Created on 01-18-2017 10:24 AM - edited 08-19-2019 02:02 AM
I have tried the example provided in the Hortonwork tutorial,
http://hortonworks.com/hadoop-tutorial/defining-processing-data-end-end-data-pipeline-apache-falcon/
I have followed the steps has provided in that tutorial but facing issue when executing "rawEmailIngestProcess", all the Jobs are KILLLED.
I have followed the below links to identify root cause,
Checked in Resouremanger Log.
Checked in Oozie log.
https://community.hortonworks.com/articles/9148/troubleshooting-an-oozie-flow.html
No error message in resource manger log, Found the error in the oozie log file.
Launcher ERROR, reason: Main class [org.apache.oozie.action.hadoop.ShellMain], exit code [1]
Created 01-18-2017 10:35 AM
Can you check an instance of shell-wf job . Do you see any link in the console URL section for this.
If a yarn job is launched , you can view the logs further using below command or by directly viewing through RM UI
yarn logs -applicationId <application_ID>.
Created 01-18-2017 10:27 AM
Please click on one of the KILLED workflow --> Click on failed action --> See if you are getting any error in error section
If not then click on Console URL --> It will land you to RM UI --> Click on succeeded mapper --> Click on logs --> Check stderr logs
You should get some hint over there.
Created 01-18-2017 10:33 AM
Yes I got the below error from Oozie,
2017-01-18 09:19:56,597 WARN ShellActionExecutor:523 - SERVER[] USER[ambari-qa] GROUP[-] TOKEN[] APP[shell-wf] JOB[0000055-170117070035733-oozie-oozi-W] ACTION[0000055-170117070035733-oozie-oozi-W@shell-node] Launcher ERROR, reason: Main class [org.apache.oozie.action.hadoop.ShellMain], exit code [1]
Created 01-18-2017 10:35 AM
But no error message in the stderr log.. Is there any workaround to fix this issue...
Created 01-18-2017 10:35 AM
Can you check an instance of shell-wf job . Do you see any link in the console URL section for this.
If a yarn job is launched , you can view the logs further using below command or by directly viewing through RM UI
yarn logs -applicationId <application_ID>.
Created 01-18-2017 12:34 PM
yarn logs -applicationId <application_ID> .. Helped me to identify the issue.. It is a permission issue for the user and I have fixed it.
Created 04-05-2017 02:37 PM
Hi! -- I am experiencing the same issue with the same tutorial (Falcon) in a Google Cloud based HDP installation. I'd be grateful if you could add some more info regarding that permission issue. Thanks in advance.
Created 07-07-2017 07:48 AM
Hi, You're my life saviour. I tried checking the logs on the schedular website but just couldnt find it.
and then I used this Yarn log and saved it into a log.txt file and there you go. Found the exact error that I was looking for.
Thank you so much. 🙂
Created 09-22-2017 09:31 AM
hi i added spark node in oozie workflow but getting this error all time {
reason: Main class [org.apache.oozie.action.hadoop.SparkMain], exit code [101]