15:00:56 #startmeeting ironic 15:00:57 Meeting started Mon Aug 12 15:00:56 2019 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:57 o/ 15:00:58 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:01 The meeting name has been set to 'ironic' 15:01:14 o/ 15:01:17 o/ 15:01:20 o/ 15:01:20 Greetings everyone, it has been a whie since we've had a meeting! 15:01:26 o/ 15:02:19 So for today, I'd just kind of like to hit the main points on our standard agenda to kind of sync up, and when we get to discussion, I'd love to hear feedback on the office hours concept 15:02:30 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:02:41 #topic Annoucements/Reminders 15:03:23 \o 15:03:26 o/ 15:03:43 #info This week is R-9 to release. 4 weeks until things begin to freeze around the openstack community. 15:03:43 o/ 15:03:52 ouch 15:03:55 Does anyone have anything else they wish to announce. 15:04:02 dtantsur: ouch++ 15:04:36 R-5 includes requirements freezes 15:04:47 R-3 is when we'll likely have to cut our stable branch. 15:04:54 #link https://releases.openstack.org/train/schedule.html 15:05:00 Anyway, moving on! 15:05:31 I don't think we had any action items since it has been a while, so we can jump to reviewing the etherpad and going through there. 15:05:37 #topic Review subteam status reports 15:05:56 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:06:13 Starting at line 269 15:06:46 Looks like stendukler would be the one to update deploy_steps 15:07:40 Merged openstack/sushy master: Unify OEM Actions with non-OEM Actions https://review.opendev.org/674882 15:07:51 As for improving deployment speeds, I've had zero time to look at that 15:08:31 I may be on a call later today regarding kexec, so maybe I can help rally some input or even contributions in that area. 15:08:56 without sidetracking, kexec is a huuuuuuge ball of wax :) 15:09:09 jroll: ++++++ 15:09:28 indeed! But I know some people super interested... something about a half terabyte of ram.... 15:09:35 heh 15:09:45 I've seen people with several terabytes :) 15:09:54 well, yeah, I just expect them to be disappointed 15:10:10 dtantsur: The moving away from wsme hope, are there any updates there that your aware of? 15:10:15 speaking of deploy steps, we need someone to work on in-band deploy steps 15:10:34 TheJulia: some activity in gerrit, but nothing ground-breaking. I may not have time this cycle.. 15:10:56 dtantsur: If there are any links, please update the etherpad with them :) 15:11:13 I saw a patch targeting replacing wsme today 15:11:15 * TheJulia votes we rename this "The cycle of being pulled in 5001 different directions" 15:11:34 s/cycle/career/? 15:11:42 heh 15:11:49 kaifeng: its commit summary is a bit of an overstatement though. it's just one step further. 15:11:56 etingof: When you get a few, can you update the section for redfish on https://etherpad.openstack.org/p/IronicWhiteBoard 15:11:59 etingof: line 331 15:12:00 jroll: LOL 15:12:38 For those that are not aware, looks like we're super close on redfish support. Support for sushy-tools merged 15:12:59 \o/ 15:13:19 do we track the ipa-builder work somewhere? 15:13:20 TheJulia, I will update that, have not looked into the doc for a long time 15:13:28 We're basically at reviews for the redfish virtual media boot interface 15:13:34 dtantsur: no, but lets! 15:13:41 FYI we're pretty close to having a DIB built image in our CI 15:13:45 okie 15:13:56 dtantsur: thanks! 15:15:00 * etingof has to reshuffle vmedia patch to merge cleanly with recently introduced iconic.conf based kernel/ramdisk config 15:15:14 Looks like some redfish sensor stuff merged \o/ 15:15:16 * kaifeng guess we'll need to raise some ram for using dib images 15:15:49 kaifeng: totally. it will be closer to the deprecated coreos images 15:16:05 For software raid! Dmitry has been working so we had CI testing of it. And people are finding/fixing issues! \o/ 15:16:06 but unlike tinyipa (and like coreos) these images will be production-ready 15:16:11 \o/ 15:16:25 \o/ 15:16:34 tssurya's work on the power sync is well advanced and seems to have the green light from the nova team now, would be great if we could finish this off 15:16:41 Power state changes will need reviews this week. Nova folks want the ironic portion to merge first. 15:16:48 :) 15:16:52 jinx? 15:16:52 check this out: software RAID job passed: https://logs.opendev.org/02/675102/1/check/ironic-standalone/5e10622/testr_results.html.gz 15:16:56 \o/ 15:16:59 \o/ 15:17:03 nice! 15:17:03 nice :) 15:17:05 This feels like a very \o/ meeting 15:17:06 reviews welcome 15:17:17 Maybe we need less often meetings.... 15:17:25 great! 15:17:37 the only problem is that cirros doesn't seem to boot on md devices, so we have to stop the test on getting to ACTIVE 15:17:48 we can try to solve it in a follow-up 15:18:10 do you think we can get the power sync patch in this week? anyone able to have a look? 15:18:14 v6 related, I've passed that off to dtantsur at the moment because of ENOTIME. We're not in good shape with our tempest stuffs, but things do fundamentally work under the hood. The issues is testing it all in CI. 15:18:19 arne_wiebalck: I'll try to get to it 15:18:28 awesom, thx dtantsur ! 15:18:37 TheJulia: my plan is to try bypassing tempest for now, e.g. by using the metalsmith CI layout 15:18:46 (which also bypasses nova, but uses neutron) 15:18:58 (this is re V6 now, not re sw raid) 15:20:02 dtantsur: I feel like that is totally acceptable and maybe we try and tackle the tempest headache next cycle.... 15:20:22 indeed 15:20:40 kaifeng: any progress with doc generation, I went back and tried to raise the image concern and didn't really get any responses :\ 15:21:04 no progress, and it seems the infra will do it in a consistent way 15:21:46 folks, I forgot to say: I'll be out starting this Wednesday and till next Tuesday 15:22:00 some hiking new Zürich :) 15:22:03 * near 15:22:28 dtantsur, will meet plenty of cows 15:22:33 I'm good with the updates, we do need to do some cleanup on that list, so if everyone is good with moving on? 15:22:49 let's 15:22:58 ++ 15:23:00 ++ 15:23:06 Excellent 15:23:16 #topic Deciding on priorities for the next week 15:24:10 So the weekly priorities list has become a bit of a advertisement board and notes for things that need visibility or reviews that we as a community have been updating/adding/removing. I _think_ we should continue this and maybe rename the section? 15:24:32 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:24:34 Line 130 15:24:39 I kind of liked having priorities 15:24:50 or at least things that we are ready to pile on 15:25:05 if we stop listing wip stuff, I think that means we can pile on it 15:26:11 "are ready" also implies actually having enough people who are ready to review and the authors who are ready to iterate 15:27:12 I am not sure "ready to iterate" is ever really a thing ;) 15:27:48 BUT, we for really minor things core reviewers do tend to do minor quick fixes 15:28:51 Anyway, I should remove merged things, I did add "things that look ready" 15:28:56 earlier this morning that is 15:29:46 if we're not ready to talk about node retirement, we should probably remove it as well 15:30:15 just did 15:30:23 * TheJulia removes the other super-wip item 15:31:33 So I think that is good for now 15:31:40 * arne_wiebalck will pick up node retirement at some point 15:31:55 "use openstacksdk to connect to ironic" probably pretty WIP as well. rpittau? 15:32:07 dtantsur: yeah, I was going to remove it myself 15:32:14 * arne_wiebalck said this multiple times already 15:32:15 TheJulia: and are you getting to fast-track test any soon? 15:32:18 arne_wiebalck: :) 15:32:52 should we have some more brain power to spare, there is also a review-hungry dhcp-less boot spec... -- https://review.opendev.org/#/c/672780/ 15:32:52 patch 672780 - ironic-specs - Update L3 based deployment spec - 4 patch sets 15:34:04 dtantsur: removed it, I think 15:34:25 etingof: this is a good candidate for adding 15:34:58 adding 15:35:30 someone beat me to it :) 15:35:43 Okay, Are we good to move to open discussion next? 15:35:44 ;-) 15:36:18 * TheJulia is going to take silence as yes 15:36:39 ++ 15:36:44 yep 15:36:51 #topic Open Discussion 15:37:07 So, how does everyone feel about meetings, not having meetings, etc? 15:37:25 I feel like I've been slacking, but also I've been insanely busy since the Summit 15:37:48 i prefer to have them honestly 15:37:52 qq: is the ironic CLI command deprecated in train? 15:38:05 I feel like lost my mind 15:38:11 cdearborn: it has been for a while; we're supposed to remove it in train 15:38:18 kaifeng: Cool! We can start a club! 15:38:23 dtantsur, thx! 15:38:26 I like having meetings, they're just not APAC-friendly.. 15:38:35 dtantsur: ++ 15:38:35 but otherwise we're losing a helpful point of sync-up 15:38:37 I like having meetings as well 15:39:01 I didn't see (nor write) much on the ML instead 15:39:22 maybe we should do a better job doing a human-friendly summary on the ML? 15:39:27 mgoddard: Yeah, I feel like we're all more IRC type folks than email folks :( 15:39:49 TheJulia: +1 - it's a nice active channel 15:39:53 it gets back to doing 5001 things in parallel. IRC is harder to ignore or delay... 15:40:13 dtantsur: I'm kind of worried about trying to add another thing to an already overflowing plate and making it 5002 things to do 15:40:22 indeed 15:40:30 we do have logs of meetings though, and the etherpad should also help, if it's updated consistently 15:40:36 meeting logs are available 15:40:38 I think the scheduled block of time helps the busy people ensure they make time for it 15:40:41 in the API SIG we used etherpad to collaborate on the weekly newsletter 15:40:49 jroll: +1 15:40:50 so that it's not one person's job 15:40:56 jroll: +2 15:40:58 I don't want to prefer meetings, but I think I prefer meetings, else we just wander 15:40:58 dtantsur: I actually already have a newsletter etherpad someplace 15:41:14 jroll: very good point, which is also why I don't want to move to every other week or monthly :\ 15:41:34 we may want to have virtual midcycles at a more APAC-friendly time 15:41:37 jroll: Those words are super true 15:41:51 it's easier to me to allocate 6pm-10pm once for 2 days rather than every week 15:41:57 TheJulia: I don't think every other week would be too bad 15:42:04 dtantsur: that is also super true 15:42:50 What does everyone think about holding a formal meeting every two weeks and still trying to essentially say "hey, these are generally the best times to find contributors"? 15:42:53 if the etherpad/status was kept 'up to date', it would be easier to know the status. we're just bad at updating until 3 minutes into the meeting. 15:43:17 TheJulia: every other week may exclude folks who cannot attend regularly. so dunno. 15:43:30 I feel like it is a bit of a no-win situation 15:43:43 it is indeed 15:43:48 kobayashi maru is right in front of us 15:43:59 sorry, the only reason I think every other week would be useful, is to also have an every other week meeting for apac folks 15:44:08 otherwise let's just keep weekly 15:44:23 this ^^ is good, assuming we can have enough people to show up at apac-friendly time 15:44:32 (which I'm unlikely to do) 15:44:33 we were fine with the weekly meeting. until we tried to accommodate folks that weren't avail at this time. and we still don't know who those folks are and their preferred method of communication. except for one person... 15:44:35 right 15:45:13 rloo: our team's got a guy from NZ, but he's mostly busy with ironic-in-tripleo things 15:45:35 dtantsur: although, I'm sure he would attend now, so maybe that might be the catalyst 15:45:56 TheJulia: I guess the core thing is to make sure you can attend that other meeting 15:46:20 may be easier for kaifeng as well 15:46:35 but half of people present seems to be from Europe 15:47:13 maybe we should start a poll among cores who can potentially make it for a time? 15:47:14 So lets try alternating, and I'll just pick a time slot for next week that seems apac friendly and maybe go from there? 15:47:37 so that we know that at least some cores are present? 15:47:40 dtantsur: I already created one, minimal results and no good overlap 15:47:44 I have applied the irc access weeks ago, so it's convenient for me keeping on the irc than before 15:47:46 oh, for just cores. hmmm 15:47:51 That is a possibility 15:48:04 TheJulia: yep, I mean the cores personally. without cores a meeting won't happen anyway. 15:48:10 I can send that out. 15:48:15 if i recall, the reason why the alternate time was a bust many years ago, was because there were very few cores (if any) that attended. 15:48:16 Merged openstack/ironic master: Asynchronous out of band deploy steps fails to execute https://review.opendev.org/674268 15:48:49 rloo: Yeah, I think it just never happened for a like a month and a half 15:48:59 rloo: that's why we need to find a slot that at least works for some cores. 15:49:02 Anyway, Does anyone else have anything we should discuss today? 15:50:47 seems we're good :) 15:51:55 Well everyone, thank you for coming today! 15:52:12 And now we can take over the world 15:52:21 +++ 15:52:37 kaifeng: and maybe come to a "ironic caused me to loose my mind" sticker design consensus 15:52:38 :) 15:52:46 Thanks for hosting the meeting, I feel refreshed :) 15:52:52 Dmitry Tantsur proposed openstack/ironic-python-agent master: Run a standalone job in the gate https://review.opendev.org/675937 15:53:21 One day, I'll have "Ironic deploys supercomputers too" stickers. 15:54:52 TheJulia: which ones? :) 15:55:08 mgoddard: getting there :) 15:55:14 Anyway, thanks everyone! 15:55:18 #endmeeting