You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-user@hadoop.apache.org by Pedro Costa <ps...@gmail.com> on 2010/06/24 00:13:09 UTC

How many ports and processes are open/created in MR?

Hi,

1 - Hadoop uses several ports to run. It exists ports for HDFS, for the
MapReduce JvmTasks, etc. I don't know how I can identify all the ports that
a MapReduce and HDFS uses. I'm running the wordcount example, and I would
like to see what ports are open and what are their purpose. Where can I get
this information?


2 - Running the example of the WordCount in MapReduce, I noticed that it's
created several new process. I know that it's created new process for a
TaskTracker, for a JobTracker and for a TaskInProgress. Is this all the
processes that are created on MapReduce during the execution of the
wordcount example? How many processes are created in total?


Thanks,
-- 
Pedro

Re: How many ports and processes are open/created in MR?

Posted by Allen Wittenauer <aw...@linkedin.com>.
On Jun 23, 2010, at 3:13 PM, Pedro Costa wrote:
> 1 - Hadoop uses several ports to run. It exists ports for HDFS, for the MapReduce JvmTasks, etc. I don't know how I can identify all the ports that a MapReduce and HDFS uses. I'm running the wordcount example, and I would like to see what ports are open and what are their purpose. Where can I get this information?
> 

netstat, because most ports are ephemeral.


Re: How many ports and processes are open/created in MR?

Posted by Hemanth Yamijala <yh...@gmail.com>.
Pedro,

> 1 - Hadoop uses several ports to run. It exists ports for HDFS, for the
> MapReduce JvmTasks, etc. I don't know how I can identify all the ports that
> a MapReduce and HDFS uses. I'm running the wordcount example, and I would
> like to see what ports are open and what are their purpose. Where can I get
> this information?
>
>
> 2 - Running the example of the WordCount in MapReduce, I noticed that it's
> created several new process. I know that it's created new process for a
> TaskTracker, for a JobTracker and for a TaskInProgress. Is this all the
> processes that are created on MapReduce during the execution of the
> wordcount example? How many processes are created in total?

There are 2 daemons that are launched when the MapReduce framework
starts - JobTracker and TaskTracker. Every map or reduce task runs in
a separate process called Child (you could grep for Child to identify
these). Possibly, this is what you meant by 'TaskInProgress'.
TaskInProgress is an internal class of the framework and as such does
not relate to any process launched. The number of child processes
launched will be equal to the number of Map and reduce tasks launched
on a slave node.

Thanks
Hemanth