hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gwenhael Pasquiers <gwenhael.pasqui...@ericsson.com>
Subject RE: Yarn java client
Date Thu, 24 Mar 2016 14:06:33 GMT
Yes, and everything seems fine.

The behavior just changed : now, once the container finished, the application goes into failed
state, the diagnostics says :

Application application_XXXXXXX failed 1 times due to AM Container for appattempt_XXXXXXXXX
exited with exitCode: 0.
For more detailed … see logs

There is only one attempt and the logs are all ok. Does the program has to exit with a specific
exitcode other than zero in case of success ?

From: Youzha [mailto:yuza.rasfar@gmail.com]
Sent: jeudi 24 mars 2016 14:52
To: Gwenhael Pasquiers
Cc: user@hadoop.apache.org
Subject: Re: Yarn java client


have u check the logs?
On Mar 24, 2016 8:12 PM, "Gwenhael Pasquiers" <gwenhael.pasquiers@ericsson.com<mailto:gwenhael.pasquiers@ericsson.com>>
wrote:
Hi ;

I’m sorry if this has already been ask but I didn’t find a solution.

I’ve developed a Java program that make a bunch of operations on HDFS and I now want to
execute it on yarn.
I’ve developed a Yarn client (we’re on 2.6.0); I think it’s close to working but not
entirely :

-          The yarn application is stuck on the ACCEPTED state (yes, there are enough resources
available)

-          While it’s on ACCEPTED state it starts my container, my program runs correctly
until it’s on SUCCEEDED sate.

-          The application stays on ACCEPTED state and restarts the container (my program)
infinitely…

Could anyone here give me a pointer, a clue ?

Thanks in advance,

B.R.

Gwen’
Mime
View raw message