ConfX created MAPREDUCE-7445:
--------------------------------
Summary: ShuffleSchedulerImpl causes ArithmeticException due to
improper detailsInterval value checking
Key: MAPREDUCE-7445
URL: https://issues.apache.org/jira/browse/MAPREDUCE-7445
Project: Hadoop Map/Reduce
Issue Type: Bug
Reporter: ConfX
Attachments: reproduce.sh
h2. What happened
There is no value checking for parameter
{{{}mapreduce.reduce.shuffle.maxfetchfailures{}}}. This may cause improper
calculations and crashes the system like division by 0.
h2. Buggy code
In {{{}ShuffleSchedulerImpl.java{}}}, there is no value checking for
{{maxFetchFailuresBeforeReporting}} and this variable is directly passed to
method {{{}checkAndInformMRAppMaster{}}}. When
{{maxFetchFailuresBeforeReporting }} is mistakenly set to 0, the code would
cause division by 0 and throw ArithmeticException to crash the system.
{noformat}
private void checkAndInformMRAppMaster(
...
if (connectExcpt || (reportReadErrorImmediately && readError)
|| ((failures % maxFetchFailuresBeforeReporting) == 0) || hostFailed) {
...
}{noformat}
h2. How to reproduce
(1) set {{{}mapreduce.reduce.shuffle.maxfetchfailures{}}}={{{}0{}}},
{{{}mapreduce.reduce.shuffle.notify.readerror{}}}={{{}false{}}}
(2) run {{mvn surefire:test
-Dtest=org.apache.hadoop.mapreduce.task.reduce.TestShuffleScheduler#TestSucceedAndFailedCopyMap}}
h2. Stacktrace
{noformat}
java.lang.ArithmeticException: / by zero
at
org.apache.hadoop.mapreduce.task.reduce.ShuffleSchedulerImpl.checkAndInformMRAppMaster(ShuffleSchedulerImpl.java:347)
at
org.apache.hadoop.mapreduce.task.reduce.ShuffleSchedulerImpl.copyFailed(ShuffleSchedulerImpl.java:308)
at
org.apache.hadoop.mapreduce.task.reduce.TestShuffleScheduler.TestSucceedAndFailedCopyMap(TestShuffleScheduler.java:285){noformat}
For an easy reproduction, run the reproduce.sh in the attachment.
We are happy to provide a patch if this issue is confirmed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: mapreduce-dev-h...@hadoop.apache.org
