map task failed to report status

[Ted Yu <yu...@gmail.com>]

Hi,
My map task repeatedly was cleaned up by hadoop because of:
Task attempt_200912071308_2358_m_000000_3 failed to report status for 601
seconds. Killing!

The task sends requests to another server using compressed MINA messages.

Can someone suggest where I should add more debugging information ?

Thanks

Re: map task failed to report status

[Edward Capriolo <ed...@gmail.com>]

On Mon, Jan 11, 2010 at 5:00 PM, Ted Yu <yu...@gmail.com> wrote:
> Hi,
> My map task repeatedly was cleaned up by hadoop because of:
> Task attempt_200912071308_2358_m_000000_3 failed to report status for 601
> seconds. Killing!
>
> The task sends requests to another server using compressed MINA messages.
>
> Can someone suggest where I should add more debugging information ?
>
> Thanks
>

One cause of this could be your algorithm takes longer then 600
seconds to complete an iteration, and since map/reduce does not get a
status update it assumes the task is stalled.

If you have a long running piece of code use the reporter object to
signify progress to the API.

If that is not the case, please refer to your standard hadoop logs to
try and determine what is going on:

http://www.cloudera.com/blog/2009/09/15/apache-hadoop-log-files-where-to-find-them-in-cdh-and-what-info-they-contain/

Good luck