Open Stack

On 15 February 2014 08:42, Dan Prince <dprince at redhat.com> wrote:
>

> Let me restate the options the way I see it:
>
> Option A is we do our job... by making it possible to install OpenStack using various distributions using a set of distro agnostic tools (TripleO).

So our job is to be the upstream installer. If upstream said 'we only
support RHEL and Ubuntu, our job would arguably end there. And in
fact, we're being much more inclusive than many other parts of
upstream - Suse isn't supported upstream, nor is Ubuntu non-LTS, nor
Fedora.

> Option B is we make our job easy by strong arming everyone into the same defaults of our "upstream" choosing.

Does Nova strong arm everyone into using kvm? Its the default. Or
keystone into using the SQL token store - its the default?

No - defaults are not strong arming. But the defaults are obviously
defaults, and inherited by downstreams. And some defaults are larger
than others -  we've got well defined interfaces in OpenStack, which
have the primary characteristic of 'learn once, apply everywhere' -
even though in principle you can replace them. At the low level REST
and message-bus RPCs, at a level up Keystone and more recently Nova
and Neutron have become that as we get higher order code like Heat and
Savanna that depend on them. I hope none would replace Nova with
Eucalyptus and then say they're running OpenStack - in the same way
we're both defining defaults, *and* building interfaces. *That* is our
job - making OpenStack *upstream* deployable, in the places, and on
the platforms, with the options, that our users want.

Further to that, upstream we're making choices with the thoughts of
our *users* in mind - both cloud consumers and cloud operators. They
are why we ask questions like 'is having every install have
potentially different usernames for the nova service a good idea'. The
only answer so far has been 'because distros have chosen different
usernames already and we need to suck it up'. Thats not a particularly
satisfying answer.

> Does option B look appealing? Perhaps at first glance. By taking away the differences it seems like we are making everyone's lives easier by "streamlining" our depoyment codebase. There is this one rub though: it isn't what users expect.

I don't know what users expect: There's an assumption stated in some
of the reponses that people which choose 'TripleO + Packages' do that
for a reason. I think this is likely going to be wrong much of the
time. Why? Because upstream doesn't offer someone to ring when there
is a problem. So people will grab RDO, or Suse's offering, or
Rackspace Private Cloud, or HP Cloud OS, or
$distribution-of-openstack-of-choice : and I don't expect for most
people that 'and we used a nova deb package' vs 'and we used a nova
pip package' is going to be *why* they choose that vendor, so as a
result many people will get TripleO+Packages because their vendor
chose that for them. That places a responsibility on the vendors and
on us. The vendors need to understand the consequences of their
packages varying trivially from upstream - the
every-unix-is-a-little-different death of a thousand cuts problem -
and we need to help vendors understand the drivers that lead them to
need to build images via packages.

> Take /mnt/state for example. This isn't the normal place for things to live. Why not use the read only root mechanism some distributions already have and work with that instead. Or perhaps have /mnt/state as a backup solution which can be used if a mechanism doesn't exist or is faulty?

Currently we have two options for upgrading images. A) /mnt/state, B)
a SAN + cinder. We haven't tested B), and I expect for many installs B
won't be an option. /mnt/state is 100% technical, as no other options
exist - none of the Linux distro 'read only root' answers today answer
the problem /mnt/state solves in a way compatible with Nova.

> In the end I think option A is the way we have to go. Is it more work... maybe. But in the end users will like us for it. And there is always the case that by not reimplementing some of the tools and mechanisms which already exist in distros that this ends up being less work anyways. I do hope so...

Certainly we're trying very hard to keep things we reimplement minimal
and easily swap-outable (like o-r-c which I expect some deployments
will want to replace with chef/puppet).

> As for the reference implementation part of A I do hope we choose it wisely and that as much as possible the distros take note of our choices.

+1

-Rob


-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud