[openstack-dev] Deprecation of in tree EC2 API in Nova for Kilo release

Sean Dague sean at dague.net
Mon Feb 2 16:29:30 UTC 2015


On 02/02/2015 10:55 AM, Daniel P. Berrange wrote:
> On Mon, Feb 02, 2015 at 07:44:24AM -0800, Dan Smith wrote:
>>> I'm with Daniel on that one. We shouldn't "deprecate" until we are 100%
>>> sure that the replacement is up to the task and that strategy is solid.
>>
>> My problem with this is: If there wasn't a stackforge project, what
>> would we do? Nova's in-tree EC2 support has been rotting for years now,
>> and despite several rallies for developers, no real progress has been
>> made to rescue it. I don't think that it's reasonable to say that if
>> there wasn't a stackforge project we'd just have to suck it up and
>> magically produce the developers to work on EC2; it's clear that's not
>> going to happen.
> 
> I think that is exactly what we'd would have todo. We exist as a project
> to serve the needs of our users and it seems pretty clear from the survey
> results that users are deploying the EC2 impl in significant numbers,
> so to just remove it would essentially be ignoring what our users want
> from the project. If we're saying it is reasonable to ignore what our
> users want, then this project is frankly doomed.
> 
>> Thus, it seems to me that we need to communicate that our EC2 support is
>> going away. Hopefully the stackforge project will be at a point to
>> support users that want to keep the functionality. However, the fate of
>> our in-tree support seems clear regardless of how that turns out.
> 
> If the external EC2 support doesn't work out for whatever reason, then
> I don't think the fate of the in-tree support is at all clear. I think
> it would have a very strong case for continuing to exist.

It's really easy to say "someone should do this", but the problem is
that none of the core team is interested, neither is anyone else. Most
of the people that once were interested have left being active in OpenStack.

EC2 compatibility does not appear to be part of the long term strategy
for the project, hasn't been in a while (looking at the level of
maintenance here). Ok, we should signal that so that new and existing
users that believe that is a core supported feature realize it's not.

The fact that there is some plan to exist out of tree is a bonus,
however the fact that this is not a first class feature in Nova really
does need to be signaled. It hasn't been.

Maybe deprecation is the wrong tool for that, and marking EC2 as
experimental and non supported in the log message is more appropriate.

	-Sean

-- 
Sean Dague
http://dague.net



More information about the OpenStack-dev mailing list