[openstack-dev] [all] Dependency management summit sessions

Sean Dague sean at dague.net
Tue Apr 28 10:57:32 UTC 2015

On 04/27/2015 05:53 PM, Robert Collins wrote:
> Hi, so  -
> https://libertydesignsummit.sched.org/event/4da45ec1390dadcfc1d8a73decbf3f19#.VT6urd_va00
> Is an ops track session about dependencies - focusing on the
> operational side (mysql, mongo, rabbit etc). I'd love to have some
> developer centric folk in the room, though I'll be doing my best to
> capture the ops constraints. My hope is to have a really clear story
> about what we can and should depend on, and the cost of non-boring
> tech, for our operators.
> On 16 April 2015 at 22:32, Sean Dague <sean at dague.net> wrote:
> ...
>> Possibly, the devil is in the details. Also it means it won't play nice
>> with manual pip installs before / after, which are sometimes needed.
>> Mostly I find it annoying that pip has no global view, so all tools that
>> call it have to construct their own global view and only ever call pip
>> once to get a right answer. It feels extremely fragile. It's also not
>> clear to me how easy it's going to be to debug.
> We don't need to do that. I've put some expanded thoughts together here:
> https://rbtcollins.wordpress.com/2015/04/28/dealing-with-deps-in-openstack/
> I *think* that avoids the omg refactor-devstack thing, at the cost of
> a small additional feature in pip.
> Thierry says we're going to slot this into
> https://libertydesignsummit.sched.org/event/da0f31eddd0def88c6c51fb131fe87bd#.VT6v1N_va00
> or the followup after lunch - I'd like to pin it down more than that,
> if we can.

I'm still generally suspicious of the precompute / install model because
solving that ends up being... interesting some times. I also think there
is a related issue of "dependencies for optional features" which,
because they are inconsistently dealt with, exacerbate things. This
being things like drivers, db backends.

After the giant unwind Doug, Clark, and I started writing up the
following - https://etherpad.openstack.org/p/requirements-future

I do think we need a summit discussion, I also think pip needs some
fixes, but I think this needs a lot of heads together to get to a plan,
because many individuals thought they nailed this issue in the past, and
were wrong.


Sean Dague

More information about the OpenStack-dev mailing list