Open Stack

On Wed, Mar 12, 2014 at 8:07 PM, Chris Jones <cmsj at tenshu.net> wrote:

>
> Hey
>
> I wanted to throw out an idea that came to me while I was working on
> diagnosing some hardware issues in the Tripleo CD rack at the sprint last
> week.
>
> Specifically, if a particular node has been dropped from automatic
> scheduling by the operator, I think it would be super useful to be able to
> still manually schedule the node. Examples might be that someone is
> diagnosing a hardware issue and wants to boot an image that has all their
> favourite diagnostic tools in it, or they might be booting an image they
> use for updating firmwares, etc (frankly, just being able to boot a
> generic, unmodified host OS on a node can be super useful if you're trying
> to crash cart the machine for something hardware related).
>
> Any thoughts? :)
>

+1 I like the idea and think it's quite useful.

Drivers in Ironic already expose a rescue interface[1] (which I don't think
we had put much thoughts into it yet) perhaps the PXE driver might
implement something similar to what you want to do here?

[1]
https://github.com/openstack/ironic/blob/master/ironic/drivers/base.py#L60
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140314/e53eb4f0/attachment.html>