running job without jar

[exception <ex...@taomee.com>]

Hi,

When launching a job in hadoop, usually we use "hadoop jar xxx.jar input output".
Can I run a job simply by a java program (In fully distributed mode), without packing a jar.
I know this will cause problems because the remote nodes don't have the source code and .class to run the mapper/reducer. So my next question is how to tell hadoop to copy user specified files to different nodes?

Thanks.

Re: running job without jar

[David Rosenstrauch <da...@darose.net>]

On 10/26/2010 03:01 AM, exception wrote:
> Hi,
>
> When launching a job in hadoop, usually we use "hadoop jar xxx.jar
> input output". Can I run a job simply by a java program (In fully
> distributed mode), without packing a jar. I know this will cause
> problems because the remote nodes don't have the source code and
> .class to run the mapper/reducer. So my next question is how to tell
> hadoop to copy user specified files to different nodes?
>
> Thanks.
>

Yes, this can be easily done.  See: 
http://www.mail-archive.com/mapreduce-user@hadoop.apache.org/msg01114.html

HTH,

DR

Re: running job without jar

[jingguo yao <ya...@gmail.com>]

As far as I know, there is no way to run a Java class on a cluster.

To copy user specified files to different nodes, you can use Distributed
Cache.

On Tue, Oct 26, 2010 at 3:01 PM, exception <ex...@taomee.com> wrote:

>  Hi,
>
>
>
> When launching a job in hadoop, usually we use “hadoop jar xxx.jar input
> output”.
>
> Can I run a job simply by a java program (In fully distributed mode),
> without packing a jar.
>
> I know this will cause problems because the remote nodes don’t have the
> source code and .class to run the mapper/reducer. So my next question is how
> to tell hadoop to copy user specified files to different nodes?
>
>
>
> Thanks.
>