[HBASE-9165] Improvements to addDependencyJars

• Type: Sub-task
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.95.2
• Fix Version/s: 0.98.0, 0.96.1, 0.94.14
• Component/s: mapreduce
• Hadoop Flags: Reviewed

Release Note

Introduces the method TableMapReduceUtil#addHBaseDependencyJars for adding HBase and its direct dependencies (only) to the job configuration.

This is intended as a low-level API, facilitating code reuse between this class and its mapred counterpart. It is also of use to external tools that need to build a MapReduce job that interacts with HBase but want fine-grained control over the jars shipped to the cluster. See also PIG-3285 and HIVE-2055.

• Assignee: Nick Dimiduk
• Reporter: Nick Dimiduk
• Created: 08/Aug/13 23:31
• Updated: 26/Nov/13 05:53
• Resolved: 15/Nov/13 21:57

Description

The way we support adding HBase dependencies to a MapReduce job in TableMapReduceUtils#addDependencyJars(job) is a bit monolithic. Advanced users need a way to add HBase and its dependencies to their job without us snooping around for ouput formats and the like (see PIG-3285). We can also benefit from a little more code reuse between our mapred and mapreduce namespaces.

Issue Links

• is related to: HBASE-9112 Custom TableInputFormat in initTableMapperJob throws ClassNoFoundException on TableMapper
• is related to: HBASE-9484 Backport 8534 "Fix coverage for org.apache.hadoop.hbase.mapreduce" to 0.96