Hello Alberto,
I would advise you to deploy as much as node as you can, and dedicate separate services. That would bring you the possibility to extend your knowledge on HA mechanisms, and really understand how services interact each other.
Let's say you spawn an instance from the api server, while it's is launched on a compute node - which is one another server. By having splitted architecture, you won't miss the obvious ;-)
What is the purpose of the lab ?

Nuage & Co - Razique Mahroua 

Le 26 mars 2012 à 12:56, Alberto Molina Coballes a écrit :

Hi all,

I'm a teacher at a sysadmin academy and we are planning to deploy a
private cloud with OpenStack for educational/training purposes. We are
currently selecting the optimal hardware configuration for our needs
and I'm asking for some help to the list.

Using [1] as a reference, we are clear about hardware specifications
of compute nodes (4 nodes with 2 processors), but we aren't about the
best solution for node controller and optimal storage:

Option 1:

1 server with 2 HD in RAID1 (nova controller, nova volume, glance,
...) + 1 SAN iSCSI

Option 2:

1 server with 12 HD in RAID5/6 (nova controller, nova volume, glance, ...)

We are excited about the possibilities that a private cloud brings to
us and we don't want to fail in this critical step, so any help will
be really appreciated.

Alberto Molina Coballes
IES Gonzalo Nazareno

[1] http://www.referencearchitecture.org/hardware-specifications/

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

Reply via email to