[TEZ-3699] For large dataset, pipelined shuffle throws exceptions in consumer side for UnorderedPartitioned edge

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.9.0
• Fix Version/s: 0.9.0
• Hadoop Flags: Reviewed
• Assignee: Rajesh Balamohan
• Reporter: Rajesh Balamohan
• Created: 26/Apr/17 04:56
• Updated: 22/Aug/17 00:02
• Resolved: 26/Apr/17 23:44

Description

2017-04-25 21:16:18,179 [INFO] [Fetcher_B {Map_1} #1] |ShuffleManager.fetch|: Completed fetch for attempt: {110, 0, attempt_1490656001509_1098_1_04_000110_0_10013_4, 2, 4} to MEMORY, csize=10505, dsize=32638, EndTime=1493180178179, TimeTaken=1, Rate=10.01 MB/s
2017-04-25 21:16:18,183 [INFO] [Fetcher_B {Map_1} #2] |HttpConnection.url|: for url=http://node120:13562/mapOutput?job=job_1490656001509_1098&dag=1&reduce=43&map=attempt_1490656001509_1098_1_04_000075_0_10008_15,attempt_1490656001509_1098_1_04_000035_0_10008_16,attempt_1490656001509_1098_1_04_
000055_0_10007_16,attempt_1490656001509_1098_1_04_000055_0_10007_16,attempt_1490656001509_1098_1_04_000075_0_10008_16,attempt_1490656001509_1098_1_04_000195_0_10013_1,attempt_1490656001509_1098_1_04_000075_0_10008_17,attempt_1490656001509_1098_1_04_000180_0_10013_1,attempt_1490656001509_1098_1_04_000195_0_10013_2,att
empt_1490656001509_1098_1_04_000075_0_10008_18,attempt_1490656001509_1098_1_04_000180_0_10013_2,attempt_1490656001509_1098_1_04_000195_0_10013_3&keepAlive=true sent hash and receievd reply 0 ms
2017-04-25 21:16:18,183 [INFO] [Fetcher_B {Map_1} #2] |ShuffleManager.fetch|: Completed fetch for attempt: {75, 0, attempt_1490656001509_1098_1_04_000075_0_10008_15, 1, 15} to MEMORY, csize=23437, dsize=70987, EndTime=1493180178183, TimeTaken=0, Rate=0.00 MB/s
2017-04-25 21:16:18,183 [INFO] [Fetcher_B {Map_1} #2] |shuffle.Fetcher|: Failed to read data to memory for InputAttemptIdentifier [inputIdentifier=35, attemptNumber=0, pathComponent=attempt_1490656001509_1098_1_04_000035_0_10008_16, spillType=1, spillId=16]. len=0, decomp=0. ExceptionMessage=Not a valid ifile header
2017-04-25 21:16:18,185 [WARN] [Fetcher_B {Map_1} #2] |shuffle.Fetcher|: Failed to shuffle output of InputAttemptIdentifier [inputIdentifier=35, attemptNumber=0, pathComponent=attempt_1490656001509_1098_1_04_000035_0_10008_16, spillType=1, spillId=16] from cn120-10.l42scl.hortonworks.com
java.io.IOException: Not a valid ifile header
        at org.apache.tez.runtime.library.common.sort.impl.IFile$Reader.verifyHeaderMagic(IFile.java:833)
        at org.apache.tez.runtime.library.common.sort.impl.IFile$Reader.isCompressedFlagEnabled(IFile.java:840)
        at org.apache.tez.runtime.library.common.sort.impl.IFile$Reader.readToMemory(IFile.java:608)
        at org.apache.tez.runtime.library.common.shuffle.ShuffleUtils.shuffleToMemory(ShuffleUtils.java:134)
        at org.apache.tez.runtime.library.common.shuffle.Fetcher.fetchInputs(Fetcher.java:814)
        at org.apache.tez.runtime.library.common.shuffle.Fetcher.doHttpFetch(Fetcher.java:539)
        at org.apache.tez.runtime.library.common.shuffle.Fetcher.doHttpFetch(Fetcher.java:428)
        at org.apache.tez.runtime.library.common.shuffle.Fetcher.callInternal(Fetcher.java:226)
        at org.apache.tez.runtime.library.common.shuffle.Fetcher.callInternal(Fetcher.java:73)
        at org.apache.tez.common.CallableWithNdc.call(CallableWithNdc.java:36)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
        at java.lang.Thread.run(Thread.java:745)        

Couple of issue. One is related to 0 length partition.
Second is related to duplicate entry for attempt_1490656001509_1098_1_04_000055_0_10007_16 (need to check if it has got different spill type)

For smaller datasets it works fine. But when there is large number of spills, this happens.