Fwd: How to bootstrap canonistack/openstack with trunk now?

[Andreas Hasenack]

Sorry, this was meant for the list.

---------- Forwarded message ----------
From: Andreas Hasenack <andreas at canonical.com>
Date: Tue, May 28, 2013 at 9:36 AM
Subject: Re: How to bootstrap canonistack/openstack with trunk now?
To: Ian Booth <ian.booth at canonical.com>


On Mon, May 27, 2013 at 9:03 PM, Ian Booth <ian.booth at canonical.com> wrote:

> Hi
>
> I am guessing you are running from source. For Canonistack, everything has
> been
>

Yes, otherwise many charms won't deploy ;)


> set up to "just work", but sadly as of right now only for region lcy01. ie
> the
> simplestreams metadata is automatically published to the shared Canonistack
> public bucket and so long as the public bucket url is correctly set, the
>

Where can I find the list of the two public buckets juju uses for
canonistack?

May I suggest that the "public-bucket-url" keyword be added to the output
of juju init -w in the openstack section? Maybe commented, and with a text
telling you to ask your cloud admin for this value, or else you will have
to setup simplestreams and use --upload-tools (and build from source).

I'm thinking about the random private cloud user, see more below.


> metadata should be found. NB there's a ticket raised to (for Canonistack)
> get
> the shared public bucket url published in keystone so real soon now you
> won't
> need to manually specify that anymore.
>
> With region lcy02, I believe there are issues gathering the image
> metadata, but
> could be mis-remembering that. I can't recall the reason but it's being
> worked
> on. Scott (who I've cc'ed) can fill in the gaps hopefully.
>
> There is a manual work around. You can generate your own image metadata
> and by
> putting that in your "local" public bucket, it will be found. I sent an
> email to
>

I wouldn't call it "work around", as it's necessary for all the users of
private clouds out there. Once I get this working, I'll update the internal
wiki.

So there are two "magic" things happening in the background that the user
has no immediate visibility to:
- tools download from the public bucket the juju team has to setup and
prepare in all public clouds
- simplestreams, also setup in the public bucket that the juju team has to
setup and prepare in all public clouds

If any of the above have issues, juju bootstrap fails. And it's absolutely
critical for users of private clouds. The instructions for these users are
far off from the simple "juju init -w". There is a setup step involved in
the cloud itself, right? Setting up that public bucket. Is there a script
for that? I suppose at some point you use "juju sync", but something has to
setup that first bucket so it can be copied from.

So basically there are two sets of instructions we need: for juju users
(the juju init -w bit, and add a public-bucket-url key if it's a private
cloud) and for their #is team, about setting up the bucket for tools (and a
development environment) and simplestreams.





> the list last week but forgot about lcy02 issue sorry. Here's what you
> need to
> do. You run a command to generate the image metadata files:
>
> juju image-metadata -i <your image id> -r lcy02 -e
> https://keystone.canonistack.canonical.com:443/v2.0/
>
> The arch defaults to amd64, and the series to precise. See juju help
> image-metadata.
>
> There are 2 json files generated to your .juju directory. As per the
> instructions when the above command is run, these 2 files need to be
> copied to
> the "streams/v1" path in your public bucket (you can use swift upload).
> Then it
> will start working.
>
>
Thanks for these instructions, I'll try them out and document them in the
internal wiki at first.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju-dev/attachments/20130528/0141e6ae/attachment.html>