Using HDFS with Event Persistence
Before you can store events with Event Persistence using HDFS as the storage engine, you must configure the Hadoop cluster. You must copy the custom Hive SerDe and Joda Date/Time libraries from your Event Persistence installation to your HDFS CDH 5.3.0 distribution.
To configure HDFS CDH 5.3.0 as the storage engine for Event Persistence:
1. In your Software AG installation, locate the Joda Date/Time and the Event Persistence Hive SerDe .jar files:
joda-time_2.9.3.jar - available in the
Software AG_directory \common\runtime\bundles\platform\eclipse\plugins directory.
com.softwareag.evp.hive.serde_10.0.0.0000-
nnnn.jar - available in the
Software AG_directory \common\runtime\bundles\evs\eclipse\plugins directory, where
nnnn is the build number of your
Event Persistence installation.
2. Copy both files to the Hive library directory on all nodes in the Hadoop cluster where Hive is running, for example CDH5.3.0_directory/var/lib/hive/lib.
3. Copy both files to the Yarn library directory on all data nodes in the Hadoop cluster, for example CDH5.3.0_directory/var/lib/hadoop-yarn/lib or CDH5.3.0_directory/var/lib/hadoop-mapreduce/lib, if you are using MapReduce MRv1.
4. Restart Hive.
