The following property in my cluster was set to false. This needs to be
true for being able to read data from timeline server.
The setting that controls whether yarn system metrics is
published on the timeline server or not by RM.
yarn.resourcemanager.system-metrics-publisher.enabled
true
To add, below configuration also needs to be true to retrieve data using
yarn client.
Indicate to clients whether to query generic application
data from timeline history-service or not. If not enabled then application
data is queried only from Resource Manager.
yarn.timeline-service.generic-application-history.enabled
true
Thanks,
Ajay
On Wed, Nov 30, 2016 at 8:22 PM, AJAY GUPTA wrote:
> Hi
>
> I am running Hadoop 2.6.0 on a cluster setup. I have setup the timeline
> server on this cluster.
> My code wants to fetch details of KILLED containers for an application. I
> have used the yarnClient.getContainerReport() call to get
> details of containers. This call throws an exception for KILLED containers,
> despite timeline server being up.
> Is there some configuration I am missing here which is resulting in this
> issue or is this expected in hadoop 2.6.0?
>
> I have a local setup with Hadoop 2.7.2. It is able to give details of
> KILLED containers without any exception.
>
> Thanks,
> Ajay
>