[HIVE-6409] FileOutputCommitterContainer::commitJob() cancels delegation tokens too early.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.12.0, 0.13.0
• Fix Version/s: 0.13.0
• Component/s: HCatalog
• Tags: FileOutputCommitterContainer delegation tokens
• Assignee: Mithun Radhakrishnan
• Reporter: Mithun Radhakrishnan
• Created: 11/Feb/14 23:52
• Updated: 03/Mar/14 21:53
• Resolved: 03/Mar/14 21:53

Description

When HCatalog's FileOutputCommitterContainer::commitJob() is run, it calls the underlying OutputCommitter and then attempts to register partitions in HCatalog.

If the commit fails (for example, because of HIVE-4996), commitJob() cancels delegation tokens retrieved from HCatalog before the exception is rethrown.

java.io.IOException: java.lang.reflect.InvocationTargetException
    at
org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigOutputCommitter.commitJob(PigOutputCommitter.java:185)
    at
org.apache.hadoop.mapreduce.v2.app.commit.CommitterEventHandler$EventProcessor.handleJobCommit(CommitterEventHandler.java:249)
    at
org.apache.hadoop.mapreduce.v2.app.commit.CommitterEventHandler$EventProcessor.run(CommitterEventHandler.java:212)
    at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:722)
Caused by: java.lang.reflect.InvocationTargetException
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:601)
    at
org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigOutputCommitter.commitJob(PigOutputCommitter.java:183)
    ... 5 more
Caused by: org.apache.hcatalog.common.HCatException : 2006 : Error adding
partition to metastore. Cause :
MetaException(message:java.lang.RuntimeException: commitTransaction was called
but openTransactionCalls = 0. This probably indicates that there are unbalanced
calls to openTransaction/commitTransaction)
    at
org.apache.hcatalog.mapreduce.FileOutputCommitterContainer.registerPartitions(FileOutputCommitterContainer.java:712)

The problem is that this happens before abortJob() has had a chance to run, thus yielding the following error:

MetaException(message:Could not connect to meta store using any of the URIs provided. Most recent failure: org.apache.thrift.transport.TTransportException: Peer indicated failure: DIGEST-MD5: IO error acquiring password
	at org.apache.thrift.transport.TSaslTransport.receiveSaslMessage(TSaslTransport.java:190)
	at org.apache.thrift.transport.TSaslTransport.open(TSaslTransport.java:266)
	at org.apache.thrift.transport.TSaslClientTransport.open(TSaslClientTransport.java:37)
	at org.apache.hadoop.hive.thrift.client.TUGIAssumingTransport$1.run(TUGIAssumingTransport.java:52)
	at org.apache.hadoop.hive.thrift.client.TUGIAssumingTransport$1.run(TUGIAssumingTransport.java:49)
	at java.security.AccessController.doPrivileged(Native Method)
...

I'll have a patch out that only cancels delegation tokens if the commitJob() has succeeded.