[openstack-dev] [db][all] (Proposal) Restorable & Delayed deletion of OS Resources

Jesse Pretorius jesse.pretorius at gmail.com
Mon Mar 17 23:23:56 UTC 2014


On 17 March 2014 20:54, Jay Pipes <jaypipes at gmail.com> wrote:

> After hearing from Tim Bell and others, I think that a two-pronged
> approach is most useful. First, the above-mentioned UI changes to
> prevent common mistakes, and second, using a consistent, standardized
> way of undoing certain operations (Boris' proposal).
>

I'm very much in favor of this proposal.

The primary defense against end-user mistakes should be within Horizon -
where currently the 'Reboot' and 'Terminate' buttons for an instance are
perilously close to each other, among other issues. There are already UI
improvement activities in play to help with this, but the UX team could
probably do with more input. The CLI, in my view, is less prone to 'finger
trouble' and is perhaps less worthy of attention in this regard.

A secondary defense is some sort of systematic protection against actions
like deleting an image from Glance when it's still used by an instance,
thus preventing actions like an instance migration. In a situation like
this, soft deletion isn't exactly the right answer - but instead perhaps
the right method is to remove the image from general visibility until such
time that the instance(s) using it are no longer around. Then the image is
actually removed from the registry.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140318/ad6a15b5/attachment.html>


More information about the OpenStack-dev mailing list