15:23:52 <j^2> #startmeeting openstack-chef
15:23:53 <openstack> Meeting started Thu Mar 19 15:23:52 2015 UTC and is due to finish in 60 minutes.  The chair is j^2. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:23:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:23:57 <openstack> The meeting name has been set to 'openstack_chef'
15:24:05 <j^2> Hey everyone!
15:26:11 <cmluciano> o/
15:27:18 <wenchma_> hello
15:27:29 <j^2> I think the main thing is reviews. Other than that I've confined multi-nova is broken. Those are the only thing I have to discuss
15:27:45 <j^2> Any other topics?
15:30:11 <j^2> I'll give it a couple mins
15:31:08 <wenchma_> #topic add bare-metal cookbook in launchpad/openstack-chef
15:31:16 <wenchma_> #link https://launchpad.net/openstack-chef
15:31:36 <wenchma_> miss bare-metal cookbook info
15:32:01 <wenchma_> need someone to add who has the permission
15:32:03 <wenchma_> ;)
15:32:19 <j^2> Yeah that cookbook has been a thorn in our side
15:32:39 <j^2> I truly believe we should make it optional
15:33:40 <j^2> I don't think it's something most will use, nor I wouldn't want to say chef OpenStack suggests using it
15:34:26 <j^2> Officially chef the company is going to say MAAS is what you want to use for bare metal pro
15:34:46 <j^2> But that's me just ranting
15:35:10 <wenchma_> MAAS
15:35:33 <j^2> We should keep development of ironic, but not make it part of the "core"
15:35:58 <j^2> #link http://Maas.ubuntu.com
15:36:14 <wenchma_> maybe not part
15:37:33 <j^2> I mean how should we test it in the testing stack? We don't bootstrap it in there. It isn't used
15:38:46 <j^2> Thoughts?
15:38:49 <wenchma_> it is not easy to test for physical deployment,
15:39:02 <j^2> Yep
15:39:18 <wenchma_> but we do some successful cases
15:40:04 <cmluciano> We're just asking how we want to test it?
15:40:36 <j^2> I think until we can we shouldn't be part of the testing-stack
15:40:38 <wenchma_> or use virtual maching  to simulate it
15:41:13 <cmluciano> Can it be tested by itself?
15:41:52 <wenchma_> maybe not
15:42:02 <cmluciano> Like with test-kitchen?
15:42:30 <j^2> ™ assumes you have an is
15:42:39 <j^2> Tk*
15:42:50 <j^2> os*
15:42:52 <cmluciano> Assumes what?
15:42:55 <cmluciano> Ah
15:43:13 <j^2> Autocorrect is a bitch
15:44:24 <cmluciano> I guess he question is who is the audience for this. If this isn't something the majority of people will use it should be tested by itself in a standalone stack
15:45:32 <cmluciano> for the most part I have resorted to commenting it out completely. And IMO this is a nice to have
15:45:46 <j^2> Agreed I'm in that same boat
15:45:55 <cmluciano> I think having working neutron and nova multi is a more pressing matter
15:46:22 <cmluciano> According to the openstack meetup, 75% of people are using neutron
15:46:32 <j^2> As with centos7 support
15:46:43 <j^2> Yep
15:49:07 <j^2> I don't think we have to decide now, but I wanted to make my personal view clear
15:50:11 <openstackgerrit> Mark Vanderwiel proposed stackforge/openstack-chef-repo: New test_patch helper script  https://review.openstack.org/164884
15:50:14 <j^2> And I think it's pretty close in line with the major use cases
15:50:49 <cmluciano> ok so do we have any other topics?
15:51:03 <j^2> Reviews
15:51:21 <j^2> Need them done
15:51:38 <j^2> And multi nova
15:52:05 <cmluciano> is broken?
15:52:18 <markvan> Yeah for bare-metal, if we just pull from roles, it still can be used, just will need different run list
15:52:49 <markvan> I already pulled the conductor side, just need the api side, and its then "optional"
15:52:56 <j^2> Yep multi-nova needs that rabbitmq fix in dues to guest user
15:53:34 <cmluciano> this one https://review.openstack.org/#/c/165506/ ?
15:53:58 <j^2> Yep
15:56:32 <markvan> j^2: for that one, I still don't understand the compute change, that's already handled in Common, I verified that in a test.
15:57:01 <j^2> The inheritance doesn't work
15:57:37 <j^2> That's why it's guest/mypass
15:58:33 <j^2> And compute happens after common so it over rides the changes
16:00:30 <j^2> yep
16:00:38 <j^2> yay on a keyboard now i can actaully type
16:02:34 <markvan> I don't see anyting in compute that overrides the Common attributes
16:03:33 <j^2> run it with out my changes and it’s broken, i walked through the changes that happens and made them and verified with both aio and multi and now rabbitmq can recieve messages from the remote machines
16:04:28 <j^2> also these are override attributes on the testing stack. in a real production environment you’d do something completely differnt
16:04:52 <j^2> this is an override attribute to get our stack working again, this isnt even a change to the recipes
16:08:35 <markvan> humm, I think the issue is how the override is being done, as this will be broken for all the other components in a multi node setup
16:08:55 <j^2> it’s a very strong possibility
16:09:29 <j^2> but try multi-nova in the state it’s in, then we can continue this discussion. i’m telling you it’s broken and we need to fix it. and the changes that i made fixes them
16:09:47 <markvan> so, I think we should figure out the root cause here before injecting this into the repo
16:10:06 <j^2> i disagree
16:10:24 <j^2> i think we should have a working stack; then investigate without the change
16:10:25 <markvan> yup, I would like that fixed, but want to do it without a hack, as that is hiding another issue
16:10:53 <j^2> if it’s a huge issue it’ll take a long time to fix, and it’s more important to have a working stack then a fix that could take months to find
16:11:09 <j^2> and the only reason why aio is working is because of the guest user
16:12:06 <j^2> ah one other thing before we break
16:12:07 <j^2> https://bugs.launchpad.net/openstack-chef/+bug/1433808
16:12:08 <openstack> Launchpad bug 1433808 in OpenStack + Chef "[all] Specs have a deprecation notice" [Medium,In progress] - Assigned to JJ Asghar (d-jj)
16:12:16 <markvan> maybe give this a couple days to figure out, and then if still not resolved, we can put in, but will need extra doc to explain the issue and the current workaround.
16:12:44 <j^2> i’m voting for ServerRunner due to you should be running chef-server not chef-solo
16:13:25 <j^2> markvan: that’s the problem, i cant explain it other than it’s broken. rabbitmq has auth creds issues and the only way to get around it i’ve found is this override
16:14:22 <markvan> yup, but we can easily dive into the attribute issue to see why compute attr are not being updated properly.
16:14:40 <markvan> As for the ServerRunner, can we change that without the Rake Chef 12 gates in place?
16:15:06 <j^2> yep, serverrunner just needs to be called
16:15:09 <j^2> instead of server
16:16:27 <markvan> sound ok to me if it will get thru our existing gates
16:16:34 <j^2> cool-
16:18:49 <j^2> any other pressing matters?
16:19:09 <j^2> otherwise i’ll make the meeting officiall closed
16:19:11 <markvan> as you mentioned eariler, reviews....
16:19:14 <j^2> yep
16:19:17 <j^2> we need to do em
16:19:20 <j^2> we need more cores
16:20:02 <j^2> if you are interested please contact me at jj@chef.io and we can start the process of making you one
16:20:25 <wenchma_> wondeful news
16:20:39 <markvan> oh, and we're going to have meeting next week on Kilo blueprints
16:20:58 <wenchma_> need more reviews
16:21:40 <j^2> yep,
16:22:01 <j^2> i believe we’re aiming for wednesday, but i can be talked out of it
16:22:37 <j^2> come to think of it maybe Tuesday our normal meeting time would be better
16:29:26 <j^2> anything else?
16:29:29 <j^2> going once?
16:29:36 <markvan> n
16:29:43 <j^2> #endmeeting