[PKG-Openstack-devel] Status?

[Thomas Goirand]

On 07/15/2017 12:11 AM, Turbo Fredriksson wrote:
> On 14 Jul 2017, at 21:08, Thomas Goirand <zigo at debian.org> wrote:
> 
>> The point of using Debian live is:
> 
> Yes, but you still need to install OS in it. And configure it..
> 
>> Oh! So you have written a layer on top of the Debian installer that
>> installs OpenStack thanks to preseed?
> 
> Yes I did, because yours was shit. I’m sure it works fine in a very simple,
> test-me setup

It was its primary goal: testing. That it didn't fit a broader scope, I
am well aware of it, but I had no resources to do more.

>> even though my Debian live method maybe has better performance for
>> tempest test provisioning).
> 
> Yeah, but your testing isn’t good enough. It might test individual parts
> very well, but not in a “complete” setup. Which is where I’m guessing
> most users will end up sooner or later.
> 
> Your script(s) also doesn’t setup everything and doesn’t allow them
> to work together in the same cluster. I took one look at that setup
> and deemed it shit in a few minutes...
> 
> I’ve found numerous bugs and issues with your packages, stuff you’re
> not interested in dealing with.

I do agree there was issues, but I don't agree I was (or I am) not
interested dealing with it and fixing things. But for this, I was
expecting people like you to communicate and contribute more.

> My suspicion is that it didn’t fit into your
> (or your employers) “agenda”

My employer used puppet, and therefore, didn't care having the
configuration (or even config files) right. Debian is the only
distribution that cared for correctness of configuration files. I know
it well because I was the only one reporting configuration file
generation bugs to upstream, while Ubuntu and Red Hat people didn't even
care for them.

> You made a good start and the whole Debian GNU/Linux community
> is better for it, but I find your work a little short-sighted.

Could you expand on this, and explain what you would like to see improved?

>> The initial install of Debian doesn't even exist in the case of Debian
>> live, it just boots in 1 minute 30 seconds. Then installing OpenStack
>> takes about 20 minutes from the local repository.
> 
> That’s what I mean, you’re shortsighted and narrow-minded. You’re not
> installing and setting up a whole cluster with all services in twenty minutes!
> 
>> Yeah, I agree. Though it has been a resource issue on my side, otherwise
>> I would probably used a 2 nodes setup.
> 
> Fair enough, but I’m sure resources is a-plenty if one only asks.

Unfortunately, this was not the case. During a year at Mirantis, I was
using my own Xen VM to do the testing, and it took over a year to ask
for 2 servers (one for Jenkins, one for tempest testing). And the result
right now is that we have nothing as these servers have been taken away
by my ex-employer.

> And
> considering that your former employer did sponsor your OS work, finding
> some resources shouldn't have been that hard to find.

It unfortunately was.

>> That's probably because you're not setting-up the number of thread,
> 
> Because this is a production environment, I’m more concerned with
> performance than memory. So I’ve bumped up threads, memory usage
> etc, etc on all services to reflect this.

Right. For testing scope, we don't really care this.

>> you end up with really too many daemons, which takes too much
>> RAM for no reason (at least no reason for such tempest validation).
> 
> This is where I STRONGLY disagrees with your previous work. Testing
> a few services individually isn’t any where near enough what’s needed.

Have you looked into what tempest does? It's all but "testing individual
service". There's full and complex scenarios that get everything involved.

> It IS needed, but an all-in setup needs to be tested as well, to iron out all
> those communication and interoperability issues. Which there’s plenty
> of unfortunately :(.

Could you explain more here?

> Now, as I’ve said all along, your work on the packages have been invauable
> and I really WILL miss you on this!!

Thanks.

> I’m not sure that’s what’s happening here, but you left. Voluntarily.

What happened is that I have been searching for the last 6 months
someone to sponsor my Debian packaging work, so I can continue to be
full time on it. So far, I haven't found a company that is willing to do
so. Therefore, I am stepping down, because I don't believe I can do a
good enough work if I am just doing this on my week-ends. So you can't
really say it's "voluntarily". I really would like to continue, but I
refuse to be taken as responsible for a lower quality of packaging.

> But we’re *ALL* on the same page regarding testing. I want much, MUCH
> more and I want it NOW (before anything is uploaded to the repos)!
> Orlov want it later, you seems to be happy the way it is.

I would also prefer *more* testing, especially in a multi-node setup,
and if possible improving openstack-deploy to handle it, possibly with
HA setup and all, if we can.

So, again, please share your setup scripts. Please do contribute to
openstack-deploy, so we can improve it. Just complaining "your stuff is
shit" and keeping your scripts private wont help. One thing which I
though we could do was even contributing to the Debian installer, and
have tasks to install OpenStack. One task for a controller, and one for
a compute node. Your thoughts?

Cheers,

Thomas Goirand (zigo)