ticket#2673 - nitrate deployment

[James Laska]

Greetings,

I am in search of sponsorship for hosting space to deploy a test
instance of nitrate [1].  Some of you may recall several years back when
we deployed an instance of Testopia.  That exploration was cut short
when we discovered license incompatibilities between testopia and
Fedora.  At that time, a small group of Red Hat associates began
creating a new Django-based test management system to remove the
components that had conflicting licenses.  That new system became
nitrate [1].

The Fedora QA team would like to experiment/explore this newer
Django-based test management system.  At the very least, I believe we'll
need a publicly accessible server where we can host the nitrate instance
and a mysql database for it.

I filed ticket#2673
(https://fedorahosted.org/fedora-infrastructure/ticket/2673) to track
this request.  Please don't hesitate with any
questions/comments/concerns.

Thanks,
James

[1] https://fedorahosted.org/nitrate/

[James Laska]

On Tue, 2011-03-15 at 16:54 -0400, James Laska wrote:
> Greetings,
>
> I am in search of sponsorship for hosting space to deploy a test
> instance of nitrate [1].  Some of you may recall several years back when
> we deployed an instance of Testopia.  That exploration was cut short
> when we discovered license incompatibilities between testopia and
> Fedora.  At that time, a small group of Red Hat associates began
> creating a new Django-based test management system to remove the
> components that had conflicting licenses.  That new system became
> nitrate [1].
>
> The Fedora QA team would like to experiment/explore this newer
> Django-based test management system.  At the very least, I believe we'll
> need a publicly accessible server where we can host the nitrate instance
> and a mysql database for it.
>
> I filed ticket#2673
> (https://fedorahosted.org/fedora-infrastructure/ticket/2673) to track
> this request.  Please don't hesitate with any
> questions/comments/concerns.

Hurry (rhe), Danqing (dli) and myself will be able to assist with the
draft RFR guidelines proposal at noted in ticket #2674.  This
includes ...

1. Applying rpm updates to the service (and to any underlying
        pieces of the application stack) if necessary.
     2. Applying hotfixes via the puppet hotfix module if there's a
        securiry fix or bugfix that needs to go in and it's not worth
        spinning a new rpm.
     3. Keeping up with upstream development
     4. Answering questions about whether a yum update (to your app or
        to the underlying stack) might break your app.
     5. Fixing things should an app start throwing errors in production
        for unknown reasons
     6. Work on deployment problems

Don't hesitate if there are any other concerns or questions on this
pilot and RFR request.

Thanks,
James

[1] https://fedorahosted.org/fedora-infrastructure/ticket/2674

[He Rui]

Greetings,

I haven't received any update on this project since the ticket filed, is
it under progressing or is there anything we need help with? We planed
to set it up this month or next, so please keep an eye on it.

Many Thanks,
Hurry

On Mon, 2011-03-28 at 15:14 -0400, James Laska wrote:
> On Tue, 2011-03-15 at 16:54 -0400, James Laska wrote:
> > Greetings,
> >
> > I am in search of sponsorship for hosting space to deploy a test
> > instance of nitrate [1].  Some of you may recall several years back when
> > we deployed an instance of Testopia.  That exploration was cut short
> > when we discovered license incompatibilities between testopia and
> > Fedora.  At that time, a small group of Red Hat associates began
> > creating a new Django-based test management system to remove the
> > components that had conflicting licenses.  That new system became
> > nitrate [1].
> >
> > The Fedora QA team would like to experiment/explore this newer
> > Django-based test management system.  At the very least, I believe we'll
> > need a publicly accessible server where we can host the nitrate instance
> > and a mysql database for it.
> >
> > I filed ticket#2673
> > (https://fedorahosted.org/fedora-infrastructure/ticket/2673) to track
> > this request.  Please don't hesitate with any
> > questions/comments/concerns.
>
> Hurry (rhe), Danqing (dli) and myself will be able to assist with the
> draft RFR guidelines proposal at noted in ticket #2674.  This
> includes ...
>  1. Applying rpm updates to the service (and to any underlying
>     pieces of the application stack) if necessary.
>  2. Applying hotfixes via the puppet hotfix module if there's a
>     securiry fix or bugfix that needs to go in and it's not worth
>     spinning a new rpm.
>  3. Keeping up with upstream development
>  4. Answering questions about whether a yum update (to your app or
>     to the underlying stack) might break your app.
>  5. Fixing things should an app start throwing errors in production
>     for unknown reasons
>  6. Work on deployment problems 
>
> Don't hesitate if there are any other concerns or questions on this
> pilot and RFR request.
>
> Thanks,
> James
>
> [1] https://fedorahosted.org/fedora-infrastructure/ticket/2674
> _______________________________________________
> infrastructure mailing list
> [e-mail]
> https://admin.fedoraproject.org/mailman/listinfo/infrastructure

[Kevin Fenzi]

On Thu, 07 Apr 2011 13:36:05 +0800
He Rui  wrote:
> Greetings,
>
> I haven't received any update on this project since the ticket filed,
> is it under progressing or is there anything we need help with? We
> planed to set it up this month or next, so please keep an eye on it.

I could probably help you guys out with this. ;)

I guess the first step is to make sure to get the package into EPEL
(which it looks like it's already under review).

After thats done, we could look at testing it on a publictest instance. 
I'll add this information to the ticket as well, but do ping me when we
are ready to do more planning. ;)
> Many Thanks,
> Hurry

kevin
--
> On Mon, 2011-03-28 at 15:14 -0400, James Laska wrote:
> > On Tue, 2011-03-15 at 16:54 -0400, James Laska wrote:
> > > Greetings,
> > >
> > > I am in search of sponsorship for hosting space to deploy a test
> > > instance of nitrate [1].  Some of you may recall several years
> > > back when we deployed an instance of Testopia.  That exploration
> > > was cut short when we discovered license incompatibilities
> > > between testopia and Fedora.  At that time, a small group of Red
> > > Hat associates began creating a new Django-based test management
> > > system to remove the components that had conflicting licenses.
> > > That new system became nitrate [1].
> > >
> > > The Fedora QA team would like to experiment/explore this newer
> > > Django-based test management system.  At the very least, I
> > > believe we'll need a publicly accessible server where we can host
> > > the nitrate instance and a mysql database for it.
> > >
> > > I filed ticket#2673
> > > (https://fedorahosted.org/fedora-infrastructure/ticket/2673) to
> > > track this request.  Please don't hesitate with any
> > > questions/comments/concerns.
> >
> > Hurry (rhe), Danqing (dli) and myself will be able to assist with
> > the draft RFR guidelines proposal at noted in ticket #2674.  This
> > includes ...
> >  1. Applying rpm updates to the service (and to any underlying
> >     pieces of the application stack) if necessary.
> >  2. Applying hotfixes via the puppet hotfix module if there's a
> >     securiry fix or bugfix that needs to go in and it's not
> >
> > worth spinning a new rpm.
> >      3. Keeping up with upstream development
> >      4. Answering questions about whether a yum update (to your app
> > or to the underlying stack) might break your app.
> >      5. Fixing things should an app start throwing errors in
> > production for unknown reasons
> >      6. Work on deployment problems
> >
> > Don't hesitate if there are any other concerns or questions on this
> > pilot and RFR request.
> >
> > Thanks,
> > James
> >
> > [1] https://fedorahosted.org/fedora-infrastructure/ticket/2674
> > _______________________________________________
> > infrastructure mailing list
> > [e-mail]
> > https://admin.fedoraproject.org/mailman/listinfo/infrastructure

[James Laska]

On Fri, 2011-04-08 at 11:49 -0600, Kevin Fenzi wrote:
> On Thu, 07 Apr 2011 13:36:05 +0800
> He Rui  wrote:
> > Greetings,
> >
> > I haven't received any update on this project since the ticket filed,
> > is it under progressing or is there anything we need help with? We
> > planed to set it up this month or next, so please keep an eye on it.
>
> I could probably help you guys out with this. ;)
>
> I guess the first step is to make sure to get the package into EPEL
> (which it looks like it's already under review).
>
> After thats done, we could look at testing it on a publictest instance. 
> I'll add this information to the ticket as well, but do ping me when we
> are ready to do more planning. ;)

Packaging of nitrate is progressing.  Libraries bundled within nitrate
have been identified and submitted as separate package reviews.  While
the process isn't complete, I'm curious if we can pipeline a bit.  Is it
possible to setup a publictest instance while packaging progresses?  The
intent for this instance wouldn't be for production use ... more to
allow us to explore use cases for Fedora and to identify any other
nitrate features/changes.

While I still think we should wait until all packaging work is completed
before officially deploying and supporting an instance, I'm hoping to
avoid waiting for packaging work to complete, deploying a test instace,
then discovering some other use-case (or upstream functionality)
roadblock.

Any thoughts/concerns with this idea?

Thanks,
James

[Kevin Fenzi]

On Mon, 01 Aug 2011 13:29:03 -0400
James Laska  wrote:
> Packaging of nitrate is progressing.  Libraries bundled within nitrate
> have been identified and submitted as separate package reviews.  While
> the process isn't complete, I'm curious if we can pipeline a bit.  Is
> it possible to setup a publictest instance while packaging
> progresses?  The intent for this instance wouldn't be for production
> use ... more to allow us to explore use cases for Fedora and to
> identify any other nitrate features/changes.

Well, possibly, but you may well have to redo things as the packaging
changes, so it might be more work than it's worth in the end. ;)
> While I still think we should wait until all packaging work is
> completed before officially deploying and supporting an instance, I'm
> hoping to avoid waiting for packaging work to complete, deploying a
> test instace, then discovering some other use-case (or upstream
> functionality) roadblock.
>
> Any thoughts/concerns with this idea?

Well, the other side of the coin is that you install now, get it all
working, then have to just re-install and get it working again when the
packaging is done. Of course there might not be too many changes from
the packaging side that require redoing things. I'm not sure where the
packaging is sitting at this point.

So, I would say:

* How far out does the packaging look? Is it getting close? or still
  much to do?

* How close is whatever you would install today to what the final
  packaging will be? Just the same thing but split out into seperate
  packages?

kevin

[James Laska]

On Tue, 2011-08-02 at 11:11 -0600, Kevin Fenzi wrote:
> On Mon, 01 Aug 2011 13:29:03 -0400
> James Laska  wrote:
> > Packaging of nitrate is progressing.  Libraries bundled within nitrate
> > have been identified and submitted as separate package reviews.  While
> > the process isn't complete, I'm curious if we can pipeline a bit.  Is
> > it possible to setup a publictest instance while packaging
> > progresses?  The intent for this instance wouldn't be for production
> > use ... more to allow us to explore use cases for Fedora and to
> > identify any other nitrate features/changes.
>
> Well, possibly, but you may well have to redo things as the packaging
> changes, so it might be more work than it's worth in the end. ;)
> > While I still think we should wait until all packaging work is
> > completed before officially deploying and supporting an instance, I'm
> > hoping to avoid waiting for packaging work to complete, deploying a
> > test instace, then discovering some other use-case (or upstream
> > functionality) roadblock.
> >
> > Any thoughts/concerns with this idea?
>
> Well, the other side of the coin is that you install now, get it all
> working, then have to just re-install and get it working again when the
> packaging is done. Of course there might not be too many changes from
> the packaging side that require redoing things. I'm not sure where the
> packaging is sitting at this point.

I suspect it will be as you say, some re-{install,config} would be
needed as the package progresses through review.  I'm accustomed to that
workflow already, so I wouldn't be upset about those changes.  Also, am
I correct in thinking that this publictest instance is just for testing
purposes.  Anything considered production would need to be handled
differently, or in a separate request?
> So, I would say:
> How far out does the packaging look? Is it getting close? or still
> much to do?

It'd say it's progressing, I don't know if I'd be comfortable giving an
accurate percentage complete.  I'll defer to rhe (cc'd) for more up2date
guidance there.  As I understand it, upstream is involved and actively
responding to issues raised during review.  The current challenge
involves identifying bundled libraries, and moving them off into
separate package reviews.
> How close is whatever you would install today to what the final
> packaging will be? Just the same thing but split out into seperate
> packages?

I *think* so.  There will likely be a new upstream release (or two)
while this happens.  But I don't expect a tremendous amount of
package-related churn for that.

Thanks,
James

[Kevin Fenzi]

On Tue, 09 Aug 2011 16:00:35 -0400
James Laska  wrote:
> I suspect it will be as you say, some re-{install,config} would be
> needed as the package progresses through review.  I'm accustomed to
> that workflow already, so I wouldn't be upset about those changes.
> Also, am I correct in thinking that this publictest instance is just
> for testing purposes.  Anything considered production would need to
> be handled differently, or in a separate request?

Correct. The plan is basically:

https://fedoraproject.org/wiki/Infrastructure/RFR
and the workflow is: 
https://fedoraproject.org/wiki/Request_for_resources_SOP

So, things start with a dev instance. In this nothing is saved, ideas
and setup are tried out, the packages are tested, process for
installing/managing is written up, thoughts are gathered about caching
and possible setup moving forward, a sysadmin-resourcename is setup
with all responsible folks working on the resource (which allows us to
know who is working on it and easily contact them), etc.
> > So, I would say:
> > How far out does the packaging look? Is it getting close? or still
> > much to do?
>
> It'd say it's progressing, I don't know if I'd be comfortable giving
> an accurate percentage complete.  I'll defer to rhe (cc'd) for more
> up2date guidance there.  As I understand it, upstream is involved and
> actively responding to issues raised during review.  The current
> challenge involves identifying bundled libraries, and moving them off
> into separate package reviews.
> > How close is whatever you would install today to what the final
> > packaging will be? Just the same thing but split out into seperate
> > packages?
>
> I *think* so.  There will likely be a new upstream release (or two)
> while this happens.  But I don't expect a tremendous amount of
> package-related churn for that.

ok.

So, I guess if you guys are ready for a dev instance we could set one
up and get the ball rolling? Or wait until packaging is done to reduce
the amount of churn/changes. ;) Up to you.

kevin