Hello Jasson
We probably have to update the documentation, but when using distributed
scheduling, the mapreduce task must be submitted with the following extra
configuration:
-Dyarn.resourcemanager.scheduler.address=127.0.0.1:8049
Essentially, the AM should talk to the DistributedScheduler
Hi Wei,
Can you please share your yarn-site.xml and the command you used for
running wordcount?
Thanks,
Konstantinos
On Tue, Jun 13, 2017 at 11:44 Jasson Chenwei
wrote:
> hi all
>
> I have set up a distributed scheduler using a new feature in Hadoop-3.0. My
> Hadoop
hi all
I have set up a distributed scheduler using a new feature in Hadoop-3.0. My
Hadoop version is hadoop-3.0.0-appha3. I have enabled the
opportunistic container and distributed scheduler in
yarn-site.xml following the guide. But wordcount application master failed
to launch as follow: