Stefan's documentation greatly helped

http://wiki.media-style.com/display/nutchDocu/Home

Since we configure all the three machines with the same configuration files

The task tracker knows the name of the job tracker.I overlooked this fact

If the namenode and jobtracker is also a  datanode , will the task tracker
copy the files to the other datanode using NDFS

or will it use its local filesystem to write the fetched segments

Rgds.
Prabhu


On 1/7/06, Raghavendra Prabhu <[EMAIL PROTECTED]> wrote:
>
> Hi Stefan
>
> Thanks for helping out .
> I wanted to try out a mapred system after reading your tutorial .It was
> simple and that is why i wanted to try it out
>
> I guess the when u start the jobtracker it listens on a port ,and the job
> tracker will send instructions through this port
>
> Thats is why we dont have any configuration for job-tracker in
> nutch-default
>
> Am i right ?
>
> With Regards,
> Prabhu
>
>
>  On 1/7/06, Stefan Groschupf <[EMAIL PROTECTED]> wrote:
> >
> > May be this can help you:
> > http://wiki.media-style.com/display/nutchDocu/Home
> > http://wiki.media-style.com/display/nutchDocu/setup+a+map+reduce+multi
> > +box+system
> >
> > Stefan
> > Am 07.01.2006 um 18:30 schrieb Raghavendra Prabhu:
> >
> > > I have only three machines which i can use (one of them is a server)
> > >
> > > The server i plan to set it up as namenode ,jobtracker and datanode
> > >
> > > The other two machines can be setup as tasktracker datanodes.
> > >
> > > Is this the right configuration .If anyone can throw some light , i
> > > would
> > > gladly appreciate it.
> >
> > ---------------------------------------------------------------
> > company:         http://www.media-style.com
> > forum:        http://www.text-mining.org
> > blog:            http://www.find23.net
> >
> >
> >
> >
>

Reply via email to