Focus for the next three weeks

John Arbash Meinel john at arbash-meinel.com
Tue Sep 10 14:43:04 UTC 2013


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2013-09-10 7:39, Tim Penhey wrote:
...

> 
> 1. MAAS Tag support
> 
> This seems to be a key need for a number of engagements we have,
> and also a reason why some MAAS/Juju demos are still using pyJuju.
> We have talked about this a lot, but I haven't seen any other
> movement or progress on it.
> 
> What is the current state of play with the provider or cloud
> specific constraints?  Does anyone see real blockers to having MAAS
> tag deployment constraints done in the next three weeks?  Who is
> going to be the lead on this?  William?

Tags end up in a position where they might sometimes be placement
directives, and might sometimes be service constraints. That is the
nature of free-form tags. (I want the next unit of this service to be
on this specific machine, vs I want units of this service to be all on
the same rack.)

I think the real restriction is just that we haven't clearly defined
how we want provider-specific constraints to operate, especially with
pre-allocated machines (juju add-machine). So we could "just implement
it" and not worry too much about the wider story.
...

> 
> 3. API only non-bootstrap agents
> 
> How close are we on this one?  I know we have been making good
> progress, but what is the likelihood of this being done in the next
> three weeks?
> 

Uniter is very close (we have the API and the client side, just need
to update the actual worker which Dimiter is already actively working on.)

Provisioner is a bit more unclear, but it should be possible. It is
more unclear because sometimes the provisioner needs provider
credentials (to spin up new ec2 instances), and sometimes it doesn't
(to spin up new LXC containers).

And there has been a few rounds of fallout because of the transition.
(Like Upgrader triggering restarts until you had run 'juju status' to
give the secret provider credentials, etc).

I think if we honestly spike on getting Provider done, it will be done
in time.

> 8. What else do we have that is new or note worthy that we are
> working on in the next three weeks?
> 

ssl-hostname-verification: false

This is necessary for CTS in their sales discussions. (Private
openstack clouds tend to have self-signed certificates.) I have the
first pass of it working, but the tentacles reach deep. The new
simplestreams stuff brings up a new avenue where we are doing
potential HTTPS requests (we read the public-bucket as
just-another-http-site). We also need to fix up cloud-init to allow
downloading the tools at startup time. And this is slightly hampered
by the fact that we don't actually have any clouds we have access to
that don't have valid certificates. (So while I can reason about where
we might be accessing it, we could do the work and then find that we
missed a place that makes it so you *still* can't use juju-core there.)

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (Cygwin)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlIvL/gACgkQJdeBCYSNAAMbAACfX34+djmug/MT4FWiDQ9AjO8L
j08An1UT5AZOpUcWYFN/GmuN6hNbtp91
=HwTx
-----END PGP SIGNATURE-----



More information about the Juju-dev mailing list