17:00:00 #startmeeting ironic 17:00:01 Meeting started Mon May 15 17:00:00 2017 UTC and is due to finish in 60 minutes. The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:04 The meeting name has been set to 'ironic' 17:00:10 o/ 17:00:11 o/ 17:00:14 o/ 17:00:19 o/ 17:00:27 o/ 17:00:34 welcome everyone :) 17:00:40 o/ 17:00:41 o/ 17:00:48 our agenda as usual can be found at 17:00:50 #link https://wiki.openstack.org/wiki/Meetings/Ironic 17:01:03 \o 17:01:11 o/ 17:01:25 * dtantsur gives everyone a minute more to join 17:01:54 o/ 17:02:23 #topic Announcements / Reminders 17:02:34 #info TheJulia has new logo stickers from the Summit. Please contact her with mailing details and she will mail them out to contributors. 17:02:50 thx TheJulia 17:02:59 thanks TheJulia! I can grab some for the Brno office, I'll send you my details later (if I don't forget) 17:03:05 o/ 17:03:10 also, many thx TheJulia for the great ironic demo at the summit last week! 17:03:13 thnx TheJulia 17:03:15 +1 17:03:28 Email and/or msg, either is fine. juliaashleykreger@gmail.com 17:03:31 yes, it was awesome, and really well received (from what I can gather downstream and upstream) 17:03:32 o/ 17:03:37 o/ 17:04:07 #info python-ironicclient 1.13.0 released with the driver composition bits 17:04:13 Thank you everyone! I was surprised how many people though ti was awesome. :) I'll try to post the scripts in the next day or so 17:04:22 yes please :) 17:04:29 +1 Thanks TheJulia :) 17:04:36 any other announcements, reminders? 17:04:42 maybe something important coming from the Forum? 17:04:57 o/ 17:05:15 Nothing that I can think of really, my brain is still unwinding the past week 17:05:34 no worries, get some rest first :) 17:05:43 #topic Review subteam status reports (capped at ten minutes) 17:05:52 TheJulia: thank you! :) 17:06:01 #link https://etherpad.openstack.org/p/IronicWhiteBoard line 93 17:06:30 Many of the deeply overall application account use related discussions end up mostly being unrelated, i.e. the mechanisms will be available via config drive regardless of what is decided. 17:06:45 o/ 17:07:05 I see 17:07:23 mjturek, do you think you could trim down the BFV section on the whiteboard to only the most important and recent stuff? 17:07:28 it's becoming huuuuge :) 17:07:42 dtantsur: sure :) 17:08:18 heh, good progress 17:08:21 thanks 17:08:59 is there currently any way to test bfv? like with virtualbmc? 17:09:46 hey, what did we decide wrt the tempest plugin? jlvillal, do you know? (L244). who's replacing soliosg? 17:09:53 vdrok: I'm not sure any of us have done an end to end test with BFV. I'm hoping to do that this week 17:10:07 I need to check in the deploy skip wire-in. 17:10:10 mjturek: would be cool! :) 17:10:34 mjturek: I should be able to look at the deploy skip change tomorrow 17:10:50 mjturek: what you mean by end to end test? :) 17:10:51 rloo: I'm going to try to get some time to work on it. 17:11:02 rloo: Tempest plugin that is. 17:11:19 jlvillal: so we can put your name down for it? 17:11:25 yes 17:11:47 jlvillal: done :) (the easy part, heh) 17:11:51 TheJulia: is that going to be separate from the detach/attach wirein patch? 17:12:00 jlvillal, we need to decide on a time frame when we freeze all changes to the in-tree tempest plugin and start the switch 17:12:23 dtantsur: Sort of. Yes. I mean we can start merging the patches we have. 17:12:29 xavierr: sorry, just mean going to see if I can actually boot from a volume with all the pieces we currently have 17:12:33 mjturek: yeah, It is the one I posted that is still flagged as a wip and has no tests 17:12:41 dtantsur: It is just when we get to the end of the process. We have to freeze and then start using the plugin repo 17:12:53 o/ 17:12:56 dtantsur: And then remove from 'ironic' the tempest code 17:12:58 got it, thanks 17:13:07 jlvillal, yeah. the patches that are up now don't have the recent changes and inspector bits. we should agree on the way forward with them, then we can start merging. 17:13:23 dtantsur: Well the patches are one each from the history of the repo. 17:13:44 dtantsur: So have to add more patches for what has changed since the last submission. 17:13:47 dtantsur: If that makes sense. 17:14:08 dtantsur: I need to setup things again with 'git-filter-branch' to get it working for me. 17:14:12 jlvillal, it does. I wonder how to do it technically. 17:14:24 dtantsur: 'git-filter-branch' 17:14:36 jlvillal, does it support going from some commit? 17:14:51 dtantsur: I'll figure that out. 17:15:06 mjturek: nice, ty! :) 17:15:26 jlvillal, yes please. well, we can always to it from scratch, then rebase on top of merged patches. 17:15:41 dtantsur: Will do 17:15:44 thanks! 17:18:06 sambetts, what do we need to start merging changes against networking-baremetal? 17:18:17 I can +2 things, but I don't understand this project mostly 17:19:59 #link https://review.openstack.org/#/c/464046/ discussing the future of the classic drivers 17:20:09 tl;dr let's deprecate them :) 17:22:17 Hmmmm 17:22:46 dtantsur: I'm not sure about that either, but I guess we need to start with that "fake" baremetal driver 17:22:57 re networking-baremetal 17:23:39 ack.. I'm not sure I even understand the difference/relationship between networking-baremetal and networking-generic-switch 17:23:53 which is a sad state to be in, I admit 17:24:17 is everyone still reviewing the statuses? 17:24:17 dtantsur: I'm kind of in the same situation... :( 17:24:54 o/ sorry I'm late... 17:24:56 dtantsur: w/r/t drivers, I kind of feel like that spec may go against the grain of what we agreed upon perviously, but pondering required. 17:25:13 sambetts, vdrok, do we have a defined work group of people who actually understand this thing? is it you two and vsaienko? and mgoddard? 17:25:17 and more reading 17:25:22 dtantsur: AFAIK networking-baremetal is going to hold things that any interaction with ironic would need. networking-generic-switch is just an implementation of ml2 plugin for specifc (dumb) switches 17:25:28 TheJulia, interesting, I don't remember not planning to deprecate them. comments appreciated! 17:25:51 on a side note, vsaienk0 is out for vacation for 2 weeks 17:25:56 good to know 17:25:58 dtantsur: I remember "leave as is, accept no new classic style drivers" 17:26:21 TheJulia: i thought we were going to deprecate at some point. 17:26:34 TheJulia, hah, we're in disagreement here :) let's move it to the spec, I guess. I don't have any strong feeling here, just don't want to leave it hanging in the air. 17:26:36 TheJulia: the 'leave as is', i thought was while we were getting driver composition to work. 17:26:37 dtantsur: the difference is, ngs does things with ovs in case of gate testing, this fake driver only will update the db and that's it :) 17:27:09 db being neutron db 17:27:16 rloo: dtantsur: That might be, I need to context switch back a year + 17:27:34 oh yeah, that lasted quite a while (the driver comp story) 17:27:46 there's also other stuff targeted for networking-baremetal, like the events thing, and api extension to use to update things for routed networks AFAIR 17:27:58 yup, that's correct 17:28:06 thanks pas-ha, this clarifies something 17:28:23 TheJulia: i just skimmed the original spec, under 'Database and Rest API', in the NOTE box, 'we plan on eventually deprecating the classic drivers.' 17:28:28 TheJulia: http://specs.openstack.org/openstack/ironic-specs/specs/7.0/driver-composition-reform.html 17:28:34 rloo: thanks! 17:28:55 these are good discussions to have, but should we wrap up with the statuses first? 17:28:55 are the DOC's up to date with the driver composition stuff? 17:28:56 TheJulia: incentive for moving faster (if only we could). easier to remember the plan :) 17:29:08 NobodyCam, nope, patches are on the whiteboard 17:29:20 * rloo is done with statuses 17:29:21 and I need to write at least one more 17:29:47 moving on? 17:29:59 #topic Deciding on priorities for the coming week 17:30:19 we got rid of one rolling-upgrades-related patch, good 17:30:34 otherwise, I expect folks to start getting back on speed after the Forum is past us 17:31:03 also, the E-tag spec is still close to landing 17:31:10 also, yet another sad anouncement - Galyna (gzholtkevich) has left Mirantis and most probably won't be able to contribute (although she promised to try :) ) 17:31:21 and that's right about e-tags 17:31:27 ouch. please wish her the best of luck 17:31:39 meaning, we need someone to take over e-tag, right? 17:31:47 so I'd take over the patches of hers 17:31:55 yes, that's me 17:32:05 pas-ha: thx for volunteering! 17:32:15 pas-ha, ok, please update the etherpad 17:32:21 will do 17:32:35 folks, I'd like to shamelessly add my driver composition install-guide changes to the priorities 17:32:54 anything that improves our docs +1000 17:33:08 I'm looking forward to no longer working on the driver composition, please help me achieve this goal :D 17:33:09 +++++ 17:33:11 yeah, I'm +1 on that 17:33:50 dtantsur: I'll review the docs after the meeting :) 17:33:53 thanks! 17:33:59 pas-ha: Sorry Galyna isn't there anymore. Met her in Barcelona and she was very nice. And did a lot of good work for Ironic. 17:34:14 it does refactor the docs quite a bit, so it may not be the easiest review in the world 17:34:20 jlvillal++ 17:34:42 dtantsur: you refactored the install guide ANd added docs for driver composition? in one patch? 17:35:14 rloo, kind of. e.g. I was going through enrollment section and making sure it reads well and up-to-date 17:35:28 dtantsur: yeah, tho we need this stuff, as people start asking about how the new things like attach/detach or driver composition work, and we have nothing in the docs :( I'll try to put up some patches too, regarding the new things 17:35:46 I'm open to splitting patches, but I do realize that it will make people review them slower. few of us like reviewing docs. 17:35:46 dtantsur: heh 17:36:03 few like +2ing them :D 17:36:09 true :) 17:36:20 i love +2ing them. i hate reviewing/commenting. 17:36:26 ok, here's the proposed priorities: next rolling upgrades patch, next BFV patch, install-guide driver-comp changes, e-tag spec. thoughts? 17:36:37 works for me 17:36:45 etags before install-guide 17:36:46 yup sounds good 17:36:50 I'd add the tags, if you think there is some space for it 17:36:56 cuz no one is reviewing the etags it seems. 17:37:00 sounds good 17:37:07 and we're just talking about the spec for etags 17:37:23 rloo, I'm following the priorities order on the etherpad, but fine with me 17:37:24 but honestly, i doubt that the order matters. we're going to review what we're going to review 17:37:33 ++ to order does not matter much 17:37:41 tags being node tags 17:38:01 vdrok, all node tags patches need rebasing 17:38:16 oh :( will do today 17:39:06 anything else on it? do we want node tags there anyway? I'm slightly -0 on that. 17:39:12 fine with leaving them out tho :) 17:40:15 ok, let's keep it as it is 17:40:33 anything else on this? 17:40:44 #topic Open discussion 17:41:02 20 minutes for your favorite part of this meeting :D 17:41:26 * rloo favorite part is the crickets 17:41:30 :) 17:41:30 hehe 17:41:31 ok, I have one topic 17:41:57 jlvillal (I think) suggested we have certain number of these meeting per cycle via phone (similar to the recent meetup) 17:41:58 i didn't attend boston summit/forum. is there any feedback from that, that we should be aware of? 17:42:13 rloo, I'm aware of 17:42:15 #link https://etherpad.openstack.org/p/BOS-forum-ironic-feedback 17:42:31 dtantsur: It wasn't just my idea :) 17:42:41 yeah, you were the last to suggest, I think :) 17:42:41 dtantsur: thx, i'll take a look. TheJulia, please let us know if there is anything else :) 17:43:00 i thought there was a suggestion to hold one eg first or last of the month? 17:43:02 rloo, also, our folks told me that ironic scalability was mentioned as a pain point 17:43:17 dtantsur: what, ironic doesn't scale? 17:43:40 apparently not as well as people want it (think, going beyond 1000 nodes per conductor) 17:44:10 dtantsur: sigh. we should make sure we have bugs open with the issues. 17:44:10 oh lots of good reading on that etherpad 17:44:35 hey ironic'ers, I have two patches for solving the some issue 1503855. I'd like your opinion if 464212 or 408298 would fit better 17:44:42 May I get feedback on this bug? https://bugs.launchpad.net/ironic/+bug/1690458 17:44:43 Launchpad bug 1690458 in Ironic "RFE for Ironic disk cleaning to ensure removal of disk labels (not just partitions)" [Undecided,New] 17:45:03 dtantsur: Yeah, I think once a month for an audio meeting would be good. 17:45:04 fultonj, was looking at it today, but did not have time to dive into man pages 17:45:16 * dtantsur adds to his "take a look" list 17:45:30 Expectation differences, because we've failed to have reference documentation on what a setup should/could look like, and theoretical numbers for what a single conductor/api can do in that setup. 17:45:39 ok, thanks dtantsur i'll be on IRC if you have questions 17:45:48 fultonj: https://review.openstack.org/#/c/424291/7 while for different driver, this kind of logic which is needed IMO 17:45:54 TheJulia, yes, this reference architecture guide, which we still hope to provide 17:45:57 Writing that documentation, even just to provide the project's context would be helpful and help us obtain clear requirements 17:46:18 TheJulia: ++ 17:46:28 jlvillal, next question would be: do we want it to be 1 hour (like this meeting) or more (like meetups)? 17:46:29 I'd like to discuss rescue mode. What is the best way to move it forward ? 17:46:43 dtantsur: I think aim for 1 hour time slot, same as this meeting. 17:46:56 izumi777, keep code patches clean and working, address comments on timely fashion 17:46:56 dtantsur: Same bat time, same bat channel 17:46:58 jlvillal, dtantsur: we could try. although i was thinking about that wrt the tc moving away from having weekly irc meetings :) 17:47:26 rloo, the TC has a slightly different problem: they have to encompass opinions of the whole community. which is barely possible with a fixed time slot. 17:47:36 we're slightly smaller 17:47:50 we also don't have a great track of working with ML, to be honest.. 17:47:54 dtantsur: different time zones, language barrier 17:48:14 true. I'm always for using ML more for making decisions. 17:48:19 dtansur, thanks. Can I take over rescue mode patchset ? 17:48:20 I vote for one meeting per month as an audio bridge. 17:48:30 jlvillal: +1 17:48:31 dtantsur: like i said, we could try. i think for the normal meetings, irc seems fine. audio is useful for discussions i think. 17:48:31 We could try it for a few months and see what people think. 17:48:36 i am fearful that moving away from irc meeting will result in loss of context for folks who are unable to attend a call in meeting.. unless good notes are taken and posted 17:48:52 speaking of which, I will bring the phone meetup question to the ML. I'd reserver 1.5 - 2 hours to account for bad sound, technical issues, etc. 17:48:52 NobodyCam: Super good point, and I agree 17:49:14 oh... the good thing with irc meeting is that it is all recorded. 17:49:14 NobodyCam, good notes is a requirement. and a summary afterwards. not sure I did great job with it the last time, but I can try better :) 17:49:45 dtantsur: someone besdies the person running the meeting hsould be taking notes. 17:49:47 rloo, well.. we can try my bluejeans, it has "recording" feature. I don't know how many people it can survive and other potential caveats though.. 17:49:48 izumi777: i think rescue mode is being worked on by stendulker & aparnav. i suggest you get in touch with them to see how you can all work on it together. 17:49:50 s/hs/sh/ 17:50:00 izumi777: definitely reviews would be great 17:50:10 TheJulia, true. we should probably find volunteers for that in advance. 17:51:40 rloo: okay. I'll contact them. Can we merge it in Pike-cycle, cant' we ? 17:52:07 are we discussing having an audio meeting for the weekly ironic meeting. or is the audio for weekly ironic meeting + more stuff? 17:52:10 izumi777: hrm, I have doubts really, there are loads of features and not that many reviewers left 17:52:23 izumi777: always possible to merge in pike cycle, if it is ready and reviewed and approved. 17:53:08 vdrok: yes, probability might be low(er), but it is possible 17:53:16 rloo, this is the question as well. for now we were discussing replacing some of weekly IRC meetings with voice ones. 17:53:36 dtantsur: if replacing, i think we should keep it to 1 hour. 17:53:55 +1 on one hour 17:53:57 yup, more than one would be too late for me (10 pm) 17:54:00 izumi777, we'd like to have rescue, but it's in the process of being taken over by a new team of developers. please get it touch with them, I'm pretty sure they'll need help. 17:54:05 dtantsur: let's just try one of them to see how it goes. 17:54:18 ok, thanks for feedback re voice. ++ to just try. May 29th? 17:54:26 To me just replace an IRC meeting with an audio meeting at same time for same amount of time. 17:54:37 may 29 sounds good to me 17:54:52 29th sounds good 17:54:53 umm, that's memorial day in the US 17:54:55 Not for me. I'll be on vacation in Moscow then :) 17:54:58 jlvillal: do you work that day? 17:55:03 * TheJulia looks at her calendar of doom 17:55:03 jlvillal: oh, you aren't around. 17:55:08 rloo: It is a holiday too 17:55:20 well, what's the use of having this w/o jlvillal :) 17:55:21 29th works for me 17:55:24 dtansur, vdrok: OK. I can contribute by testing and reviewing 17:55:43 mmm, excluding the whole US is not good 17:55:45 rloo: heh 17:55:45 izumi777: cool, thank you for that! 17:55:55 next Monday? or June 5th? 17:56:00 (next = May 22nd) 17:56:02 dtantsur: How about first meeting of the month is audio :) 17:56:05 Next monday is a no-go for me 17:56:15 next monday is a holiday for me :) 17:56:22 ok, June 5th? 17:56:25 I'm fine with any monday for the next 2 months :D 17:56:36 heh 17:56:42 jlvillal: you avail june 5? 17:56:53 rloo: I'll be on vacation :) 17:57:06 I think it will go fine without me though. 17:57:21 who else wanted the audio besides jlvillal? 17:57:23 I'm available June 5th 17:57:26 Though I know everyone will miss the dulcet tones of my voice 17:57:32 rloo, well, I wanted 17:57:36 :D 17:57:43 dtantsur: ah, ok then :) 17:57:55 I'd like to see if it can get more productive that way 17:58:03 * rloo wondered who owned that dulcet voice. 17:58:03 I really don't know, hence I wonder 17:58:09 2 minutes left... 17:58:19 dtantsur: pauses will be shorter for sure :) 17:58:23 1 minute... 17:58:24 Hi everybody... it would be great for us oneview folks to have some word on https://review.openstack.org/#/c/464212/ if you have some time :) 17:58:29 worth trying :) 17:58:31 ok, I'll propose it on the ML. I wonder if we still should have full-featured voice meetups in addition to that 17:59:08 thanks everyone! 17:59:08 ricardoas, will check (tomorrow apparently) 17:59:09 dtantsur: I could see at least once per cycle an actual big time block. Like we did last time. 17:59:13 and thanks everyone! 17:59:15 Thank you everyone! 17:59:20 jlvillal, I'd even do a few times 17:59:20 thanks all 17:59:22 Bye 17:59:24 see you 17:59:24 o/ 17:59:25 dtantsur: ++ 17:59:25 thx 17:59:36 #endmeeting ironic