19:01:27 <devananda> #startmeeting ironic
19:01:28 <openstack> Meeting started Mon Nov 18 19:01:27 2013 UTC and is due to finish in 60 minutes.  The chair is devananda. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:31 <openstack> The meeting name has been set to 'ironic'
19:01:58 <devananda> Hi lucasagomes, NobodyCam, lifeless, everyone!
19:02:05 <lucasagomes> o/
19:02:23 <shadower> hey
19:02:47 <lifeless> o/ devananda
19:02:48 <devananda> NobodyCam seems to still be getting his coffee :)
19:03:15 <GheRivero> o/
19:03:22 <devananda> in the mean time, here's our agenda for the day:
19:03:24 <devananda> #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting
19:04:19 <devananda> that giant agenda aside, I would like to take a little time and go over the ehterpad from the summit
19:04:27 <devananda> make sure we have captured BPs for all the things taht need them
19:04:36 <devananda> #link https://etherpad.openstack.org/p/IcehouseIronicNextSteps
19:04:42 <devananda> #link https://blueprints.launchpad.net/ironic
19:05:52 <devananda> #topic blueprints
19:06:28 <devananda> I sent an email out last week about it, and I see a few folks ahve put up a few BPs
19:06:59 <devananda> shall I poke and prod? or just start creating and assigning based on my notes/memory?
19:07:10 <lucasagomes> yea I did only one, I will try to write more this week (I just got myself fighting against wsme complex type validations and couldn't stop)
19:07:51 <lucasagomes> things like porting bugs doesn't need a bp does it?
19:08:05 <lucasagomes> porting bug fixes*
19:08:06 <devananda> nah
19:08:39 <lucasagomes> right, I can pick some others then, idk "Secure boot (signed firmware)" sounds interesting I will have to do some research on that
19:08:44 <lucasagomes> but I can do it and write a bp
19:09:02 <devananda> anything that we punted on should not get a BP
19:09:35 <lucasagomes> ahn.. right
19:10:22 <devananda> it's the "road to MVP" taht really needs to be addressed
19:10:46 <devananda> and I'd like to capture "important next steps" and "copied from other sessions" as well, since folks will probably be working on some of that this cycle
19:11:22 <romcheg> Is ironic meeting now?
19:11:25 <devananda> and we should somewhat of an idea of what development will be goign on during this cycle so taht I can communicate that to the release manager (thierry)
19:11:28 <devananda> romcheg: hi! yes
19:11:30 <lucasagomes> romcheg, yes
19:11:51 <romcheg> hi, sorry for being late?
19:12:25 <lucasagomes> romcheg, it's grand
19:12:37 <devananda> romcheg: no worries. blame daylight savings time change.
19:12:38 <lucasagomes> devananda, right I can take a look at the pxe integration then
19:14:09 <devananda> romcheg: we were discussing converting the etherpads from the summit int oblueprints
19:15:33 <romcheg> Oh, these are mine: https://blueprints.launchpad.net/ironic/+spec/migration-from-nova https://blueprints.launchpad.net/ironic/+spec/utility-ramdisk
19:16:08 <devananda> great, thanks!
19:16:16 <devananda> we're addign them to the etherpad here https://etherpad.openstack.org/p/IcehouseIronicNextSteps
19:16:22 <devananda> so we can see what does/n't have a BP yet
19:17:33 <romcheg> devananda: cool
19:18:14 <devananda> romcheg: the other topic with your name on it seems to be RPC routing
19:18:25 <devananda> romcheg: still want to take taht on? if so, can you file a BP?
19:18:40 <romcheg> devananda: I'm still working on the BP
19:18:45 <romcheg> 75% done
19:18:51 <devananda> great
19:19:25 <devananda> lucasagomes: for the bug fix porting, how do you feel about making a list somewhere public
19:19:35 <devananda> lucasagomes: so taht we can check off when each bug is ported?
19:19:44 <devananda> lucasagomes: and are there enough bug fixes to port to warrant doing taht?
19:19:45 <lucasagomes> devananda, sure works for me
19:20:00 <lucasagomes> I will try to map that tomorrow then
19:20:08 * lucasagomes writes it down
19:20:44 <devananda> lucasagomes: thanks!
19:21:02 <devananda> ok, that list is looking much better to me
19:21:32 <devananda> I'll keep working on general things (oslo, docs, nova glue, etc)
19:21:43 <devananda> there are more oslo changes to sync, too...
19:22:05 <devananda> any questions about BPs before we start going through the actual agenda?
19:22:38 <devananda> #topic previous action items
19:22:51 <devananda> nova-ironic driver
19:23:07 <devananda> NobodyCam has taken over that work and it seems to be going well
19:23:24 <devananda> he'll probably have things tos ay when he's back from coffee :)
19:23:39 <devananda> lucas to add cli docs
19:24:04 <devananda> if you didn't already, maybe wait a bit -- I may do some rewriting of our docs soon
19:24:14 <lucasagomes> yea shame on me, I started writting a manpage for the CLI but it's stoped in time
19:24:26 <lucasagomes> yea there's some API changes I'm doing in the moment that will also affects the CLI
19:24:43 <devananda> Ah! manpage for the CLI would be great
19:24:57 <devananda> don't worry about API docs or deployer .rst doc files
19:25:09 <lucasagomes> yea I've a patch which is now abandoned but I will revive that soon
19:25:22 <NobodyCam> did our meeting change time
19:25:24 <NobodyCam> ?
19:25:24 <lucasagomes> yea, I think a manpage would be a great doc for CLIs
19:25:25 <devananda> I think we can auto-generate all the API docs from the pecan/wsme code. just need to plumb that
19:25:38 <lucasagomes> devananda, yup, I'm counting on that for the API :)
19:25:39 <devananda> NobodyCam: meeting times are in GMT. the US changed time :p
19:25:50 <NobodyCam> lol
19:25:53 <NobodyCam> sorry
19:26:00 <NobodyCam> I was an hour behind
19:26:16 <lucasagomes> NobodyCam, welcome :D
19:26:21 <NobodyCam> :-p
19:26:50 <devananda> romcheg: on teh gating job, I think you're on the right track & got great feedback from mordred last week. anything to discuss here?
19:27:23 <romcheg> devananda: I got a task from mordred so I'm working on it
19:27:42 <romcheg> As soon as I have results, i'll let you know
19:28:00 <devananda> ack
19:28:18 <devananda> also, romcheg, it looks like transifex is up and working -- it's proposed a few translation patches arleady
19:28:50 <devananda> so everyone who wants to work on translations should be able to do that now
19:28:54 <romcheg> I got an intern so I'm going to do more translations :)
19:29:19 <lucasagomes> I joined the brazilian portuguese team, but didn't do any translation yet, will do when I get some time
19:29:21 <devananda> as far as translation patchsets go -- NobodyCam, lucasagomes -- we should eyeball the jenkins patch submissions to make sure they're sane
19:29:23 <NobodyCam> #link https://www.transifex.com/projects/p/ironic/
19:29:40 <devananda> and then feel free to +2/+A on your own. those don't need 2 +2
19:29:51 <NobodyCam> devananda: ++
19:29:57 <lucasagomes> devananda, right
19:30:42 <devananda> any other comments on previous action items?
19:31:43 <devananda> #topic testing
19:32:12 <devananda> my summary here is, i've talked with infra folks a few times
19:32:27 <NobodyCam> looks like several of the patches for devstack still need looking at
19:32:44 <devananda> the goal is to get a lot of the devstack, tempest, etc, jobs -- both check and gate jobs -- into ironic's pipeline
19:32:57 <devananda> but not add ironic to all the other project's gate pipelines yet
19:32:59 <NobodyCam> 53899 has a couple of -1 on it
19:33:20 <romcheg> I'm on it :)
19:33:58 <NobodyCam> the others look like we need to poke a few people to review them
19:34:08 <devananda> once there is confidence that ironic isn't going to break the gate for all of openstack, AND ironic is able to come out of incubation, then we can enable it in the main gate pipe
19:35:21 <devananda> NobodyCam: i'm not sure we do. some of these actually shouldn't be enabled yet
19:36:03 <NobodyCam> then should we marke them as wip so they don't land unexpectedly
19:36:07 <romcheg> NobodyCam, devananda: Let me finish the last patch before poking around pls
19:36:25 <devananda> romcheg: definitely.
19:37:36 <devananda> NobodyCam: the issue is that some of those patches would add ironic to nova's gate pipeline. which we shouldn't do until ironic is part of the integrated release.
19:37:46 <NobodyCam> i will update the dib walk thru with ltest info ...
19:38:04 <romcheg> NobodyCam: The patches that should not land now are -1ed or WIP
19:38:16 <NobodyCam> romcheg: awesome
19:38:18 <NobodyCam> :)
19:38:19 <devananda> thanks :)
19:38:53 <romcheg> 48109 and 56345 won't break anything
19:39:53 <devananda> time getting short, so i'm going to move on
19:40:03 <devananda> #topic nova driver // deploy bindings
19:40:05 <NobodyCam> :)
19:40:13 <devananda> NobodyCam: that's you :)
19:40:30 <NobodyCam> ok i've been playing with the nova driver this last week
19:41:00 <NobodyCam> bigest issue is dhcp
19:41:28 <NobodyCam> the nova baremetal driver handles "out of band" dhcp for pxe style deploys itself
19:41:49 <NobodyCam> we will need to move that logic in to the pxe driver for ironic
19:42:30 <dkehn> NobodyCam: note: is comming https://review.openstack.org/#/c/50749/
19:43:04 <dkehn> which does the move into the pxe.py
19:43:17 <devananda> dkehn: taht's great, but ....
19:43:37 <devananda> dkehn: the problem is that not all ironic depploy drivers are going to need to set DHCP PXE options
19:43:42 <devananda> dkehn: some may not use PXE at all :)
19:43:49 <dkehn> understand
19:43:52 <devananda> **some may not use DHCP at all, i mean
19:44:21 <devananda> dkehn: is it reasonable for ironic to pass this information directly to neutron (not going through nova) ?
19:44:41 <dkehn> understand, which would be eaiser
19:45:08 <NobodyCam> dkehn: are you thinking about doing the patch for ironic as well?
19:46:42 <devananda> dkehn: i'm looking at the code in nova/compute/manager.py
19:46:42 <dkehn> NobodyCam: thinking, we should talk about it, have a few other items on the plater
19:46:43 <devananda> 1017                 dhcp_options = self.driver.dhcp_options_for_instance(instance)
19:47:06 <devananda> AIUI, there are two different IPs which will be assigned to the baremetal node
19:47:23 <devananda> one during the PXE deploy phase -- this is an IP outside of the tenant network
19:47:29 <devananda> and one when the deploy is done -- this is on the tenant network
19:47:32 <NobodyCam> there is a deploy ip and and final "user" ip
19:47:40 <NobodyCam> :)
19:47:59 <devananda> dkehn: but I dont see how the nova code is removing the DHCP options for PXE booting, or moving them onto the tenant IP ?
19:48:58 <dkehn> at present there is only a setting it up when creation is performed via nova
19:49:17 <dkehn> to remove we assume you are removing the interface
19:49:38 <devananda> interface == MAC ?
19:50:04 <NobodyCam> i'm not sure Ironic can make that assumshion
19:50:20 <dkehn> bad choice or words once the port is removed the DHCP option ass to it are removed
19:51:31 <NobodyCam> his will need more time then we have here. /vote table
19:52:10 <devananda> yea, let's come back to this in channel later
19:52:42 <devananda> #topic open discussion
19:52:50 <NobodyCam> no api/
19:52:52 <devananda> there's a few things left on the agenda, but i want to also open the floor
19:52:53 <NobodyCam> :-p
19:53:02 <lucasagomes> oh
19:53:15 <devananda> we can talk about the API -- but I want folks to have a chance to jump in with other things too
19:53:15 <lucasagomes> so I've some patches upstream reworking the API for ports
19:53:48 <NobodyCam> devananda: ++
19:53:56 <lucasagomes> I'm trying to remove mostly of the complexity of validations we had before
19:54:00 * lucasagomes get the links
19:54:18 <lucasagomes> #link https://review.openstack.org/#/c/56682/
19:54:23 <lucasagomes> #link https://review.openstack.org/#/c/56984/
19:54:28 <devananda> less complexity is good :)
19:54:46 <lucasagomes> please take a look and give me suggestions/comments
19:54:50 <lucasagomes> devananda, yea
19:55:23 <devananda> lucasagomes: will do
19:55:35 <lucasagomes> thanks
19:55:52 <lucasagomes> this will affect the CLI also, because I introduced node_uuid for the API objects
19:55:52 <NobodyCam> I saw a patch to remove keystone validation form vendor pass thru
19:56:05 <lucasagomes> so we don't need to use node_id = <uuid> on ports anymore
19:56:40 <lucasagomes> NobodyCam, https://review.openstack.org/#/c/56612/
19:56:52 <NobodyCam> that the one
19:56:56 <yuriyz> yes
19:57:10 <NobodyCam> should we also look at node-list and node-show
19:57:15 <devananda> NobodyCam: yes. I think yuriyz was workign on that. i haven't had time to give it a good explanation for why I want to -1 it yet
19:57:35 <devananda> yuriyz: i dont like the potential for a public service to have no-auth like that
19:58:12 <yuriyz> devananda, take a look and comment
19:58:23 <devananda> k
19:58:45 <devananda> romcheg: you and I should talk, maybe tomorrow, about the RPC routing stuff
19:58:48 <NobodyCam> deva ack. the reason I ask is nova appears to run some administravite tasks like finding baremetal with out context
19:59:02 <romcheg> devananda: agree
19:59:10 <devananda> romcheg: and how we want to handle >1 conductor, which is why we need the routing :)
19:59:26 <devananda> NobodyCam: nova generates an admin context when it needs to do taht
20:00:18 <devananda> thanks everyone!
20:00:24 <lucasagomes> cheers
20:00:26 <NobodyCam> I couldn't get that to auth thru the client. But I may have done something incorrectly.. I need to dbl check
20:00:33 <NobodyCam> TY ll
20:00:37 <NobodyCam> TY all
20:00:38 <NobodyCam> :-p
20:00:46 <devananda> #endmeeting