20:01:52 <lifeless> #startmeeting tripleo
20:01:53 <openstack> Meeting started Mon May 27 20:01:52 2013 UTC.  The chair is lifeless. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:56 <openstack> The meeting name has been set to 'tripleo'
20:02:15 <lifeless> #link https://wiki.openstack.org/wiki/Meetings/TripleO
20:02:32 <lifeless> #topic bugs
20:02:55 <lifeless> #link https://bugs.launchpad.net/tripleo/
20:04:11 <vbannai> Hi
20:04:15 <lifeless> hi
20:04:32 <lifeless> so - we have 9 criticals
20:04:38 <lifeless> because, we found out how broken things were :)
20:05:24 <lifeless> we need some volunteers to step up and hack
20:06:11 <lifeless> devananda: are we getting you good enough info with the bm specific bugs ?
20:06:40 <devananda> lifeless: largely, yes.
20:06:57 * devananda looks for al ink
20:07:25 <devananda> #link https://bugs.launchpad.net/nova/+bug/1184445
20:07:27 <uvirtbot> Launchpad bug 1184445 in tripleo "deploy / delete fragility" [High,Triaged]
20:07:38 <devananda> i posted a cmment there with several possibly-related-but-vague bugs
20:07:46 <devananda> other than those, i think everything's been pretty clear
20:08:34 <lifeless> cool
20:08:52 <lifeless> the main themes I see are:
20:08:57 <lifeless> - broken trunks in other projects
20:09:09 <lifeless> - too much boot-stack-undercloud hardcoding in our elements (which we knew about)
20:09:15 <lifeless> - baremetal fragility
20:10:50 <lifeless> anything else to say on bugs? Anyone volunteering?
20:12:30 <lifeless> guess not :P
20:12:41 <devananda> nothing more here. I fix 'em when I can :)
20:12:47 <lifeless> ok grizzly test rack POC status
20:13:01 <lifeless> we have a rack, with 21 compute nodes deployed via heat + nova bm
20:13:12 <lifeless> we've had to hand tune the control plane
20:13:17 <lifeless> and manually setup quantum
20:13:25 <lifeless> bugs filed on every manual step.
20:13:39 <lifeless> the 21 compute nodes though were fully automatic
20:13:45 <lifeless> once we got all the bugs hammered out
20:14:00 <lifeless> everyone say \\\o////
20:14:21 <dkehn> lifeless, \\\o////
20:14:21 <lifeless> we've found mores issues in heat - in particular the hostname issue
20:14:24 <GheRivero> o/
20:14:39 <lifeless> diskimage-builder needed a couple of tiny tweaks
20:15:01 <dkehn> lifeless, as far as volunteering https://bugs.launchpad.net/tripleo/+bug/1174149, can we talk later about it
20:15:02 <uvirtbot> Launchpad bug 1174149 in tripleo "VLANs for bare metal require manual configuration" [High,Triaged]
20:15:20 <lifeless> dkehn: yes, absolutely.
20:15:40 <dkehn> lifeless, I'm assuming this will be in the quantum area?
20:16:11 <dkehn> lifeless, once I get situated in Europe will ping for more info
20:17:35 <lifeless> yes
20:17:45 <lifeless> there's another quantum bug upcoming too
20:18:36 <dkehn> there's always another bug in quantum
20:18:50 <lifeless> ok
20:18:58 <lifeless> #topic Grizzly POC
20:18:59 <lifeless> see above
20:19:05 <lifeless> #topic open discussion
20:19:19 * lifeless opens it to the floor
20:19:31 <dkehn> BTM they want to disscuss another approach for the extra_dhcp_opts using the quantum meta server, or something will get more info in the next meeting
20:19:44 <dkehn> which could spell more delays
20:21:05 <devananda> dkehn: any more info on that besides garyk's one comment on the review?
20:21:17 <dkehn> thats it
20:21:32 <dkehn> jsut concerned that these have a way of boiling over a bit
20:21:38 <devananda> aye
20:21:48 <devananda> i'll stick around for the quantum meeting and say some things
20:22:04 <dkehn> my position will be to got with what we have then deal with other needs as they arrive
20:22:09 <devananda> "lets make this more general" is sometimes the right thing, but not always
20:22:12 <dkehn> cool
20:22:17 <lifeless> yeah
20:22:26 <lifeless> this is already super general.
20:22:31 <lifeless> arbitrary options.
20:22:31 <dkehn> I think its the right thing, but would like to get this moved on
20:22:39 <lifeless> oh god please yes.
20:22:39 <dkehn> no kidding
20:22:49 <lifeless> whats the review url for it ?
20:22:51 <devananda> "This could be useful for all" (quote from his comment)
20:22:58 <devananda> #link https://review.openstack.org/#/c/30441/
20:23:00 <dkehn> https://review.openstack.org/#/c/30447/
20:23:04 <devananda> erm
20:23:10 <dkehn> worng
20:23:23 <devananda> it's 30441
20:23:27 <dkehn> devananda, right https://review.openstack.org/#/c/30441/
20:23:33 <dkehn> me wrong
20:23:48 <dkehn> the 30447 is the client
20:23:59 <dkehn> python-quantumclient
20:24:02 <dkehn> sorry
20:25:24 <devananda> i dont see dhcpopts on the agenda?
20:25:46 <dkehn> for quantum meeting
20:25:48 <dkehn> ?
20:25:50 <devananda> right
20:26:21 <dkehn> probably not, so two approachs, be quiet it goes away, bring it up and god save us
20:26:42 <devananda> um, it's not going away. this needs to get resolved, so i'm going to bring it up, i guess
20:26:46 <dkehn> I'm wondering if gary will bring it up
20:26:48 <devananda> lifeless: unless you want to?
20:27:22 <dkehn> I can bring it up, during the open discussion I guess
20:27:28 <dkehn> my thingy
20:27:55 <lifeless> devananda: you already have the discussion open with him
20:28:14 <lifeless> devananda: I would say 'dhcp options are generic but specific; hiding them in a metadata thing adds no value
20:28:22 <lifeless> devananda: which I think is roughly what you already said
20:28:31 <devananda> ack
20:28:32 <lifeless> back in a sec
20:28:56 <dkehn> In addition, if they get more completicated, whcih they are not now, they won't get used
20:33:11 <lifeless> right
20:33:47 <lifeless> so key points: DHCP options are highly relevant to ports; no reason to hide it away. Doesn't stop a generic 'metadata' thing later, if that is considered a good idea.
20:34:06 <dkehn> that how I see it
20:34:34 <dkehn> moving forward does preclude the metadata thing
20:35:24 <lifeless> dkehn: does? or does not?
20:35:40 <dkehn> doesn't sorry
20:35:45 <dkehn> does not
20:36:26 <dkehn> once this portion complete, we can move on the metadata, which will require some understanding
20:36:48 <lifeless> yeah; personally I'm skeptical of generic things
20:37:07 <lifeless> because you still need reliability and predictable locations - a well known path - for clients to be written
20:37:13 <lifeless> overly generic is death to APIs
20:40:13 <lifeless> ok
20:40:17 <lifeless> it seems like we're done.
20:40:22 <lifeless> Last call for topics
20:41:16 <lifeless> #endmeeting