[Openstack] Metadata and File Injection

Soren Hansen soren at linux2go.dk
Mon Jan 2 19:59:04 UTC 2012


2012/1/2 Ewan Mellor <Ewan.Mellor at eu.citrix.com>:
>> I disagree. I think it's very important indeed whether the reasons are
>> good or stupid. If they're good, they'll probably apply to others as
>> well, so we need to know about them, understand them, and take
>> appropriate action. If the reasons are stupid, whatever solution we
>> come up with might be deemed forbidden as well for similarly stupid
>> reasons. The only winning move is not to play.
> I didn't mean that we shouldn't try to understand the reasons.  I
> meant that it doesn't matter whether they are good reasons or bad ones
> because either way the policy is there, and we're not going to change
> it.  If you want to sell to that customer then you have to work with
> their environment.

I understand. My point is just that if we don't understand (or are even
aware of) the reasons there's no way for us to know that whatever
solution we come up with won't trigger the same (silly or reasonable)
rules.

There's no way I want to waste time (others are free to do with their
time whatever they please, of course) playing some silly game where a
company has decided (for no apparent, good reason) to disallow DHCP, and
we come up with some amputated, under-engineered,
not-tested-in-the-real-world-at-all alternative that just happens to not
be called DHCP, and then watch them disallow that, because they don't
like the colour of the box it came in, and then we scramble to develop
yet another functionally equivalent, but differently named dynamic host
configuration protocol, etc. etc. etc.  The only way to win that game is
to refuse to play.

If there are appropriately articulated, sound reasons to avoid DHCP,
let's make them heard so that we can address them. "We don't allow DHCP.
Period." isn't even a reason, and if that were all we had to go on, we
could just call it something else, but let it accidentally be compatible
with DHCP. The magic of ridiculous rules is that you can come up with
equally ridiculous workarounds.

> In the context of this discussion, that means that OpenStack needs to
> work in DHCP-free environments, because we already know of many.

The only one I'm familiar with is Rackspace where I think (and please do
correct me if I'm wrong) DHCP isn't used only because Windows instances
can't be configured with DHCP if they have multiple interfaces.

-- 
Soren Hansen        | http://linux2go.dk/
Ubuntu Developer    | http://www.ubuntu.com/
OpenStack Developer | http://www.openstack.org/




More information about the Openstack mailing list