[
https://issues.apache.org/jira/browse/SQOOP-2151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14537506#comment-14537506
]
Ted Malaska commented on SQOOP-2151:
------------------------------------
Thx Gwen,
I was able to recreate the problem repeatedly.
The problem is two part, one it is a race condition and the other is the
integration testing does a catch on exceptions where the unit tests don't.
So first the race condition. This is what happens to make the lock happen
- Thread 1 - Lock free
- Thread 2 - Exception
- Thread 2 - Released Free
- Thread 1 - Lock free
- Thread 1 - future.get
- Thread 1 - Exception thrown
- Thread 1 - Exception catched
- Thread 1 - close
- Thread 1 - Try to lock free
- Stuck
The solution is simple: Just add the following to waitForConsumer() when you
catch an exception:
if (free.availablePermits() == 0) {
free.release();
}
I also think we should add the try catch to the unit test to better align with
the integrated tests. So in short we catch exceptions when writing and then no
matter the exception we try to close the writer.
Let me know what you think
Ted Malaska
> Sqoop2: Sqoop mapreduce job gets into deadlock when loader throws an exception
> ------------------------------------------------------------------------------
>
> Key: SQOOP-2151
> URL: https://issues.apache.org/jira/browse/SQOOP-2151
> Project: Sqoop
> Issue Type: Bug
> Affects Versions: 1.99.5
> Reporter: Jarek Jarcec Cecho
> Assignee: Ted Malaska
> Priority: Blocker
> Fix For: 2.0.0
>
>
> I'm working on Kite integration tests and I've noticed that there is certain
> case where Sqoop mapreduce job gets into deadlock.
> I've get there by running Kite job after upgrading to Kite 1.0 but before
> fixing the temporary data set problem covered by SQOOP-2150. Here is the log
> output from mapper:
> {code}
> 2015-02-28 09:14:50,994 [OutputFormatLoader-consumer] INFO
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor -
> SqoopOutputFormatLoadExecutor consumer thread is starting
> 2015-02-28 09:14:51,021 [OutputFormatLoader-consumer] INFO
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor - Running loader class
> org.apache.sqoop.connector.kite.KiteLoader
> 2015-02-28 09:14:51,025 [main] INFO org.apache.sqoop.job.mr.SqoopMapper -
> Starting progress service
> 2015-02-28 09:14:51,030 [main] INFO org.apache.sqoop.job.mr.SqoopMapper -
> Running extractor class org.apache.sqoop.connector.jdbc.GenericJdbcExtractor
> 2015-02-28 09:14:51,306 [main] INFO
> org.apache.sqoop.connector.jdbc.GenericJdbcExtractor - Using query: SELECT *
> FROM FROMRDBMSTOKITETEST WHERE 1 <= "id" AND "id" <= 4
> 2015-02-28 09:14:51,627 [OutputFormatLoader-consumer] ERROR
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor - Error while loading
> data out of MR job.
> org.kitesdk.data.ValidationException: Dataset name
> temp_9975e79a-7e5d-493a-b6d4-646f3452a51f is not alphanumeric (plus '_')
> at
> org.kitesdk.data.ValidationException.check(ValidationException.java:55)
> at
> org.kitesdk.data.spi.Compatibility.checkDatasetName(Compatibility.java:103)
> at org.kitesdk.data.spi.Compatibility.check(Compatibility.java:66)
> at
> org.kitesdk.data.spi.filesystem.FileSystemMetadataProvider.create(FileSystemMetadataProvider.java:209)
> at
> org.kitesdk.data.spi.filesystem.FileSystemDatasetRepository.create(FileSystemDatasetRepository.java:137)
> at org.kitesdk.data.Datasets.create(Datasets.java:239)
> at org.kitesdk.data.Datasets.create(Datasets.java:307)
> at org.kitesdk.data.Datasets.create(Datasets.java:335)
> at
> org.apache.sqoop.connector.kite.KiteDatasetExecutor.createDataset(KiteDatasetExecutor.java:67)
> at
> org.apache.sqoop.connector.kite.KiteLoader.getExecutor(KiteLoader.java:51)
> at org.apache.sqoop.connector.kite.KiteLoader.load(KiteLoader.java:61)
> at org.apache.sqoop.connector.kite.KiteLoader.load(KiteLoader.java:36)
> at
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor$ConsumerThread.run(SqoopOutputFormatLoadExecutor.java:250)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> 2015-02-28 09:14:51,633 [main] INFO org.apache.sqoop.job.mr.SqoopMapper -
> Stopping progress service
> 2015-02-28 09:14:51,634 [main] INFO
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor -
> SqoopOutputFormatLoadExecutor::SqoopRecordWriter is about to be closed
> {code}
> But the mapper never finished, here is the relevant jstack:
> {code}
> "main" #1 prio=5 os_prio=31 tid=0x00007fedf180a800 nid=0xc07 waiting on
> condition [0x000000010b3f2000]
> java.lang.Thread.State: WAITING (parking)
> at sun.misc.Unsafe.park(Native Method)
> - parking to wait for <0x0000000127399b50> (a
> java.util.concurrent.Semaphore$FairSync)
> at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:997)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304)
> at java.util.concurrent.Semaphore.acquire(Semaphore.java:312)
> at
> org.apache.sqoop.job.mr.SqoopOutputFormatLoadExecutor$SqoopRecordWriter.close(SqoopOutputFormatLoadExecutor.java:113)
> at
> org.apache.hadoop.mapred.MapTask$NewDirectOutputCollector.close(MapTask.java:667)
> at org.apache.hadoop.mapred.MapTask.closeQuietly(MapTask.java:2012)
> at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:794)
> at org.apache.hadoop.mapred.MapTask.run(MapTask.java:341)
> at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:163)
> at java.security.AccessController.doPrivileged(Native Method)
> at javax.security.auth.Subject.doAs(Subject.java:422)
> at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1628)
> at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:158)
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
