh based on debian packaging

6.2.3.6 - 2018-10-09
  * Improve index rebuild performance
  * Fix missing hotspot_compiler

6.2.3.5 - 2018-09-24
  * Upgrade to cassandra 3.11.3
  * Improve indexing of updates not time-ordered
  * Add new index setting index_insert_only for insert only workloads.
  * Fix #178 add package name to function class if needed
  * Add a node status DISABLED for cassandra-only nodes

6.2.3.4 - 2018-08-10
  * Fix cassandra discovery for k8s
  * Delete nested documents when truncating a cassandra table
  * Fix concurrent partition read-before-indexing

6.2.3.3 - 2018-07-16
  * #211 Improve ip replacement management
  * Fix race condition when computing the routing table
  * Fix NPE when logging an error
  * Fix bug #199 with node.name

6.2.3.2 - 2018-06-25
  * Fix packaging of elasticsearch-plugin in tarball distribution
  * Fix #202 Memory lock not working

6.2.3.1 - 2018-06-09
  * Upgrade to Elasticsearch 6.2.3 + Cassandra 3.11.2
  * Allow to use the cassandra listen_address as the elasticsearch transport publish address.

5.5.0.18 - 2018-06-03
  * Fix support for range partition deletion and add support for smallint and tinyint
  * Fix #197 Deletion of a List element removes the document on ES
  * Improve catch exception to avoid cassandra write timeout

5.5.0.17 - 2018-05-22
  * Fix list settings representation
  * Make replicated actions executed locally
  * Improve completion type mapping to get closer to elasticsearch behavior

5.5.0.16 - 2018-05-10
  * #187 refactoring of partitionDelete
  * Fix #191 NPE
  * Don't discover es_query and es_options pseudo columns
  * Improves shards starting before playing commitlogs
  * Update the cassandra startup script

5.5.0.15 - 2018-05-01
  * Fix #183 by using node name as the cassandra broadcast address
  * Fix #184 disable translog sync thread
  * Fix #188 race condition when starting elassandra on kubernetes

5.5.0.14 - 2018-03-18
  * Merge #176 Don't overwrite replication setting
  * Remove asm lib from the cassandra classpath to support elasticsearch scripting
  * Improve index mapping initilization in multi-datacenter configuration

5.5.0.13 - 2018-02-28
  * Upgrade to cassandra 3.11.2
  * Set elasticsearch transport address to cassandra RPC address
  * Add a new index setting to set keyspace replication map

5.5.0.12 - 2018-02-12
  * Improve elassandra bootstrap to build elasticsearch indices when rebuilding a datacenter from another one.
  * Improve mapping of _type having an hyphen in their name
  * Fix scripting class loading issue #165

5.5.0.11 - 2018-02-03
  * Fix #163 Invalid date mapping
  * Fix #161 Search over a nested set returns wrong inner_hits
  * Fix duplicate indexing of _all and field_names
  * Back port elasticsearch #26856 to fix IndexOutOfBoundsException in histograms for NaN doubles (#26787)

5.5.0.10 - 2018-01-20
  * Fix #157 Indexing over set<text> not working
  * Fix #153 datacenter.group settings ignored

5.5.0.9 - 2018-01-06
  * Add support for completion suggester
  * Add static_document index settings from v2.4.5
  * Add support for nested empty object  #146
  * Fix elasticsearch index update when deleting a flushed wide row

5.5.0.8 - 2017-12-01
  * Fix #142 nested date object
  * Fix #141 issue with native elasticsearch transport client
  * Enable STDOUT logging by default for docker

5.5.0.7 - 2017-11-20
  * Fix case-sensitive nested object name

5.5.0.6 - 2017-11-13
  * Fix Error while writing unindexed boolean #135
  * Improve README #133
  * Upgrade to Cassandra 3.11.1.2
  * Improve search query routing for spark support

5.5.0.5 - 2017-11-01
  * Upgrade to Cassandra 3.11.1.0 to fix #132
  * Fix index build race condition when creating an index on existing cassandra data.

5.5.0.4 - 2017-10-16
  * Upgrade to Cassandra 3.11.1
  * Fix #128 fetch multiple CQL types
  * Fix dependencies for Cassandra UDF

5.5.0.3 - 2017-10-11
  * Improve performances on GET operations.
  * Fix startup when metadata is empty.
  * Update and refactor documentation.

5.5.0.2 - 2017-10-01
  * Add support for CQL timeuuid, date and time.

5.5.0.1 - 2017-09-22
  * Fix NPE on startup

5.5.0.0-beta1 - 2017-09-18
  * Upgrade to Elasticsearch 5.5.0 + Cassandra 3.11.0
  * Improved support for concurrent mapping update.

2.4.5.5 - 2017-08-23
  * Fix #112 Don't search over SETS
  * [breaking change] Does not more index static document until index_static_document=true (default is false)

2.4.5.4 - 2017-07-30
  * Upgrade to Cassandra 3.11.0
  * Escape special characters in serialized primary keys #101

2.4.5.2 - 2017-05-26
  * Upgrade to Elasticsearch 2.4.5 + Cassandra 3.10
  * Add Windows launch scripts (#97)
  * Add retry to creation of metadata keyspace/table

2.4.5.1 - 2017-05-13
  * Upgrade to Elasticsearch 2.4.5 + Cassandra 3.0.13
  * Undo CASSANDRA-13499 and now use a Cassandra custom per-row index.

2.4.2-13 - 2017-04-25
  * Elassandra is now an Elasticsearch plugin for Strapdata-Cassandra.
  * [breaking change] change _token DocValues type from SORTED_NUMERIC to NUMERIC (remove your indices before upgrade and reindex)
  * [breaking change] change _token default Lucene numeric precision (remove your indices before upgrade and reindex)
  * Add a new token_numeric_precision option to optimize search request.
  * Improve indexing on concurrent partition update.
  * Add a new index_on_compaction option to avoid lucene tombestones when compacting SSTables.
  * Add token ranges bitset caching to improve search performances.
  * Add support for pluggable CQL generic functions CASSANDRA-13267
  * Fix cluster state blocks updates in a multi-node cluster.
  * Add rebuild+cleanup command to rebuild and clean index from cassandra on all nodes.
  * #92 Add support for dynamic mapping for Cassandra map.

2.4.2-12 - 2017-04-11
  * Fix #87 null_value support
  * Fix #91 update of a C* mas as a nested field.

2.4.2-11 - 2017-03-10
  * Fix indexing in debug mode only #84

2.4.2-10 - 2017-03-02
  * Add support for blob and inet CQL types in primary key (in elasticsearch _id).
  * Fix partitioned index #83 and provides customizable partition function implementations.
  * Include optimized Cassandra serializers CASSANDRA-13271

2.4.2-9 - 2017-02-26
  * Significant write performance improvement.
  * New optimized version less internal engine (don't store any more version number in lucene files).
  * Fix erroneous deletion when compacting SSTables.
  * Fix a routing table issue in a multi-node single-rack cluster.

2.4.2-8 - 2017-02-15
  * Fix support for uuid and timeuuid Cassandra Types #74

2.4.2-7 - 2017-02-09
  * Enable support for multi-threading rebuild_index on Cassandra 3.
  * Add new _meta fields in order to index static columns.
  * Fix a minor bug on type checking.

2.4.2-6 - 2017-01-25
  * Fix ES Auto-Discovery fails when Cassandra table has at least one UDT field (#77)
  * Allows a text columns containing a geohash to be mapped to a geo_point.

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