Error Running MapReduce2 job using YARN

[#1]

We installed and configured HDP-2.5.0.0. When we run a mapreduce job it fails with the diagnostics message "We crashed durring a commit". The status on Map and Reduce tasks indicate that they were successful. Following is the error observed in syslog output. How do we troubleshoot this error?Any tips will be appreciated.

2016-10-03 15:19:20,017 FATAL [main] org.apache.hadoop.mapreduce.v2.app.MRAppMaster: Error starting MRAppMaster
java.io.IOException: Was asked to shut down.
	at org.apache.hadoop.mapreduce.v2.app.MRAppMaster$5.run(MRAppMaster.java:1559)
	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:1724)
	at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.initAndStartAppMaster(MRAppMaster.java:1553)
	at org.apache.hadoop.mapreduce.v2.app.MRAppMaster.main(MRAppMaster.java:1486)
2016-10-03 15:19:20,021 INFO [main] org.apache.hadoop.util.ExitUtil: Exiting with status 1

[#2]

wrong mapreduce.jobhistory.address property's value. try to copy mapred-site.xml from the cluster

[#3]

Thank you for the suggestion. I have verified the hostname and port for mapreduce.jobhistory.address and they are accurate. We are using the copy of the config from the cluster.

[#4]

@Vijay Lakshman

Can you check the container size allocated ? Seems you are hitting memory issue.

[#5]

Thank You for the suggestion. This appears to be plausible. Will allocate additional memory and post an update.

[#6 · Accepted solution]

This issue was a result of misconfiguration of one of the nodes having an ip address conflict with another server. Once the ip address was corrected, we were able to review the logs and address classpath errors. Thanks everyone for their suggestions.

[#7]

I faced similiar issue.

We had a HDP upgrade, still my jobs are pointing to the custom library (which are all of previous version). I have updated my custom library with latest jars. The issue is fixed.