15:00:33 #startmeeting ironic 15:00:33 Meeting started Mon Oct 12 15:00:33 2020 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 The meeting name has been set to 'ironic' 15:00:37 Good morning everyone! 15:00:40 o/ 15:00:40 o/ 15:00:43 o/ 15:00:44 o/ 15:00:53 o/ 15:00:54 Our agenda this week can be found on the wiki. 15:00:56 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:00:58 o/ 15:00:58 o/ 15:01:16 We seem to be announcement heavy this morning, so if you have announcements be ready to raise them! 15:01:34 #topic Announcements / Reminders 15:01:52 #info Open Infrastucture Summit is *NEXT* week! October 19-23rd! 15:02:19 #link https://www.openstack.org/summit/ 15:02:26 o/ 15:02:40 \o 15:02:47 o/ 15:02:49 #info The Project Teams Gathering is the following week, October 26th through 30th. 15:02:53 \o 15:02:53 we will probably skip next 2 weeks of meetings? (or maybe only next week?) 15:03:16 * dtantsur still has no clue about the schedule 15:03:17 iurygregory: lets discuss that during the discussion period, if you'd quickly edit the agenda on the wikie :) 15:03:22 #link https://www.openstack.org/ptg/ 15:03:31 TheJulia, ++ 15:04:03 #info We have a PTG etherpad, please add items in the next couple of days. After that I will try to break things into logical time blocks for scheduling during the PTG. 15:04:06 #link https://etherpad.opendev.org/p/ironic-wallaby-ptg 15:04:37 #info Expect Victoria final releases this week. Hopefully nothing needs to merge to the stable branch and I just submitted ironic 16.0.1's release. 15:04:45 And last for me! 15:04:56 #info PTL on vacation the first week of November. 15:05:14 Does anyone else have anything to announce or remind us of this week? Or maybe for the next few weeks? 15:06:36 * TheJulia takes silence as no... 15:07:22 Looks like we had no action items last week 15:07:49 Do we feel the need to review the subteam status reports? 15:07:55 since we're at the end of cycle and all 15:09:02 * TheJulia hears silence and guesses "no" ? 15:09:05 probably not 15:09:45 Dmitry Tantsur proposed openstack/ironic master: [WIP] Do not use RPC continue_node_deploy/clean in agent_base https://review.opendev.org/756354 15:09:45 likely no 15:09:48 Okay then 15:10:00 #topic Priorities for the coming week 15:10:07 Above and beyond adding topics to the PTG etherpad! 15:10:16 #link https://etherpad.opendev.org/p/IronicWhiteBoard 15:10:19 * iurygregory is wondering if we should change Victoria Priorities to Wallaby Priorities... 15:10:20 Starting at line 124 15:10:27 iurygregory: Likely :) 15:11:15 It looks like only a couple items merged over the past week, which makes sense for end of cycle. But we do have a number of proposed items starting at line 176 15:11:55 If there are no objections I'd like to add them to the list for this next week 15:12:05 sounds good 15:12:18 sounds good to me 15:12:31 no objections 15:12:41 looks good 15:13:09 In particular, vendors who allow/enable users to use IPMI, I'd like some eyes on https://review.opendev.org/757198 specifically because to validate if the list selector number that is default (0x08) is correct or if it needs to be modified for your hardware in UEFI mode. 15:13:10 patch 757198 - ironic - WIP: IPMI: Handle vendor set boot device differences - 1 patch set 15:13:41 * dtantsur sighs 15:14:01 funny =) 15:14:46 it is a very le sigh topic but it does actually explain some various vendor headaches with ipmi, sadly 15:14:54 yeah 15:15:00 "standards" 15:15:01 TheJulia: I can check with our hardware experts what they think. 15:15:01 evil numbers 15:15:09 arne_wiebalck: would be very very very much appreciated 15:15:36 our users have IPMI access 15:15:37 * TheJulia started to add "le sigh" to the etherpad in reponse to dtantsur's sigh 15:16:23 unfortunately there is almost no documentation out there of variations, beyond the occassional blog post so itmay be most folks just adhered to the standard and we're hitting ipmitool quarks then 15:16:34 the world "protocols" seems to have a better reputation :( 15:16:43 heh 15:17:04 * TheJulia will not suggest editing the wikipedia entry... 15:17:10 https://xkcd.com/927/ 15:17:24 Anyway, silliness aside, it seems like we're goot to proceed to the Baremetal SIG 15:17:48 First meeting took place, relatively low attendance, 8-10 people. 15:17:58 #topic Baremetal SIG 15:18:08 Nice discussion none the less (I think). 15:18:10 I think 8-10 is actually not bad 15:18:10 Is the hope to do the meetings monthly or every two weeks? 15:18:19 Will make some changes to the format. 15:18:25 Monthly. 15:18:42 If you're off first week of Nov, how about Nov 10? 15:18:50 Also some distance to the summit. 15:19:05 Okay, If we can have the time and maybe discussion topic agenda available the week before, I'd encourage people tweeting it and posting to other social media to see if we can get more people to join our SIG meetings 15:19:08 ++ 15:19:16 Right. 15:19:34 we need to define the next topic and the one who starts the discussion 15:19:35 Yeah, at least the next two weeks will be meeting heavy for most people 15:19:53 that is what I thought, therefore Nov 10 may be ok 15:20:18 ++ 15:20:24 ok, I will try to identify the topic and a victim ... erm, speaker 15:20:29 victims! 15:20:32 that is it 15:20:46 Redundant Array of Independent Speakers 15:21:03 woot 15:21:05 this is what we had last time 15:21:09 Anyway, I'm clearly hitting maximum silly 15:21:10 worked 15:21:20 Are we good to proceed to Open Discussion? 15:21:24 +1 15:21:26 let's 15:21:29 #topic Open Discussion 15:21:37 So iurygregory added two topics for open discussion 15:21:43 the first is should we skip the next two meetings 15:21:44 yup o/ 15:22:38 do we have anything on Mondays? 15:22:58 I guess on the PTG week we'll have enough talking to each other :) 15:23:11 Yes, there is actually talk at 15 minutes into our meeting on Monday 15:23:33 so I guess we should just cancel both meetings. Any objections? Concerns? Thoughts? 15:23:35 *shrug* let's skip them 15:23:43 Ack, I'll send an email to the mailing list. 15:24:09 Next topic appears to be a brewing epic battle of "openstack baremetal" vs "baremetal" 15:24:19 iurygregory: would you like to walk us through this? 15:24:33 sure 15:25:08 fight! 15:25:18 We had a plan to update our docs to use `baremetal` since it wouldn't be openstack specific (we have the standalone case) 15:25:44 so just to see if there is a consensus of our community if we should or not change our docs 15:26:09 what are the arguments against? 15:26:09 https://review.opendev.org/#/c/756708/4/doc/source/admin/portgroups.rst 15:26:10 patch 756708 - ironic - Update .rst files - 4 patch sets 15:26:12 dtantsur: I was thinking more like we would compare the output of work product like in the show "Forged in Fire" here in the states...... 15:26:39 is Iury speaking of https://review.opendev.org/#/c/756708/ ? 15:26:40 oh, so I should put back my battle axe? 15:26:40 patch 756708 - ironic - Update .rst files - 4 patch sets 15:26:49 kaifeng, correct =) 15:27:20 kaifeng: re your question: the problem is: 'baremetal' works for both standalone and nova cases, 'openstack baremetal' - only for nova case 15:27:23 dtantsur: well, maybe not? 15:27:25 okay, I am not against s/openstack baremetal/baremetal/, totally 15:27:37 s/nova/full openstack/ but you get an idea 15:27:47 kaifeng: your raising a good point and I kind of feel like maybe the opposite idea might be good. 15:27:48 just I feel the doc is in an inconsistent state 15:27:48 Merged openstack/ironic-python-agent stable/stein: Fix: make Intel CNA hardware manager none generic https://review.opendev.org/753730 15:27:53 Merged openstack/ironic-python-agent stable/rocky: Fix: make Intel CNA hardware manager none generic https://review.opendev.org/753785 15:27:54 people read our docs and like 'what, I need the whole openstack to use ironic'? 15:27:57 Merged openstack/ironic-python-agent stable/queens: Fix: make Intel CNA hardware manager none generic https://review.opendev.org/753786 15:28:02 Merged openstack/ironic-prometheus-exporter stable/ussuri: Fallback to `node_uuid` if`node_name` is not present https://review.opendev.org/753065 15:28:22 the pure 'baremetal' command should work in all cases, so I guess we can remove the word 'openstack', just be consistent everywhere 15:28:35 I think it is ultimately good for us to peel back the openstack command references in the documentation and highlight that there is integration with the openstack command 15:29:09 Ultimately so we reach consistency in our docs and hopefully can get to a point where we're publishing some form of them on the newer website 15:29:25 which is where people would surely be confused by the "openstack" command. 15:29:43 yeah 15:30:11 if the two command can be used interchangeably, I agree to use `baremetal` instead of `openstack baremetal`, we just need a disclaimer that, hey, if you want to type more characters that is also fine 15:30:15 it's already a problem now e.g. among metal3 folks who don't realize it's safe to remove the 'openstack' prefix 15:31:39 on the consistency topic, I agree that we probably shouldn't leave openstack commands hanging around in random places of our documentation 15:31:40 kaifeng: ++ They can be 15:31:50 ++ 15:31:54 i.e. we should concentrate openstack-related aspects together 15:32:06 I agree with that 15:32:30 Do we seem to have consensus then? It feels like we kind of do in terms of meaning and basic direction 15:32:34 I also agree with that 15:33:22 excellent 15:33:51 So what else shall we discuss? Taking over the world? 15:34:03 for outreachy I think we can check if openstack baremetal was switched to baremetal, and we can have other task for outreachy or any contributor to concentrate the openstack commands in other place 15:34:08 this would make sense? 15:34:41 ++ 15:34:46 I think that makes sense, it would be a slighly harder task, but a good evolution 15:34:51 TIL `baremetal` exists independently of `openstack baremetal` 15:34:53 yeah 15:35:37 JayF: new in ussuri 15:36:13 Verification of a change to openstack/ironic-prometheus-exporter failed: Fallback to `node_uuid` if`node_name` is not present https://review.opendev.org/753067 15:36:20 dtantsur: of openstackclient? 15:36:26 JayF: of ironicclient 15:36:31 ah 15:36:48 it was a good solution for those that missed running "ironic" commands 15:37:00 there's the ipa-builder conundrum aka we can't build rpm-based ipa ramdisks on focal using diskimage-builder 15:37:00 I sent a mail to the mailing list and got some good replies, although we might want to decide which direction to take 15:37:00 we already build centos8 on centos8, although we still use bionic to test devstack 15:37:52 * kaifeng would love to have it backported to train, it would relieve some finger movements in daily lifes 15:37:54 The suggestion seemed to be that not using the minimal element should allow it to work 15:38:50 I'm not sure, but I'll give that a try 15:38:53 kaifeng: the baremetal command? cannot you just pull in a newer ironicclient? 15:39:54 rpioso: I think it would be worth it, it may mean we get to identify some extra "junk" in the images that can be purged before final packaging 15:39:57 err 15:40:01 that was meant for rpittau 15:40:15 :) 15:40:21 dtantsur we are running train at the moment, upgrade client would introduce requirements 15:40:26 not using the minimal element has a risk of increasing the size quite substantially 15:40:37 yeah, I'm mainly worried about the size of the final image, but I guess it's the easiest path to take at the moment 15:41:04 we need to check how much it increases the size 15:41:14 if the increase is so big that we cannot longer use it in the CI 15:41:16 I'll check that this week 15:41:20 then it's pointless 15:41:24 I suspect even with the minimal element we have things we can remove from that DIB image. 15:41:38 JayF: we already remove a bunch of stuff 15:41:39 JayF: we remove a lot of stuff by hand 15:41:48 most of the image right now is the linux-firmware package 15:42:01 which is several hundreds of megabytes 15:42:18 That makes a lot of sense 15:42:24 * arne_wiebalck has just removed this downstream ... 15:42:36 arne_wiebalck: you know your hardware, you can do it 15:42:41 I have it on my 'eventually' list to see why they were so huge, that makes a LOT of sense 15:42:41 arne_wiebalck: you removed the linux-firmware ? oO 15:42:43 dtantsur: exactly 15:42:46 guys, I've got to go 15:42:47 and I don't know my hardware, so I probably can't :D 15:42:48 #endmeeting