[one-users] VM Provisioning, opennebula and virt-install/kickstart

[Javier Fontan]

Hello,

Maybe a way to do it is preparing base images with base REHL installed
and scripts that autoinstall packages and configures your machine. One
way of passing parameters is CONTEXT parameter in VM description.
There you can put information on what to install or maybe point to an
URL with chef recipes/configurations to do the job for you.

Does it make sense?

Bye


On Thu, Mar 18, 2010 at 3:16 AM, Mick Pollard <lists at lunix.com.au> wrote:
> G`day,
>
> On Wed, 17 Mar 2010 19:00:32 -0500
> Jack Jill <jacknjill111 at gmail.com> wrote:
>
>> Mick,
>>
>> <quote>
>> I am wanting to build a 'self serve' VM provisioning tool so my devs can
>> pick a client, have that clients environment deployed to a VM for
>> dev/testing/staging purposes and it be exactly the same as the clients
>> current production environment.
>> </quote>
>>
>> When you say you want to replicate your client's production environment,
>> what exactly are you attempting to replicate? Is it the application software
>> such as Ruby, Tomcat, etc. Or even lower level specific stuff such as a
>> specific version of openSSH server, etc.
>>
>> --
>> Thanks,
>> Rosh
>> http://blogs.plexibus.com
>>
> I have a yum repos that all my servers point to instead of the RHEL ones. This way I can test updates outside of this environment before adding them to my yum repos.
> I the use virt-install and kickstart to install a new VM.
>
>    virt-install --accelerate -n cust01-test -r 256 --vcpus=1 --disk pool=lvm,bus=virtio,size=10 --vnc --os-type linux --os-variant=rhel5 \
>    --network=network:default -l http://192.168.122.1/os/CentOS/5.4/os/x86_64/ -x "ks=http://192.168.122.1/ks/cust01.cfg"
>
> The kickstart does a very minimal install and sets the hostname and then installs the puppet client.
> Puppet then takes care of building out the environment to match the production environment of the customer.( selects webserver etc )
> Currently I monitor this and let the devs know when its ready. I would like to eventually add cucumber-nagios in there or the like to alert devs without me be involved.
> The typical lifecycle for a VM for us is measured in days and not months. deploy, modify customers software, test, then push to production and destroy the VM.
> I understand opennebula uses images for a reason and it makes sense for the most part (it's distribution agnostic) but it would be nice to have the choice to not always use images.
> It would be nice if I could move from running virt-install/virsh and having wiki pages to remember which client is on what node to making use of opennebula.
>
> I am only new to ruby and not quite comfortable at writing something myself although I have looked at the ruby-libvirt bindings and they seem to lack the whole virt-install command set.
>
>
>
> --
> Regards
> Mick Pollard ( lunix )
> ------------------------------------------------
> BOFH Excuse of the day:
> Resignalled Integrity NMI Warning
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>



-- 
Javier Fontan, Grid & Virtualization Technology Engineer/Researcher
DSA Research Group: http://dsa-research.org
Globus GridWay Metascheduler: http://www.GridWay.org
OpenNebula Virtual Infrastructure Engine: http://www.OpenNebula.org