You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by Sandhya E <sa...@gmail.com> on 2009/04/01 09:58:20 UTC
datanode but not tasktracker
Hi
When the host is listed in slaves file both DataNode and TaskTracker
are started on that host. Is there a way in which we can configure a
node to be datanode and not tasktracker.
Thanks
Sandhya
Re: datanode but not tasktracker
Posted by Owen O'Malley <om...@apache.org>.
On Apr 1, 2009, at 12:58 AM, Sandhya E wrote:
> Hi
>
> When the host is listed in slaves file both DataNode and TaskTracker
> are started on that host. Is there a way in which we can configure a
> node to be datanode and not tasktracker.
If you use hadoop-daemons.sh, you can pass a host list. So do:
ssh namenode hadoop-daemon.sh start namenode
ssh jobtracker hadoop-daemon.sh start jobtracker
hadoop-daemons.sh -hosts dfs.slaves start datanode
hadoop-daemons.hs -hosts mapred.slaves start tasktracker
-- Owen
Re: datanode but not tasktracker
Posted by Sandhya E <sa...@gmail.com>.
This is just a temporary requirement for me, ideally I always use same
set of hosts for tasktrackers and datanodes.
On Wed, Apr 1, 2009 at 8:22 PM, Bill Au <bi...@gmail.com> wrote:
> Not sure why you would want a node be a datanode but not a tasktracker
> because you normally would want the map/reduce task to run where the data is
> stored.
>
> Bill
>
> On Wed, Apr 1, 2009 at 3:58 AM, Sandhya E <sa...@gmail.com> wrote:
>
>> Hi
>>
>> When the host is listed in slaves file both DataNode and TaskTracker
>> are started on that host. Is there a way in which we can configure a
>> node to be datanode and not tasktracker.
>>
>> Thanks
>> Sandhya
>>
>
Re: datanode but not tasktracker
Posted by Bill Au <bi...@gmail.com>.
Not sure why you would want a node be a datanode but not a tasktracker
because you normally would want the map/reduce task to run where the data is
stored.
Bill
On Wed, Apr 1, 2009 at 3:58 AM, Sandhya E <sa...@gmail.com> wrote:
> Hi
>
> When the host is listed in slaves file both DataNode and TaskTracker
> are started on that host. Is there a way in which we can configure a
> node to be datanode and not tasktracker.
>
> Thanks
> Sandhya
>
Re: datanode but not tasktracker
Posted by Jean-Daniel Cryans <jd...@apache.org>.
Sandhya,
You can specify the file to use for slaves so instead of start-all you
can start-dfs with the normal slave file and start-mapred with a
specified file on the command line.
J-D
On Wed, Apr 1, 2009 at 3:58 AM, Sandhya E <sa...@gmail.com> wrote:
> Hi
>
> When the host is listed in slaves file both DataNode and TaskTracker
> are started on that host. Is there a way in which we can configure a
> node to be datanode and not tasktracker.
>
> Thanks
> Sandhya
>