Hi Edmar,

I am checking it now. In theory your approach should work.

Cheers,

-- Andrei Savu / andreisavu.ro

On Fri, Jan 20, 2012 at 6:45 PM, Edmar Ferreira <
[email protected]> wrote:

> Hi,
>
> I want to use whirr to launch a Hadoop cluster on EC2 to run some
> stremming jobs using ruby.
> I am using whirr 0.7.0
> The default ubuntu images that are being used don't have ruby installed.
>
> First I followed this tutorial :
>
> http://whirr.apache.org/docs/0.7.0/quick-start-guide.html
>
> I can start my cluster but my job fails because ruby is not installed.
>
> Following the instructions in the faq:
> http://whirr.apache.org/faq.html
>
> I created a functions directory with the scripts. The FAQ tell me to
> change these scripts to add my own packages.
>
> I changed the update_repo() function inside my functions/install_hadoop.sh
> script addint one line :
>
>  17 function update_repo() {
>  18   if which dpkg &> /dev/null; then
>  19     sudo apt-get update
>  20     sudo apt-get install ruby rubygems
>  21   elif which rpm &> /dev/null; then
>  22     yum update -y yum
>  23   fi
>  24 }
>
>
>
> When I log in the machines ruby is not installed even with this change.
> I tested inserting a echo statement inside this function and checked the
> whirr.log to make sure that it's being called and it is.
> If I log in the machines and run :
>
> sudo apt-get install ruby rubygems
>
> I can install without problems.
>
>
> What can I do to automatically install ruby in my cluster ?
>
>
> Thanks
>
> --
> Edmar Ferreira
> Co-Founder at Everwrite
>
>

Reply via email to