Apache Solr 4.4 发布了，这是跟进 Lucene 4.4 的更新版本。
下载地址：
http://lucene.apache.org/solr/mirrors-solr-latest-redir.html
值得关注的改进内容：
* Solr indexes and transaction logs may stored in HDFS with full read/write
capability.
* Schemaless mode: Added support for a mode that requires no up-front schema
modifications, in which previously unknown fields' types are guessed based
on the values in added/updated documents, and are then added to the schema
prior to processing the update. Note that the below-described features
are also useful independently from schemaless mode operation.
* New Parse{Date,Integer,Long,Float,Double,Boolean}UpdateProcessorFactory
classes parse/guess the field value class for String-valued and unknown
fields.
* New AddSchemaFieldsUpdateProcessor: Automatically add new fields to the
schema when adding/updating documents with unknown fields. Custom rules
map field value class(es) to schema fieldTypes.
* A new schemaless mode example configuration, using the above-described
field-value-class-guessing and unknown-field-schema-addition features,
is provided at solr/example/example-schemaless/.
* Core Discovery mode: A new solr.xml format which does not store core
information, but instead searches for files named 'core.properties' in
the filesystem which tell Solr all the details about that core. The main
example and the schemaless example both use this new format.
* Schema REST API: Add support for creating copy fields.
* A merged segment warmer may now be plugged into solrconfig.xml.
* New MaxScoreQParserPlugin: Return max() instead of sum() of terms.
* Binary files are now supported in ZooKeeper.
* SolrJ's SolrPing object has new methods for ping, enable, and disable.
* The Admin UI now supports adding documents to Solr.
* Added a PUT command to the Solr ZkCli tool.
* New "deleteshard" collections API that unloads all replicas of a given
shard and then removes it from the cluster state. It will remove only
those shards which are INACTIVE or have no range.
* The Overseer can now optionally assign generic node names so that
new addresses can host shards without naming confusion.
* The CSV Update Handler now supports optionally adding the line number/
row id to a document.
* Added a new system wide info admin handler that exposes the system info
that could previously only be retrieved using a SolrCore.
来自:开源中国社区
