What's New in Cloudera Search
Learn about the new features of Cloudera Search in Cloudera Runtime 7.1.6.
Single-step metadata transition
The solr-upgrade.sh
script now validates and transforms all your
configuration metadata in a single step. You no longer need to individually run the script
for each and every collection configuration set and then copy the files manually to the
destination directory. Besides transitioning configuration files to the format of the
upgrade target, the script also produces a set of *_validation.html
files
for all the steps of the migration that you can investigate, should the script stop on
error. After fixing the incompatibility in the input folder, you can simply rerun the script
as it overwrites the contents of the output directory on each run.
Local file system support
Cloudera Search now supports using local file system instead of HDFS. For more information, see Local file system support.
SolrCell reintroduced
SolrCell is reintroduced to Cloudera Search with limited functionality, which means that it can only be used with Morphlines as the JAR is not added to the Solr server classpath.
Crunch Indexer socket timeout is configurable
The Tomcat HTTP client library uses a socket timeout of 10 minutes. Spark Crunch Indexer does
not override this value, and in case indexing a single batch takes more than 10 minute the
entire indexing job fails. This can happen especially if the morphlines contain
DeleteByQuery
requests.
As of this release, you can configure the socket timeout for the connection in the morphline
file by adding the solrClientSocketTimeout
parameter to the
solrLocator
command.
For example:
SOLR_LOCATOR :
{
collection : test_collection
zkHost : "zookeeper1.example.com:2181/solr"
# 20 minutes in milliseconds
solrClientSocketTimeout: 1200000
# Max number of documents to pass per RPC from morphline to Solr Server
# batchSize : 10000
}