15:01:08 #startmeeting ironic 15:01:08 Meeting started Mon Feb 3 15:01:08 2020 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:09 o/ 15:01:09 o/ 15:01:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:11 o/ 15:01:13 o/ 15:01:14 The meeting name has been set to 'ironic' 15:01:15 Good morning everyone! 15:01:18 o/ 15:01:19 \o 15:01:21 o/ 15:01:24 o/ 15:01:37 o/ 15:02:06 o/ 15:02:06 Our agenda can be found on the wiki. Hopefully we should get through it fairly quickly. 15:02:11 \o 15:02:12 o/ 15:02:16 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:02:18 o/ 15:02:37 #topic Announcements/Reminders 15:03:07 #info Ironic Mid-cycle February 25-26 @ CERN 15:03:14 #link https://indico.cern.ch/event/863986/ 15:03:26 #link https://etherpad.openstack.org/p/ironic-ussuri-midcycle 15:03:34 I'll try and send out a mail to the registered participants some time this week. 15:03:42 arne_wiebalck: awesome 15:03:48 With some logistics details. 15:03:59 * iurygregory probably won't go to the mid cycle =( 15:04:10 If anyone is looking at tickets from the states, it seems like flying in to CDG and taking a train is semi-reasonable. 15:04:21 iurygregory: :( 15:04:53 #info Opendev + PTG Vancouver - June 8th through 11th 15:05:15 #info TheJulia: Announcing Dell 3rd Party Ironic CI now functional for Openstack/sushy library. Please take a look at https://review.opendev.org/#/c/705289/ 15:05:16 patch 705289 - sushy - Testing thirdparty sushy CI - 3 patch sets 15:06:15 rajinir: can you please a DNM to the commit message title? So it's clear we shouldn't merge that :) 15:06:28 Additional info for the Opendev + PTG event 15:06:29 rpittau: sure 15:06:32 thanks! 15:07:01 Rajini Karthik proposed openstack/sushy master: Testing thirdparty sushy CI ( DON"T MERGE) https://review.opendev.org/705289 15:07:31 #info The foundation is looking for people interested in being part of the programming committee. If your interested in helping out, there is a form linked from the email Ashlee Ferguson sent out. 15:07:32 #link http://lists.openstack.org/pipermail/openstack-discuss/2020-January/012239.html 15:07:42 Does anyone else have anything to announce or remind us of? 15:08:03 TheJulia: Please see rajinir's message above. 15:08:28 #info Announcing Dell 3rd Party Ironic CI now functional for Openstack/sushy library. Please take a look at https://review.opendev.org/#/c/705289/ 15:08:28 patch 705289 - sushy - Testing thirdparty sushy CI ( DON"T MERGE) - 4 patch sets 15:08:43 Hopefully that will show up in the summary correctly :) 15:08:50 :) 15:09:03 rpioso, thank you 15:09:08 It looks like we had no action items last week. 15:09:16 rajinir: yw 15:09:33 I think everyone has been fairly heads-down so I'm wondering if it even makes sense to visit the subteam status reports? 15:09:46 thank you rajinir! that is great that we have more testing :) 15:10:42 rloo: thank you 15:11:17 I propose we jump to priorities for the week since it appears to not have made much progress the last week. 15:11:44 yep 15:11:59 #topic Deciding on priorities for the coming week 15:12:16 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:12:22 Line 169 15:13:36 It would be awesome if a couple people could review Ilya's L3 deployment spec revision today... It is actually also open on my other screen at the moment so I think he would appreciate feedback and maybe even an approved spec when he gets back from FOSDEM 15:14:19 Does anyone have anything to add to this list? 15:14:37 TheJulia: yes 15:14:38 I guess one question is a stable/rocky fix. I'm wondering if we just abandon it and move on? 15:14:55 can we add this? https://review.opendev.org/705395 15:14:55 patch 705395 - ironic - Fix jsonpatch related tests - 3 patch sets 15:15:06 rpittau: Sure, please add! 15:15:53 Anyone else? 15:15:55 done, FYI CI is broken without that :P 15:16:02 \o/ 15:16:05 Okay 15:16:35 I guess we're good then? 15:16:39 rocky is going into extended maintenance soon? 15:17:02 rloo: yes, like a week or two I think 15:17:15 rloo: yes, the 24th 15:17:22 don't we need rocky CI working at least before then? 15:17:36 or do we decide now that we aren't going to 'support' rocky? 15:17:43 or whatever the term is... 15:17:59 It is not completely broken, it is a single job that breaks... If memory serves for completely unrelated reasons 15:18:42 i just did a recheck on a patch for rocky and i was wondering about the state of rocky CI: https://review.opendev.org/#/c/692516/ 15:18:43 patch 692516 - ironic (stable/rocky) - Kill misbehaving `ipmitool` process - 1 patch set 15:18:51 In the extended maintenance path, we have the option of killing jobs that are too much overhead to maintain. We could do that and merge the fix. 15:20:23 wondering if we want to table that question until after Feb 24. For now, the situation is that we can't merge anything cuz CI rocky is broken. So unless someone fixes the CI, we're in the current state we're in, right? 15:20:41 after the 24th, I'd just remove the job anyway 15:21:16 ok, we're good then. we'll wait til after 24th, the job will be removed, then we can merge. yes? 15:21:20 yes, we're stuck unless we fix the issue which I believe is elsewhere, get that merged in the next two weeks, and then.. yeah 15:21:31 Yes, that is entirely reasonable 15:21:39 umm. when do we do a (last?) release of rocky? 15:21:54 whoops rechecked wrong patch :/ 15:22:26 rloo: the release team does a final tag I believe around the 24th 15:22:26 (ie, do we want these changes in the last official release of rocky) 15:22:59 Nobody is screaming for it, so I suspect it is fine if it is something people can pickup later if needed 15:23:00 TheJulia: ok, so I think that's really the only diff. Changes to Rocky avail in the last tag, or not. 15:23:34 yup 15:23:35 And yeah, if no one is 'asking', then let's go with our plan, unless someone wants to fix the CI for Rocky. 15:23:48 Iury Gregory Melo Ferreira proposed openstack/ironic master: WIP - Native zuulv3 grenade multinode multitenant https://review.opendev.org/705030 15:24:22 sounds good 15:24:26 Moving on! 15:24:30 #topic Discussion 15:25:10 #info stevebaker has proposed copying a subset of WSME type validation code into ironic. Steve's timezone really prohibits him attending this meeting as he should be quite asleep at the moment. 15:25:20 #link https://review.opendev.org/704486 15:25:21 patch 704486 - ironic - WIP do all serialization in the expose decorator - 1 patch set 15:26:53 WSME is MIT licensed, and he is proposing moving about a thousand lines into ironic. I'm not really objecting to it as long as it is properly attributed. My only concern is really maintenance of the code, but copying the validations would at least put them into an active code base, but honestly I'd also rather see a more widely used piece of code out there for type validations 15:27:46 I don't know enough about it, but it seems odd that we'd need to move (copy/paste) that code to ironic. is ironic the only project that needs that code? 15:27:53 in general is a good idea and sure helps the transition to Flask, as that's the final goal 15:28:09 WSME was not only for ironic/pecan 15:28:28 It supports other web frameworks also 15:28:30 but I think we're one of the last projects using wsme? 15:28:34 rpittau: so you're saying this wsme code is only temporary for ironic? 15:28:39 rloo: exactly 15:28:52 rpittau: how 'temporary' are we thinking? 15:28:59 did keystone use wsme before migration? 15:29:05 I didn't percieve it as temporary 15:29:11 kaifeng: I don't think so 15:29:13 rloo: hard for me to answer that :) 15:30:00 I think keystone was using wsme for some time.... 15:30:02 i guess i'd like to know if any other openstack project is doing what ironic is doing, and whether this wsme copy is the way they've done it, and whether IF we copy, it should be in a library. 15:30:18 but they moved to flask, wondering how they do validation now? 15:30:26 I'm speaking/thinking from ignorance though 15:30:58 "temporary" nearly always means forever, fwiw 15:31:12 jroll: that's why i asked :) 15:31:16 in my opinion, of course :) 15:31:36 keystone uses jsonschema for validation now, afaik 15:32:24 note that is strictly for validation, not for serialization to/from json/python objects 15:33:11 validation was only one reason for WSME using at ironic, actually ironic used WSME's types and base classes also 15:33:49 right, serialization is the other thing we used it for 15:33:59 And expose function also 15:34:18 kaifeng: I believe they wrote their own validation code... 15:34:19 Did we would use flask in future? 15:34:21 it feels like we could/should rip out the intermediate objects (not sure if now or later) 15:34:36 If yes, current wsme code would change also 15:34:58 for differnt web frame work WSME code was diffent 15:34:59 wrote their own originally 15:36:20 TheJulia: How will driver vendor passthru APIs be converted? 15:36:29 jroll: I feel like that might be reasonable, but I am struggling to visualize what that would look like 15:37:04 TheJulia: I was just skimming code and wondering to myself if we could just use our existing o.vo objects or something. or just dicts in the api layer. 15:37:05 rpioso: they would just be serialized into json and passed through, I believe that is what happens presently 15:37:18 TheJulia: Thank you! 15:38:35 dict seem like the would be the most native and easiest to work with, but I suspect that also brings risk. Anyway, Steve has been focused on this, I suspect this is all feedback for him on thoughts and maybe he has ideas. There is also a thread on the mailing list that may be worth looking at/replying to 15:40:05 I guess we should move on 15:40:11 Any objections? 15:41:20 Moving on! 15:41:24 #topic Baremetal SIG 15:41:30 Regarding the whitepaper! 15:42:09 * arne_wiebalck has to get going on this one 15:42:43 I chatted with the foundation staff about getting the baremetal logo program submissions. The tl;dr is that they can't share them because of privacy issues and advised us that we should use content that was already published and can be attributed to that publication or that is directly submitted to the whitepaper. 15:43:27 The tl;dr of that is that if there is content to be submitted, please do so soon. I'm working on getting Red Hat's submission released so we can put it in. I think that will help bolster the whitepaper, tbh. 15:44:26 I won't be able to work much on this before the mid-cycle I'm afraid. 15:45:25 Nor will I sadly, but I should at least have publication data. I also have an interesting reference that might work, but I need a few days to hunt through NSF proposals.. that is if I can find the public versions. 15:46:08 Anyway, I guess we can proceed to Open Discussion if there is nothing else 15:46:09 ? 15:47:56 #topic Open Discussion with ?coffee? 15:48:03 * TheJulia suspects coffee is required again 15:48:19 * rpittau sipping chocolate and pear tea 15:48:38 * arne_wiebalck needs a coffee as well 15:50:58 Well, I hope everyone has a wonderful week 15:51:26 thanks! 15:51:47 crickets 15:51:47 thanks, TheJulia 15:51:49 thx! 15:51:54 Indeed! 15:52:11 I'll end the meeting early if everyone promises to go review the L3 Deployments spec revision :)\ 15:52:39 ok, promise :D 15:52:39 https://review.opendev.org/#/c/672780/ 15:52:40 patch 672780 - ironic-specs - Update L3 based deployment spec - 31 patch sets 15:52:43 :) 15:52:48 Thanks everyone 15:52:49 \o 15:54:13 have to take a look tomorrow :) 15:55:00 Thanks, I am just hoping we finally reach some level of consensus and are able to move forward since it is possibly the most impactful feature we could deliver this cycle. 15:55:06 And with that, Thanks! 15:55:09 #endmeeting