17:00:07 #startmeeting Ironic 17:00:07 #chair devananda 17:00:07 Meeting started Mon Sep 14 17:00:07 2015 UTC and is due to finish in 60 minutes. The chair is NobodyCam. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:07 Welcome everyone to the Ironic meeting. 17:00:08 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:10 The meeting name has been set to 'ironic' 17:00:11 Current chairs: NobodyCam devananda 17:00:18 o/ 17:00:19 Of course the agenda can be found at: 17:00:19 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 17:00:22 o/ 17:00:22 o/ 17:00:23 hello 17:00:24 \o 17:00:27 #topic Greetings, roll-call and announcements 17:00:27 o/ 17:00:27 Roll-call: Who's here for the Ironic Meeting? 17:00:31 o/ 17:00:32 o/ 17:00:36 \o 17:00:38 o/ 17:00:41 o/ 17:00:44 o/ 17:00:55 welcome all 17:00:56 \o 17:00:57 o/ 17:01:19 o/ 17:01:31 o/ 17:01:32 o/ 17:01:33 I haz announcements when we're ready for them :) 17:01:41 lets jump into it 17:01:42 #topic announcements: 17:01:47 o/ 17:01:58 hi, a few things: 17:02:15 we released python-ironicclient 0.8.1 today with a whopping 3 changes 17:02:23 that will be stable/liberty for the client 17:02:30 w00t 17:02:55 I hear tail we may also cut the first bifrost release this week 17:03:06 That is my hope *hopes* 17:03:07 TheJulia: ^^^ ???? 17:03:23 #link https://gist.github.com/jimrollenhagen/d626f93472ff8bda9759 17:03:30 changelog (if people are interested) ^ 17:03:32 we'd like to enter a soft freeze for ironic beginning thursday - this means we shouldn't land anything risky. the goal is to release 4.2.0, to be stable/liberty, next thursday 17:03:45 lucasagomes: for the client? 17:03:51 NobodyCam, yup 17:03:55 even better link about the client: 17:03:57 #link http://lists.openstack.org/pipermail/openstack-announce/2015-September/000622.html 17:04:13 awesome 17:04:42 anyway, cores please honor this soft freeze and don't land anything risky after thursday 17:04:45 also devananda should be back 17:04:54 yup - i'm back from vacation 17:05:03 \o/ 17:05:10 and for the third announcement... we're taking summit session ideas on this etherpad here: 17:05:11 #link https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas 17:05:25 jroll: are you thinking we should we block with -2 until afterthe freeze? 17:05:30 jroll, ack, it's a bit odd to release the client before the server 17:05:37 but we can/should discuss it later 17:05:38 NobodyCam: yes, if needed 17:05:44 I don't have confirmation from ttx yet about how many slots we'll have -- but I am hoping it will be roughly the same # as last summit 17:06:00 lucasagomes: yeah, let's discuss later, I'd like to think about how we feature freeze (sadface) more for Mitaka 17:06:06 devananda: i thought ttx sent out email about the slots 17:06:11 lucasagomes: as we don't control the library freeze 17:06:16 jroll, sure +1 17:06:22 rloo: oh - i may have missed it then when combing through all the backlog ... 17:06:43 devananda: i htink it was 4, 4, 1/2 day, but not sure. 17:06:48 devananda: http://lists.openstack.org/pipermail/openstack-dev/2015-September/073654.html 17:06:55 it isn't set in stone yet, that's "proposed" 17:06:56 rloo: ahh yes, there it is 17:07:01 yea 17:07:31 apologies for the noise. i apparently hvae more emails to catch up on 17:07:46 :) 17:08:09 are there anyother announcements? 17:08:40 * jroll is fresh out 17:09:04 ok then we'll move on... Thank you jroll and everyone 17:09:08 #topic SubTeam: status report 17:09:09 Posted on Whiteboard 17:09:09 #link https://etherpad.openstack.org/p/IronicWhiteBoard 17:09:55 * jroll updates neutron thing 17:10:57 awesome Ty :) 17:11:33 comments / question on status's? 17:12:15 jroll, others: are those 11 high bugs stuff we want fixed in liberty release? 17:12:24 giving folks a little time to review 17:12:43 jroll: i haven't looked at your list of bugs in the etherpad (that's why i picked on you) 17:12:55 rloo: oh, I should have mentioned the bug list before :) 17:13:09 rloo: I listed the ones I'd like to see fixed, others may have varying opinions 17:13:14 and i see one HI nova bug there too 17:13:17 (under release status) 17:13:23 jroll: ok, will take a look later, thx. 17:13:40 * lucasagomes would like the clean* stuff in 17:14:08 lucasagomes: the abort? 17:14:10 the bugs seems to be all in progress as well 17:14:11 lucasagomes: the abort stuff? +1 17:14:17 rloo, jroll yes sir 17:14:18 clean + zapping + RAID are my favourites 17:14:29 * NobodyCam refreshes his agenda and see a new line item 17:14:34 maybe folks can annotate/add to jroll's list then 17:15:01 yes, feel free 17:15:15 do you guys think that a review jam would help with some of it? 17:15:26 maybe? 17:15:32 when? 17:15:35 I know rloo and myself have just been constantly review jamming 17:15:42 by ourselves :P 17:15:48 if we have a quorum, idk tomorrow or wed. we should try get the fixes in 17:15:53 most (if not all bugs) already have a fix proposed 17:15:59 I'd be down for a jam though 17:16:00 jroll: i was wondering. yeah, by ourselves. Seems to be working fine :) 17:16:19 I can attempt to make it 17:16:24 i'd like to focus on feature patches this week. fine with a review jam for bugs after that. 17:16:25 yeah we can ask cores to prioritize those stuff and they can review at their own time 17:16:27 that works too 17:16:58 * lucasagomes will focus on review more tomorrow 17:17:03 don't we have a hi proitory rewiew list (etherpad) somewhere? 17:17:03 I think we already do :) 17:17:14 (that was for " prioritize those stuff") 17:17:45 i think for sure we should regroup on Monday, to prioritize/review jam... 17:17:58 rloo: ++ 17:18:07 +1 for create a priority review list 17:18:11 NobodyCam: under release status on the whiteboard, I put my view of the priorities for the rest of this cycle... deva seemed to agree with them 17:18:29 awesome then I'll start with that 17:19:07 ++ 17:19:15 good to move on to the first item on the agenda 17:20:08 #topic Config drive support for hybrid whole disk images 17:20:16 stendulker: thats you 17:20:29 Thanks NobodyCam 17:20:49 This is about hybrid whole disk images. We cannot add config drive to such images 17:21:11 parted/sfdisk recognise these images as GPT based while creating config drive 17:21:31 this feels kinda like a bug 17:21:32 these images cannot be booted in BIOS boot mode once config dive is created 17:21:48 No error is thrown during config drive ceration 17:21:51 I have an unnatural aversion to using expect scripts 17:21:58 +1 heh 17:22:12 I have removed expect and using sgdisk instead... 17:22:16 stendulker, I remember we investigate it a bit, do you still have the pre/pos partition layout ? 17:22:28 sgdisk ++ 17:22:31 was trying to use expect to use 'parted' 17:22:36 but that did not work 17:22:46 lucasagomes: remember when you switched things to parted, I believe some folks raised concerns about things like this and the need to support sgdisk .... 17:22:50 have used sgdisk/gdisk combination and poated a patch today 17:22:54 so... is there a question on this topic, other than "can people review this pls?" 17:22:59 it sounds like the gpt table just needs to be repaired 17:22:59 but that will not help in bios boot mode 17:23:11 devananda, yeah. There was some discussion about whether use parted or sgdisk 17:23:29 #link https://bugs.launchpad.net/ironic/+bug/1486887 17:23:30 Launchpad bug 1486887 in Ironic "Config drive creation fails in UEFI boot mode for whole disk image " [High,In progress] - Assigned to Shivanand Tendulker (stendulker) 17:23:35 i openend it for discussion as i was not able to find any solution for bios boot mode for such images 17:24:08 stendulker, so did sgdisk work? 17:24:16 currently coreos, opensuse creates such hybrid images 17:24:33 sgdisk works only in UEFI boot mode. No approach works for bios 17:24:53 stendulker, right, have you tried to manually add a partition on the image 17:24:57 Are there any Python libraries to manipulate partition tables? 17:25:07 and boot it in BIOS mode? just to see if that works 17:25:09 so this might be a terrible suggestion, but one way to solve this is to pre-create the configdrive partition in the image, label it config-2. the agent will pick this up, as will cloud-init 17:25:14 jlvillal, blivet 17:25:20 thx 17:25:34 jlvillal, that's what anaconda uses underneath, it's neat 17:25:43 lucasagomes: manually in the sense? 17:25:51 stendulker, yes 17:26:28 lucasagomes: I tried the steps manually to ensure drive is created. It creates, but boot fails 17:26:55 jroll: that looks better... 17:27:21 stendulker: do you have a paste of the actual error? 17:27:34 or when boot fails 17:27:44 also there is no config drive support for any of iscsi based drivers when whole disk image is deployed 17:27:46 stendulker: it would be nice to work with stock coreos images, but this is a workaround that works... that's what the coreos onmetal image does 17:28:27 NobodyCam: CoreOs gives red screen on HP Proliant. Opensuse no error is returned, but node goes for reboot perpetually 17:28:57 should we continue this discussion on the bug(s) ? seems like it may take more investigation than meeting time permits 17:29:21 NobodyCam: I will capture screen shot of coreos boot failure 17:29:27 devananda: yes... but this is the only item on the agenda to 17:29:29 day 17:29:35 devananda: +1, I don't think there's anything contentious here... I don't see any major questions 17:29:40 lots of time for open discussion then :) 17:29:45 it's just "people doing work" 17:29:49 jroll: right 17:30:16 stendulker: can you also see if you can dump partition table data for what is on disk when that boot failed, for additional context 17:30:26 are we good with taking this to the review / bug? 17:30:42 or main channel even :P 17:30:46 ++ 17:30:49 thT TOO 17:30:53 gah 17:30:58 ++ 17:30:58 that too :) 17:30:58 TheJulia: sure 17:31:02 stendulker: Thanks 17:31:08 thanks! 17:31:15 and thank you stendulker for working on it :) 17:31:23 ok then let jump into : 17:31:26 #topic Open Discussion 17:32:11 anyone want to discuss summit ideas? 17:32:23 I like summits 17:32:25 just FYI.. I will be making the trek back to cali starting on friday 17:32:26 :P 17:32:37 I definitely want to do a retro on this cycle 17:32:39 * jlvillal is glad he got approved last week to attend summit :) 17:32:48 whether that's a session or part of the group thing friday 17:32:50 so I'll be afk fri thru prob tues / wed 17:32:50 jlvillal: \o/ 17:32:55 :) 17:33:12 NobodyCam: woot 17:33:20 you just missed the heat wave :P 17:33:29 ya :) 17:33:31 jroll: retro on releases in particular would be great 17:33:34 Not sure if in retro, maybe discussion a little on gate breakages. For like two weeks it seemed our gate repeatedly got broken. Can we do anything to help prevent that better? 17:33:45 devananda: +1 17:33:47 jlvillal: ++ 17:33:51 jlvillal: is there a summary somewhere of what/how/why it broke? 17:34:12 I don't think so. I know Neutron hit us. oslo.versioned.objects hit us. 17:34:15 there are notes on the whiteboard i beleieve 17:34:21 Maybe other people remember other reasons. 17:34:31 ubuntu problem 17:34:35 Oh yeah, I think on the whiteboard we have been archiving them. 17:34:38 k, i think that'd be a great session 17:34:41 ++ 17:34:57 maybe more generally "how do we improve our gate" starting with the breakages 17:35:00 we should start the discussion with a post-mortem report // details of each breakage 17:35:09 jroll, + functional testing, I guess 17:35:14 yep 17:35:40 +1 though realizing I have made little progress :( 17:36:00 jlvillal: lekha should be at the summit iirc 17:36:02 we have implicit dependencies on things like distros, but being implicit, i suspect we could all do with better awraeness of them 17:36:17 and ew have explicit dependencies on other projects (eg, oslo, neutron, etc) -- some of which maybe those projects aren't aware of 17:36:18 as long as the session isn't about 'we need functional testing'. we know that. 17:36:23 jroll: Great. 17:36:24 rloo: ++ 17:36:45 "test all the things" makes a great tshirt, but a very short session 17:36:57 lol ++ 17:37:01 hah 17:37:16 A hackathon type session might be useful. 17:37:27 For functional testing. 17:37:32 yep, sounds like a good friday thing 17:37:38 i'd like our summit sessions to have goals, so that there are action items or concrete things out of them. 17:37:44 thats what friday is for? hackathon stuff no? 17:37:59 friday is for "whatever" 17:38:01 Please make it early on Friday. My flight is at 4pm I think :( 17:38:05 we only have a room for fri AM, so that isn't much time. 17:38:28 oh 17:38:31 i think it took us about an hour just to figure out what we were going to do on Fri. in last summit. 17:38:36 * lucasagomes got disconected :-( 17:38:40 yea - friday is always awkward 17:38:47 lucasagomes: we're talking about summit sessions 17:38:56 we've never really used hte afternoon, so this time it looks like we won't have one 17:39:06 rloo: sounds like a good reason to get organized ahead of time :) 17:39:10 stendulker, (sorry all) I lost the scrollback. Is it a public image? 17:39:21 jlvillal: by friday we'll all be fried -- hacking on code is not likely, based on past experience 17:39:24 lucasagomes: yes. 17:39:32 stendulker, we talk about it on #openstack-ironic 17:39:37 yeah, i think the more organized we are, the more we'll get out of the sessions. my hope anyway :) 17:39:47 rloo: ++ 17:39:50 lucasagomes: http://eavesdrop.openstack.org/irclogs/%23openstack-meeting-3/%23openstack-meeting-3.2015-09-14.log.html 17:39:51 rloo: ++ 17:39:51 lucasagomes: ok 17:39:58 jlvillal, thanks 17:40:00 devananda: Okay, makes sense but :( 17:40:02 lucasagomes: thank you 17:40:47 IF we have time and can take a look at specs before summit, would be good to have one session to hash out contentious/disagreements with particular specs. 17:40:53 jlvillal: i mean, you're more than welcome to code on friday! 17:41:01 jlvillal: just don't expect me to :p 17:41:29 devananda: Okay :) 17:41:41 it is true that by friday /me 's brain is cooked 17:41:52 jlvillal, re gate breakage https://review.openstack.org/#/c/220516/ this will help (already merged) 17:41:55 it could a bit better this time, as we don't have Monday 17:42:00 since now we can test the swift module with ironic 17:42:01 NobodyCam: at least it's only 4 days this time 17:42:10 :) 17:42:32 wrt the liberty etherpad: https://etherpad.openstack.org/p/mitaka-ironic-design-summit-ideas 17:42:35 dtantsur: monday is the board meeting, so some of us will still be busy ... 17:42:48 oh, I feel sorry for some of us 17:42:54 i always find it useful (or would think it would be useful if it were there) if folks put their names beside the stuff they enter there 17:43:19 rloo: ++ just like on the agenda 17:43:33 rloo, done 17:44:45 anything else for OD? 17:45:03 one thing 17:45:10 :) 17:45:12 devananda: when/how do we want to decide on sessions? :) 17:45:23 in a meeting, separate thing, in a secret cave? 17:45:49 shh. don't tell anyone about the secret cave :p 17:45:57 wait, you have a secret cave?!?!?! 17:46:04 lol 17:46:24 oops 17:46:37 bet it's really a hollow volcano 17:46:40 we can change them up to roughly the week before the summit 17:46:48 * NobodyCam now unders stands the "same bat time/channel" reffs 17:47:09 so let's keep discussing on the 'pad and in channel for another month 17:47:30 sounds good 17:47:36 hopefully we can distill things down to the right # of slots by mid october 17:47:44 gives us time to review specs and such too 17:47:47 yup 17:48:05 ++ 17:48:07 cool 17:48:15 I'd rather keep the official lists empty until then, than have to keep changing them as we go 17:48:36 makes sense 17:48:37 also this gives us time to earmark joint sessions // work with other project teams to decide which track will host it 17:49:05 yeah, sounds good 17:49:21 For joint sessions, are their other projects besides Nova and Neutron we want to work on things with? 17:49:36 maybe oslo? 17:49:37 Those are the first two that pop to mind for me. 17:49:46 tripleo? not sure here though 17:49:51 people keep asking about an ironic/neutron session, what do we need to talk about there? the design is done/solid/approved... 17:49:58 kolla? magnum? 17:49:59 last summit, we talked with cinder but not sure what happened with that work 17:50:08 devananda, oh, that might be interesting 17:50:20 kolla's with guns? 17:50:21 tripleo is using ironic in their cloud region 17:50:28 :p 17:50:39 unfortunately magnum on baremetal isn't terribly useful without neutron networking stuff :( 17:50:42 roo: I posted the spec for cinder integration 17:50:43 fyi *ten* minutes 17:50:55 jroll: right. but folks are def interested in magnum-on-ironic 17:51:03 ++ 17:51:04 smoriya: ah, thx 17:51:05 rloo: I posted the spec for cinder integration 17:51:16 smoriya: link? 17:51:19 rloo: so I talked about in-band volume stuff with nova at their midcycle... they don't love the methods but are cool with the feature. just needs to happen 17:51:20 there's cinder too, but I think we ironed out a good part of it last cycle 17:51:26 thiagop: https://review.openstack.org/#/c/200496/ 17:51:30 smoriya: awesome. I'd love to see that work happen during Mitaka 17:51:31 yeah, idk if we need a session with cinder 17:51:34 smoriya: thanks 17:51:57 and I'd like you guys to review it. 17:52:03 I am interested in Kolla in general. Ironic integration sounds good too. 17:52:11 mgoddard and I are discussing the target_info format in instance_info. 17:52:20 L119 in the spec. 17:52:27 We'd like to know other guys opinions. 17:52:43 thiagop: np 17:52:49 jroll: wdyt about trying to land the cinder spec after the 4.2 release, but before the summit, so that folks can start on the code? 17:53:02 devananda: I would like to do that for every feature :) 17:53:11 jroll: there's a lot of features :p 17:53:17 you get what I mean :P 17:53:19 yup 17:53:21 devananda: oh thats a really good idea 17:53:37 we should be doing lots of spec reviews after 4.2 17:53:50 smoriya, added to my list will take a look at it! 17:53:52 which reminds me... 17:53:54 thanks for the spec 17:53:58 devananda: we should have a larger stable maint team 17:54:03 so yah, i agre with jroll -- if we can land that spec, then we probably don't need a joint session with cinder 17:54:06 devananda: more than just you, at least 17:54:10 jroll: yes! 17:54:17 lucasagomes: Thanks! 17:54:43 sounds like jroll just volentereed 17:54:44 jroll: does RS do anything with stablemaint? 17:54:59 my impression is "no" but I could be totally wrong on that 17:55:06 *five* minutes 17:55:21 devananda: not that I'm aware of, but who knows 17:55:23 what's RS? 17:55:28 rackspace 17:55:30 rack space 17:55:31 (I presume0 17:55:46 jroll: thx. (d'oh) 17:55:51 devananda: but my real need is to not be blocked on you to approve stable branch bugfixes :P 17:55:58 s/my/the/ 17:56:18 jroll: i'm happy to share +A rights on that with you, as you've been doing a lot of the release things anyway 17:56:18 we should probably have at least 3-4 cores with stable branch permissions 17:56:42 devananda, jroll ++ 17:56:46 historically, stable-maint was kept really small 17:56:53 ++ 17:56:56 right 17:57:02 adam is not around as well so, it would be good to have jroll as a core there too 17:57:06 i'm happy with 2. if more than two, I'm fine with that too. 17:57:09 maybe I should chat with the stable team about this stuff... or doug or someone 17:57:17 whether we intend to keep it small or what 17:57:19 jroll: yah. there's some stuff to think about 17:57:26 i agree it should be more than just me, tho 17:57:35 adam has basically bowed out so he doesn't count. 17:57:36 yep 17:57:41 rloo: right 17:58:13 *two* minutes 17:58:33 * NobodyCam would vote for adding jroll 17:58:37 ++ 17:58:46 yay, more work \o/ 17:58:49 lol 17:59:02 jroll: look on the bright side, more POWER! 17:59:07 lol 17:59:10 begin the corruption 17:59:16 30 second mark 17:59:19 jroll, also would be nice to have a formal release of IPA around Liberty... 17:59:29 dtantsur: is trown not working on that? 17:59:29 reminder - PTL and TC voting is coming up -- dont forget to go out and vote! 17:59:36 (next week and hte week after) 17:59:46 jroll, I think he does 18:00:03 dtantsur: I'm happy to push the release tags or whatever if it's set up to do so 18:00:06 * jroll back to channel 18:00:08 thanks all! 18:00:09 last comments 18:00:12 thanks 18:00:12 Speaking of broken gate.... 18:00:13 cheers 18:00:13 Thank you all 18:00:17 I think we are broken again :( 18:00:17 Great meeting 18:00:25 uggh 18:00:29 #endmeeting