Open Stack

On 10/28/2013 06:17 AM, John Garbutt wrote:
> Is there a reason why you could not just use a Cinder Volume for your
> data, in this case?

Yeah, he was saying that's the long term goal, and an Ironic-targeted
feature.

> While at a first glance, it feels rather wrong, and un-cloudy, I do
> see something useful about refreshing the base disk, and leaving the
> data disks alone. Prehaps it's something that could be described in
> the block device mapping, where you have a "local volume" that you
> choose to be non-ephemeral, except for server terminate, or something
> like that?

Agreed that my first reaction is that it feels wrong.  It's making
ephemeral disks not so ephemeral.  Knowing this isn't the ideal
long-term solution doesn't make me feel better about it.

But, we can err on the side of practical instead of technical purity
here.  I'm OK with that.  A new block device attribute sounds like a
good way to do it, too.  There's a somewhat similar flag
'delete_on_termination' for block device mappings.  It's used to specify
that a cinder volume should be automatically deleted when an instance is
terminated.  We could have 'preserve_on_rebuild' or something that would
just apply to ephemeral disks.

-- 
Russell Bryant