17:00:02 #startmeeting ironic 17:00:02 o/ 17:00:02 Meeting started Mon Apr 10 17:00:02 2017 UTC and is due to finish in 60 minutes. The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:03 o/ 17:00:04 o/ 17:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:06 The meeting name has been set to 'ironic' 17:00:07 o/ 17:00:09 o/ 17:00:14 o/ 17:00:18 \o 17:00:20 Hi everyone! 17:00:25 howdy 17:00:35 o/ 17:01:00 our agenda (quite light) is as usual at: 17:01:07 #link https://wiki.openstack.org/wiki/Meetings/Ironic 17:01:11 o/ 17:01:33 #topic Announcements / Reminders 17:01:56 I/ 17:02:00 #info Boston Forum schedule was published 17:02:08 \o 17:02:08 (with some mistakes to be corrected, apparently) 17:02:13 o/ o/ 17:02:16 * dtantsur does not have a link at hand 17:02:29 o/ 17:02:30 o/ 17:02:53 I don't have other announcements. Anyone? 17:03:05 ah, hmm 17:03:38 #info dtantsur is on PTO starting from Fri, Apr 14. Back on Mon, Apr 24 17:03:42 yes, again :) 17:03:48 woohoo 17:03:51 :) 17:03:58 dtantsur: wow, lucky you! 17:03:59 I think a few more folks are going to be out the same week, right? TheJulia at least 17:04:01 I'll be out tues-thurs this week, btw 17:04:05 julia as well 17:04:12 (leadership training) 17:04:23 o/ 17:04:29 thanks! let's keep it in mind when defining priorities for the coming weeks 17:04:53 #info a suggestion for a virtual meetup: https://etherpad.openstack.org/p/ironic-virtual-meetup 17:04:54 jroll: would you chair next Mon's meeting? 17:05:02 jroll: please :) 17:05:22 rloo: sure, why not 17:05:27 thanks jroll! 17:05:31 * rloo going to see if there is any diff with jroll's leadership... 17:05:54 3 cycles of history not enough? :P 17:06:03 hah, rloo is already preparing for the next elections :) 17:06:07 jroll: but the leadership course...!! 17:06:12 oh, that 17:06:18 ah, yeah. fair :) 17:06:23 I've been once already, I'm helping connect dots 17:06:39 :) 17:06:42 please have one in EU, and I'll attend. otherwise you'll have to live with my terrible leadership for some more time :-P 17:07:03 anyway, another request to take a look at the virtual meetup proposal: https://etherpad.openstack.org/p/ironic-virtual-meetup 17:07:16 wrt virtual meetup. with dtantsur and TheJulia gone next week. Am wondering whether we should punt it another week. 17:07:26 rloo, it's proposed for week of Apr 24 17:07:31 o/ 17:07:41 as it was suggested on the previous meeting 17:07:44 dtantsur: right, you and TheJulia will have just come back. will you be prepared to participate? 17:07:53 I will be 17:07:59 * TheJulia re-engages the cloaking device 17:08:00 always prepared :) I'm not going anywhere, just hanging out here in Brno 17:08:17 trying to reduce stress by examining the nearby bars 17:08:36 dtantsur: That sounds like an awesome plan 17:08:37 dtantsur, +2000 17:08:53 we can get back to the meetup topic in the open discussion. any other announcements? 17:09:26 * milan announces lack of sleep :P 17:09:33 :) 17:09:44 #topic Review subteam status reports (capped at ten minutes) 17:09:54 #link https://etherpad.openstack.org/p/IronicWhiteBoard 17:10:02 starting with line 75 17:10:15 (oh, was it last week that we did an ironic release?) 17:10:50 yeah, probably 17:10:58 yes, april 5 17:11:04 8.0.0 specifically 17:11:26 Only three more major releases and we'll be better than windows! /s 17:11:41 JayF: huh, aren't we better already? 17:11:52 rloo: Windows is 10, we're only 8! 17:12:02 Firefox is 52.. 17:12:07 lol 17:12:08 JayF: but we're better, regardless ;) 17:12:24 rloo, agreed 17:12:31 I was joking, obviously :) 17:12:40 "jroll promises to look at this week of march 27" I wonder if he did ;) 17:12:45 (line 154) 17:12:56 damn, no I did not 17:13:18 * jroll bumps on todo list, sorry 17:14:27 rolling upgrades: "make a change to the grenade job to only upgrade conductor" (from the priority list) 17:14:29 was it done? 17:14:33 jroll: related to last week and these priorities. you were going to review the nova stuff to see if there is anything we need to do b4 nova's feature proposal freeze 17:15:09 rloo: yes, I instead did paperwork all week. I've got it docketed for today 17:15:28 * jroll hasn't had much time for upstream, sorry 17:15:29 vdrok: ^^ do you know what the status is with multi-node grenade testing changes? 17:15:42 jroll: ok, i'll try to remember to ask next week :) 17:17:22 2 more minutes to review the statuses 17:19:17 done reviewing? 17:19:28 yuppers 17:19:44 #topic Deciding on priorities for the coming week 17:20:02 I'm inclined to leave it as it is now (with moving BFV to next patches) 17:20:05 what are your thoughts? 17:20:16 fine with me 17:20:35 +1 17:20:40 + 17:20:43 ++ 17:20:54 ++ 17:21:04 good 17:21:27 before we move on. who was leading the bug triage previously and who wants to? 17:21:50 * dtantsur thinks we should also start fixing bugs, not only triaging.. 17:22:22 * jroll looks back 17:22:24 I don't know who was doing it before, but I can take it this week if there are no volutneers 17:22:34 awesome! thanks JayF 17:22:43 #info JayF to help with bug triaging this week 17:22:52 I think we missed it last week 17:23:08 was TheJulia and vdrok previously 17:23:12 yeah, my bad. was just out of PTO :) 17:23:31 #topic #Trello fate 17:23:36 wut? 17:23:38 #undo 17:23:39 Removing item from minutes: #topic #Trello fate 17:23:45 #topic Trello fate 17:23:56 looks good now :) rloo, your turn 17:24:15 well, dtantsur brought it up at a previous meeting. we had been using trello. do we want to continue? 17:24:16 * jroll pulls out a knife 17:24:26 * jroll waits for the kill word 17:24:30 trello hasn't been updated so if no one will volunteer to use it, i think we should NOT 17:24:35 I'm very pleased with how we track things on the etherpad 17:24:38 i mean, volunteer to update it 17:24:57 so might be easiest to ask: does anyone want to continue using trello? 17:25:11 we've been asking about it for weeks, nobody has stood up and said they want it, afaik 17:25:38 I think somebody wanted it on the previous meeting.. but I'm not sure 17:25:44 ok, let's get rid of it. fwiw, the etherpad doesn't cover stuff that used to be i trello. trello had links to nova patches and ironic patches. 17:26:02 dtantsur: i wanted it cuz i found it useful, but only if it is keep up to date. 17:26:03 rloo, 1-2 gerrit topics should cover it 17:26:06 any reason the etherpad doesn't? I feel like it should 17:26:33 no reason tht the etherpad cannot. i'm just not sure it does have that info. 17:26:38 let's just add all related topics to subteam items on the etherpad 17:26:57 anyone wants to? otherwise I can do it 17:27:51 #agreed abandon our Trello board in favor of the WhiteBoard 17:28:24 dtantsur: while you're at it, i'm not sure cuz it has been a while, but trello had some projects that weren't completed that I don't think are in our priorities. if taht's the case, would you please make sure there are bugs open for those? 17:28:32 rloo, sure 17:28:41 thx dtantsur 17:28:51 #action dtantsur to move all useful information from Trello to the WhiteBoard or Launchpad (for non-priorities) 17:28:57 anything else here? 17:29:27 ok, before we jump into open discussions 17:29:32 #topic Virtual meetup 17:29:43 #link https://etherpad.openstack.org/p/ironic-virtual-meetup 17:30:03 if it does happen on the week of Apr 24, I need to organize it *this* week. or I need somebody to organize it the next week. 17:30:29 what all needs to be organized? time/place/? 17:30:35 I suggest to start with a doodle for specific dates 17:30:54 then getting a conference number (or how was it called?) 17:31:09 actually, we can try using my bluejeans, they work with latest FF and Chrome without plugins 17:31:19 but I can't promise :) so we'd better have SIP as a backup 17:31:41 and finally, to announce the details on the ML 17:31:45 i thought we used something that openstack provided. that stopped working during the meetup 17:31:53 ++ to SIP backup! 17:32:02 * jroll wonders if there's anyone that will refuse to / cannot use bluejeans 17:32:08 rloo, this is what I just call SIP 17:32:16 dtantsur: ah. 17:32:19 jroll, it probably involves non-free javascript, so there may be 17:32:34 dtantsur: right, also curious if it's reachable from china etc 17:32:39 oh yeah 17:32:54 I think I had meetings with folks from China, but the point is still valid 17:32:55 anyway, openstack SIP is always open, so that's an easy backup 17:32:57 I'll say that personally, I find it extremely difficult to understand what folks are saying without an accompanying video to lip read from 17:33:06 but I understand if folks freedom-concerns prevent that 17:33:13 JayF++ 17:33:22 fwiw I think you can call into bluejeans from a phone 17:33:27 well, i guess that's why we have PTG 17:33:35 which is of course completely free and open source ;) 17:33:43 (the phone, I mean) 17:33:45 hah 17:34:03 do we have an idea of how many days/hours? 17:34:09 jroll, should we reserve something in OpenStack SIP for specific days? 17:34:22 dtantsur: there's no reservation system, afaik 17:34:26 rloo, so, my idea was: two slots in one day to accommodate different timezones 17:34:31 similar to the midcycle 17:34:49 dtantsur: this would be the closest thing to reservations https://wiki.openstack.org/wiki/VirtualSprints 17:35:10 oh my, this was the previous midcycle: https://etherpad.openstack.org/p/ironic-newton-midcycle 17:35:14 so, back to planning, ideally if I can create the doodle this week, the next week somebody makes it official 17:35:34 but I can rush everything in this week, if there are no volunteers 17:36:22 * jroll doesn't want to promise anything right now, hopes someone steps up to coordinate 17:36:23 #link https://wiki.openstack.org/wiki/VirtualSprints 17:36:50 do we really want a midcycle? i'm half joking. just looking at the previous one and what we have gotten accomplished since then... 17:37:02 git clone https://git.openstack.org/openstack-dev/devstack.git devstack 17:37:07 whoops, sorry 17:37:19 lol 17:37:20 mariojv, access denied ;) 17:37:25 am wondering whether we should 'just' have days/times, regardless of what week, to *do* something 17:37:46 * jroll wonders if rloo is implying we haven't done much since, or have and so we don't need them 17:37:49 rloo, I was thinking into breaking the midcycle into such smaller sprints. and I want to have the first happening before I can suggest do it regularly 17:38:03 i mean, just pick feature X, prepare, and meet at and do it. 17:38:35 planning is hard though: timezones, other meeting, etc 17:39:06 if sprints work for folks then that is fine. i was just wondering if it might work better if we just say, eg. rescue is there, we want to just sit down and go over it. or spec Foo is there, lets just meet and go over it... 17:39:29 so eg, etags spec looks like it is almost there. let's just meet this week at .. and get it done. 17:39:51 this is a bit different think, I suspect 17:40:06 in your etags example, we only need cores and the author 17:40:16 rloo: something like regular spec review meeting with specific timeslot and list of specs? 17:40:16 this can indeed just happen one day this week 17:40:18 I like rloo's example, honestly 17:40:35 dtantsur: well, i was looking at your #2 in the meetup. what are we going to accomplish there? 17:40:40 and it mirrors some stuff we've done before with some features, i.e. a couple of weeks ago mario, annie, anup and I met and worked on rescue for a couple of hours 17:40:55 rloo, something close to what you suggest, I guess 17:41:44 I believe we had a similar conversation when dtantsur originally proposed this. and agreed we should do targeted meetups like rloo mentioned, *and* occassional syncups like dtantsur proposed. am I remembering this wrong/ 17:41:49 dtantsur: of those things listed in the meetup, i would be only interested in #3. 17:41:49 s/\//?/ 17:42:08 jroll, I thought the same 17:42:17 dtantsur: i think it'd be great if we could do 1, 2 in a diff way. 17:42:28 dtantsur: i was hoping that our weekly meetings could address 1 17:42:49 rloo, my point is, such meetup can collect bigger audience (e.g. our QE got interested when I talked about my proposal last week) 17:43:05 for the benefit of people who don't attend every meeting we have 17:43:23 3rdparty CI conversation will attract specific audience who is not necessary here 17:43:32 but dunno. I agree with jroll, we should have both 17:43:47 and I think targeted meetups do not have to be driven by the PTL ;) 17:43:54 ++ 17:44:08 if people aren't interested in the 'midcycle' topics, they don't need to join 17:44:09 dtantsur: i'm fine if others want to have this meetup and will get something from it. and i agree the 3rd party ci will be a useful meeting to have. 17:44:12 let's give it a shot 17:44:41 ok, so I'll start with a doodle 17:44:56 (or those people can add more topics, of course) :) 17:45:00 as I don't see volunteers for organizing it, the deadline for the doodle will be Friday 17:45:24 dtantsur: thanks for doing that 17:45:50 we'll use SIP, but we can also experiment with Bluejeans, if folks will feel like. co-ordination through our regular channel. any other thoughts? 17:46:54 I'm ready to move on. And you? :) 17:47:39 * jroll is 17:47:47 #topic Open discussion 17:47:53 hey, I've got an issue to share regarding oneview drivers :) 17:48:04 sure 17:48:06 hey ironic, we from oneview have a patch where we are discussing about non-persistent boot in our drivers we want your opinion if our idea for implement that is fair. suggestions are welcome! link: https://review.openstack.org/#/c/436469/ 17:48:07 we'd love to replace python-oneviewclient (and deprecated it) by hpOneView, it's been more actively maintained and supporting newer oneview api versions and we are already using it at oneview ml2 driver 17:48:29 want opinion on whether to show driver 'type' column in https://review.openstack.org/#/c/419274/ 17:48:53 rloo, in driver-list or driver-show? 17:49:00 but we need some advice on what to do with specific validations like the ones at https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/oneview/common.py#L150 17:49:09 rloo: oh, this is that microversion inconsistency topic? 17:49:19 ricardoas, what's the problem with that? 17:49:19 our best guess is to move it upstream, wdyt? 17:49:28 mariojv: yeah. 17:49:41 ricardoas, "move it upstream" gets my support any day :) 17:50:06 One item here. Wonder if we could bring forward 'move ironic tempest plugin to the new repository' task 17:50:09 dtantsur, i've talked with some of you during the week and it was suggested to raise it here... so just checking if it is ok to move this validations upstream 17:50:11 dtantsur: in driver_list 17:50:29 dtantsur: \o/ 17:50:35 rloo: i think it should be consistent with what driver show does 17:50:36 I did some initial work, porting patches from ironic/ironic_tempest_plugin to the new repo 17:50:49 dtantsur: in driver_show, it will show 'type' with no value, but it is in a row. in driver_list, 'type' will be a column with nothing in it. 17:50:54 rloo, hmm, in list. I suspect the answer is "no, because of microversions", right? 17:50:56 right 17:51:12 rloo, I suggest we don't show it. I hope the day will come when all drivers are new-style. 17:51:15 rloo: but if it's a huge refactor to get it hidden or something, i might be ok with the small inconsistency 17:51:43 They're still waiting for review. I think cores could review & approve. The real review would be when making changes to project config and devstack-gate 17:51:47 so the right way, is not to show info that isn't avail in a microversion but we didn't do that for information. to date, that info is usually some field, so it shows up as a row in eg -show. 17:51:48 soliosg, thanks! this is one thing I constantly forget about. We have to freeze our tempest plugin and start merging the patches. 17:52:01 this list one is the first time (I think) the info is a column. 17:52:27 also here are some links illustrating the issue in case anyone has trouble visualizing this: http://paste.openstack.org/show/603982/ vs http://paste.openstack.org/show/603983/ 17:52:32 soliosg, with project-config and devstack-gate, I think you can rely on help of #openstack-qa and mtreinish personally :) they are very keen on projects moving to branchless plugins. 17:52:58 regarding this patch that xavierr pasted... i think we should just store something at the node internal info to warn the boot change function that the next operation will be non persistent. Any better idea, folks? :D 17:53:12 dtantsur: yes, also jlvillal_pto has provided great support! 17:53:13 rloo, why do we need to show this column at all? 17:53:34 dstanek: heh, actually both project-config and devstack-gate are the domain of -infra. I'll gladly look and review patches, but I don't have the big rubber stamp on either :) 17:53:43 dtantsur: well, that is a good question. it isn't a need. do we want to or not? 17:53:56 dtantsur: i suppose we don't, but then we'd be not exposing something potentially useful returned by the api 17:53:57 mtreinish, yeah, I specifically mean "look and review" part 17:54:17 dtantsur: with the driver stuff, i think it is in the spec that driver-list shows the type. 17:54:23 rloo, mariojv, I hope one day all drivers are new-style, and this "Type" column will be always-true 17:54:31 oh, i see your point 17:54:43 even worse, if we remove it in one microversion, then this column will have to go away too.. 17:55:01 rloo, how many times have we amended this spec already? one more won't hurt :) 17:55:29 dtantsur: I'm fine with amending the spec. we just need to decide. (and i wanted a decision before i reviewed more). 17:55:38 dtantsur: keep in mind, the API is returning 'type' . 17:55:55 yeah, but we don't have to expose everything in a list CLI call 17:56:10 especially when we have reasons to believe that it will become redundant one day 17:56:26 dtantsur: that is true. i'm fine with that. so no type in CLI except in show or if we add 'filter'ing on driver-list... 17:56:55 dtantsur: if there is some time left, i'd like to talk about this patch too: https://review.openstack.org/#/c/408298/ 17:57:10 i think that's reasonable for now 17:57:17 dtantsur: or if driver-list --detail is specified. i think we have --detail... 17:57:18 (and everyone else...) 17:57:42 rloo, ++ 17:57:49 *Two (2) minutes to go* 17:58:37 dtantsur: ok, no one is in disagreement, so i'll convey the news to the author of the patch :) 17:58:58 rloo, thanks! good point to raise indeed, the author of the driver composition spec clearly did not think about it ;) 17:59:17 I suggest we move the discussion of the specific patches to the channel, if you don't mind 17:59:21 ricardoas, ^^^ 17:59:24 dtantsur: can't expect the author of the spec to think about everything. but we're close to the finish line now :) 17:59:25 ok :) 17:59:33 no problem 17:59:51 #agreed Do not expose "Type" column in the output of driver-list without --detail 17:59:58 thanks everyone! 18:00:02 ty 18:00:04 thanks :) 18:00:04 thanks all 18:00:07 o/ 18:00:08 thanks! 18:00:10 #endmeeting