[HBASE-9533] List of dependency jars for MR jobs is hard-coded and does not include netty, breaking MRv1 jobs

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.95.2, 0.96.1
• Fix Version/s: 0.98.0, 0.96.0
• Component/s: mapreduce
• Hadoop Flags: Reviewed
• Assignee: Matteo Bertozzi
• Reporter: Aleksandr Shulman
• Created: 14/Sep/13 00:54
• Updated: 20/Nov/15 11:55
• Resolved: 20/Sep/13 03:41

Description

Observed behavior:

Against trunk, using MRv1 with hadoop 1.0.4, r1393290, I am able to run MRv1 jobs (e.g. pi 2 4).

However, when I use it to run MR over HBase jobs, they fail with the stack trace below.

From the trace, the issue seems to be that it cannot find a class that the netty jar contains. This would make sense, given that the dependency jars that we use for the MapReduce job are hard-coded, and that the netty jar is not one of them.

https://github.com/apache/hbase/blob/trunk/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.java#L519

Strangely, this is only an issue in trunk, not 0.95, even though the code hasn't changed.

Command:

/bin/hbase org.apache.hadoop.hbase.mapreduce.RowCounter sampletable

TT logs (attached)

Output from console running job:

13/09/13 16:02:58 INFO mapred.JobClient: Task Id : attempt_201309131601_0002_m_000000_2, Status : FAILED
java.io.IOException: Cannot create a record reader because of a previous error. Please look at the previous logs lines from the task's full log for more details.
	at org.apache.hadoop.hbase.mapreduce.TableInputFormatBase.createRecordReader(TableInputFormatBase.java:119)
	at org.apache.hadoop.mapred.MapTask$NewTrackingRecordReader.<init>(MapTask.java:489)
	at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:731)
	at org.apache.hadoop.mapred.MapTask.run(MapTask.java:370)
	at org.apache.hadoop.mapred.Child$4.run(Child.java:255)
	at java.security.AccessController.doPrivileged(Native Method)
	at javax.security.auth.Subject.doAs(Subject.java:396)
	at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1149)
	at org.apache.hadoop.mapred.Child.main(Child.java:249)

13/09/13 16:03:09 INFO mapred.JobClient: Job complete: job_201309131601_0002
13/09/13 16:03:09 INFO mapred.JobClient: Counters: 7
13/09/13 16:03:09 INFO mapred.JobClient:   Job Counters 
13/09/13 16:03:09 INFO mapred.JobClient:     SLOTS_MILLIS_MAPS=29913
13/09/13 16:03:09 INFO mapred.JobClient:     Total time spent by all reduces waiting after reserving slots (ms)=0
13/09/13 16:03:09 INFO mapred.JobClient:     Total time spent by all maps waiting after reserving slots (ms)=0
13/09/13 16:03:09 INFO mapred.JobClient:     Launched map tasks=4
13/09/13 16:03:09 INFO mapred.JobClient:     Data-local map tasks=4
13/09/13 16:03:09 INFO mapred.JobClient:     SLOTS_MILLIS_REDUCES=0
13/09/13 16:03:09 INFO mapred.JobClient:     Failed map tasks=1

Expected behavior:
As a stopgap, the netty jar should be included in that list. More generally, there should be a more elegant way to include the jars that are needed.