Reply
Highlighted
Contributor
Posts: 95
Registered: ‎08-07-2017

mapreduce job not going beyond acceptance state

Hello All,

 

I have one application for which I am facing issue. When I start the application, it will start one job, resources get allocated to it an dsecond job starts. But second job is not going beyond acceptance state.

 

I am using fair scheduler and there has been no change to the queue.

 

When I kill first job, second job runs.

 

Please suggest.

 

Thanks,

Priya

Posts: 1,673
Kudos: 329
Solutions: 263
Registered: ‎07-31-2013

Re: mapreduce job not going beyond acceptance state

If it is a single-node or very small cluster (< 3-5 nodes), see https://community.cloudera.com/t5/Batch-Processing-and-Workflow/JOB-Stuck-in-Accepted-State/td-p/294...

Visiting the application page in RM Web UI will typically show the outstanding resource demands and the availability.
Contributor
Posts: 95
Registered: ‎08-07-2017

Re: mapreduce job not going beyond acceptance state

Hi Harsh,

 

Thanks for reply. I did check the resource WebUI, its showing 2 active nodes out of 3 nodes and their resources like memory and cores. Third node is in bad health due to less disk space. So, does that mean no container can be launced on that node?

 

Also,  I cannot allocate more container memory too.

 

Kindly suggest.

 

Thanks,

Priya

Posts: 1,673
Kudos: 329
Solutions: 263
Registered: ‎07-31-2013

Re: mapreduce job not going beyond acceptance state

Yes, if the NodeManager is placed under a bad-health state by the
ResourceManager, it will no longer be considered for any more container
allocation requests until the health check is resolved (more free space or
revised config for space check).
Contributor
Posts: 95
Registered: ‎08-07-2017

Re: mapreduce job not going beyond acceptance state

 

 

Thanks for the reply Harsh. So, for container allocation/launch, along with memory and cores, hard disk space also plays significant role.

 

Thanks,

Priya

Announcements