This is the way I understand your question:
"if I ssh into a slave node and want to start jaql, what do I need to
do?"
So, my answer to this question is yes, however its configured on the
client machine, it must also be configured the same
way on the slave nodes.
However, I don't really know what you want to do... the typical way I
run Jaql is to set it up on a client machine, e.g., my laptop,
and point it to a hadoop cluster-- there is no requirement to install
jaql on all nodes, slave or master and still have it work
with map-reduce. To the hadoop cluster, Jaql just looks like some
program that submits map-reduce jobs and passes in a jaql.jar
+ all other udf/module jars that you've specified in your jaql
session.
On Mar 23, 6:25 am, KRISHNA RAO CHEPYALA
<
chepyalakrishna...@gmail.com> wrote:
> So, what you say is the slave node jaql distribution settings and the master
> node jaql distribution settings should be the same.
>
> On Wed, Mar 23, 2011 at 12:25 AM, Vuk Erecegovac <
vuk.ercego...@gmail.com>wrote:
>
>
>
> > Sure, you just need to make sure that that the jaql distribution is present
> > from where you launch the jaql client.
> > In particular, you'll need the jaql.jar, bin, conf, and lib dirs. Launching
> > it from a slave node should be no different than
> > launching it from a non-slave node.
>