[ https://issues.apache.org/jira/browse/TEZ-3699?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Rajesh Balamohan reassigned TEZ-3699: ------------------------------------- Assignee: Rajesh Balamohan > For large dataset, pipelined shuffle throws exceptions in consumer side for > UnorderedPartitioned edge > ----------------------------------------------------------------------------------------------------- > > Key: TEZ-3699 > URL: https://issues.apache.org/jira/browse/TEZ-3699 > Project: Apache Tez > Issue Type: Bug > Reporter: Rajesh Balamohan > Assignee: Rajesh Balamohan > Attachments: TEZ-3699.1.patch > > > {noformat} > 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) > {noformat} > 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. -- This message was sent by Atlassian JIRA (v6.3.15#6346)