Open Stack

> -----Original Message-----
> From: Clayton Coleman [mailto:ccoleman at redhat.com]
> Sent: Thursday, February 13, 2014 4:34 PM
> To: OpenStack Development Mailing List (not for usage questions)
> Subject: Re: [openstack-dev] [Solum] Question about Zuul's role in Solum
> 
> 
> 
> ----- Original Message -----
> > From: Julien Vey < vey.julien at gmail.com >
> > Reply-To: "OpenStack Development Mailing List (not for usage
> > questions)" < openstack-dev at lists.openstack.org >
> > Date: Thursday, February 13, 2014 7:18 AM
> > To: "OpenStack Development Mailing List (not for usage questions)" <
> > openstack-dev at lists.openstack.org >
> > Subject: Re: [openstack-dev] [Solum] Question about Zuul's role in
> > Solum
> >
> >
> >
> >
> > Hi,
> >
> > I have some concerns about using Zuul in Solum
> >
> >
> >
> >
> > I agree gating is a great feature but it is not useful for every
> > project and as Adrian said, not understood by everyone.
> > I think many Solum users, and PaaS users in general, are
> > single-project/single-build/simple git worklow and do not care about
> gating.
> >
> > Agreed, this is my major concern, I also think majority of users will
> > be after quick app creation from a single git repository. For those
> > people having to jump through any additional hoops around gating will
> > give them a barrier that will cause them to evaluate if this is the
> > right solution for them.
> >
> >
> >
> >
> >
> > I see 2 drawbacks with Zuul :
> > - Tenant Isolation : How do we allow access on zuul (and jenkins) for
> > a specific tenant in isolation to the others tenants using Solum.
> >
> > Tenant isolation is a big problem for service providers, but maybe not
> > for enterprises running their own Openstack/Solum.
> >
> >
> >
> >
> > - Build customization : One of the biggest advantage of Jenkins is its
> > ecosystem and the many build customization it offers. Using zuul will
> > prohibit this.
> >
> >
> > Agreed. If we make CI/CD pluggable and provide two reference
> > implementations showing both use cases ( which should be achievable with
> M1 workflow ).
> >
> > * `if it builds its good` and provide example code repos that use
> > travisCI we can show external tooling CI/CD workflow. [1]
> > * full code gating functionality of Gerrit->Zuul->Solum. I see this as
> > external to the core of Solum with the user's source control being the
> > middleware, but could be something that solum could assist with
> > building ( provide heat template ). [2]
> >
> > [1] attached Solum-M1.png
> > [2] attached Solum-Gerrit.png
> >
> >
> >
> >
> >
> > About Gerrit, I think it is also a little too much. Many users have
> > their own reviewing system, Pull requests with github, bitbucket or
> > stash, their own instance of gerrit, or even a custom git workflow.
> > Gerrit would be a great feature for future versions of Solum. but only
> > as an optionnal one, we should not force people into it.
> >
> > Agreed. The M1 workflow should make it easy to provide examples of
> > plugging in common external CI/CD tooling, or running your own
> Gerrit+Zuul.
> >
> > Using a gating system like Gerrit/Zuul could make for some very
> > interesting advanced testing, for example one of the tests run could
> > be to hit the solum API and deploy the application to it and provide
> > the URL to reviewers, then when the code is merged it can kill that
> > test. But this is just as doable as an pluggable integration as it is tightly
> coupling Zuul to solum.
> 
> I think it's worth noting that the things Zuul can do (outside of gating or being
> dependent on Gerrit) is allow a reasonable extensible workflow solution for
> source -> outcome flows.  We should be careful to separate the value zuul
> brings as part of a Solum/OpenStack deployment (not inventing an
> extensible workflow story for source) out from gating.
> 
> I agree with the statement that we should not tightly couple Zuul to Solum.
> The M1 workflow does not necessarily need to require zuul if the right
> interaction points (which Zuul would require anyway) are exposed for a
> much simpler solution.  However, it would be a mistake to build a simple
> service now and then to continue to expand it until it looks a lot like Zuul
> without making effort to improve Zuul's multi tenancy, customizability, and
> extensibility stories.  There are certainly pieces of Zuul that don't apply to
> most Solum workflows - it's just that we have to resist the temptation to
> create a standalone solution for source -> outcome flows outside of Zuul.
> 
> Simple, practical, example solution for M1 without gating?  Good.
> Building a zuul equivalent?  Bad.
> Demonstrating a simple integration with Solum build flows that proves out
> Solum endpoints for receiving updates about new deployment units being
> available?  Good.
> 
> >
> > This also gives service providers a good way to productize Solum to
> > different ( or learning ) audiences. Assuming a service provider
> > creates a product based off Solum called 'Bananas' they could offer
> >
> > * 'Banana Lite' which will simply deploy an app ( single instance
> > free, pay for more ) from a git-push/git-pull. Only gating is 'did the build
> succeed'
> > * 'Banana Pro' Adds basic CI/CD flow ( Solum calls into a test
> > framework as part of the Build )
> > * 'Banana Enterprise' Provides full code review, git hosting,
> > automated testing (via Gerrit+github Enterprise), deploy to test instance,
> deploy App.



I like the characterizations above.

This is my understanding of the scope and order of priority for Solum, as outlined in the high level roadmap @ : https://wiki.openstack.org/wiki/Solum/HighLevelRoadmap
1. Priority 1: implement Banana Lite (scoped into Solum M1 milestone) - simple git based deploy with build success as the "gating"
2. Priority 2: implement Banana Pro (scoped into the M3 milestone) - introduce CI/CD by introducing test gating.

The question: is Zuul the right tool to implement the above items in the current Solum roadmap? I think we should focus the discussion on this FIRST.

Banana Enterprise (i.e. the Code gating), does sound like useful, but needs more validation, and certainly not part of our current roadmap. We should pursue Code gating as a separate topic, and focus this thread on choosing the right tool for items currently on the roadmap.

Thanks,
Roshan



> >
> > Julien
> >
> > 2014-02-13 5:47 GMT+01:00 Clark Boylan < clark.boylan at gmail.com > :
> >
> >
> >
> > On Wed, Feb 12, 2014 at 7:25 PM, Noorul Islam K M < noorul at noorul.com
> > >
> > wrote:
> > > "devdatta kulkarni" < devdatta.kulkarni at rackspace.com > writes:
> > >
> > >> Hi,
> > >>
> > >> I have been looking at Zuul for last few days and had a question
> > >> about its intended role within Solum.
> > >>
> > >> From what I understand, Zuul is a code gating system.
> > >>
> > >> I have been wondering if code gating is something we are
> > >> considering as a feature to be provided in Solum? If yes, then Zuul
> > >> is a perfect fit.
> > >> But if not, then we should discuss what benefits do we gain by
> > >> using Zuul as an integral part of Solum.
> > >>
> > >> It feels to me that right now we are treating Zuul as a conduit for
> > >> triggering job(s) that would do the following:
> > >> - clone/download source
> > >> - run tests
> > >> - create a deployment unit (DU) if tests pass
> > >> - upload DU to glance
> > >> - trigger the DU deployment workflow
> > >>
> > >> In the language-pack working group we have talked about being able
> > >> to do CI on the submitted code and building the DUs only after tests
> pass.
> > >> Now, there is a distinction between doing CI on merged code vs.
> > >> doing it before code is permanently merged to master/stable branches.
> > >> The latter is what a 'code gating' system does, and Zuul is a
> > >> perfect fit for this.
> > >> For the former though, using a code gating system is not be needed.
> > >> We can achieve the former with an API endpoint, a queue, and a
> > >> mechanism to trigger job(s) that perform above mentioned steps.
> > >>
> > >> I guess it comes down to Solum's vision. If the vision includes
> > >> supporting, among other things, code gating to ensure that
> > >> Solum-managed code is never broken, then Zuul is a perfect fit.
> > >> Of course, in that situation we would want to ensure that the
> > >> gating functionality is pluggable so that operators can have a
> > >> choice of whether to use Zuul or something else.
> > >> But if the vision is to be that part of an overall application
> > >> lifecycle management flow which deals with creation and scaling of
> > >> DUs/plans/assemblies but not necessarily be a code gate, then we
> > >> should re-evaluate Zuul's role as an integral part of Solum.
> > >>
> > >> Thoughts?
> > >>
> > >
> > > Is Zuul tightly couple with launchpad? I see that most of the
> > > information that it displays is coming from launchpad.
> > >
> > > If it is, is it a good idea to force launchpad on users?
> > >
> > > Regards,
> > > Noorul
> > >
> > > _______________________________________________
> > > OpenStack-dev mailing list
> > > OpenStack-dev at lists.openstack.org
> > > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> >
> > I can't think of any places that Zuul requires launchpad (or displays
> > launchpad info for that matter). It is a bit coupled to Gerrit on one
> > end and Gearman on the other, but not in an extreme way (the use of
> > Gearman makes a bunch of sense imo, but having additional triggers
> > instead of just Gerrit sounds great to me).
> >
> > Clark