15:01:43 #startmeeting ironic 15:01:44 Meeting started Mon May 20 15:01:43 2019 UTC and is due to finish in 60 minutes. The chair is dtantsur. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:45 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:45 #chair TheJulia 15:01:47 The meeting name has been set to 'ironic' 15:01:47 o/ 15:01:48 \o 15:01:49 Current chairs: TheJulia dtantsur 15:01:50 \o 15:01:52 o/ 15:01:54 o/ 15:01:57 okay, I could not resist, sorry :) 15:01:57 Julia Kreger proposed openstack/ironic-specs master: Add train priorities https://review.opendev.org/660145 15:02:03 o/ 15:02:04 o/ 15:02:05 woot! 15:02:05 \o/ 15:02:07 o/ 15:02:10 TheJulia: the mic is yours now :) 15:02:16 dtantsur: thanks, I was just finishing up the priorities doc 15:02:20 false start! 15:02:20 o/ everyoen 15:02:21 \o 15:02:34 Sorry sorry! 15:02:39 * jroll sneaks in the back \o 15:02:43 o/ 15:02:44 no worries, last Monday I was 10 mins late :D 15:02:48 lol 15:02:51 it's hard to get back in schedule after PTO.. 15:03:01 Yeah, and I literally was on PTO last week 15:03:15 :-) 15:03:24 Our agenda this week can be found on the wiki. 15:03:30 #link https://wiki.openstack.org/wiki/Meetings/Ironic 15:03:35 o/ 15:03:49 #topic Announcements/Reminders 15:04:15 #info Dmitry has written up notes from the PTG (Thanks Dmitry!) and shared them with the world. 15:04:24 #link https://dtantsur.github.io/posts/ironic-denver-2019/ 15:04:34 0/ 15:04:57 #info The Next PTG will be in Shanghai on November 6th through 8th. 15:05:16 Does anyone have anything to add or remind us of? 15:05:33 Wow, look like good notes. 15:06:01 :) 15:06:44 I guess not... 15:06:55 items to raise that is :) 15:07:08 dtantsur: I'm sorry, I've not had time to look at your notes yet. Hopefully later today :) 15:07:27 No problem, they will be waiting for you :) 15:07:31 They're awesome! 15:07:31 \o/ 15:07:39 thx :) 15:07:40 Moving on then! 15:07:43 #topic Review action items from previous meeting 15:08:07 #info We had no formal action items from last week, but I know there was post meeting discussion of someone emailing regarding the baremetal sig. 15:08:20 Since I've barely even started looking at my email, did that take place? 15:08:27 arne_wiebalck: I think it was you ^^ 15:08:34 and I think I've seen some email 15:08:40 I contacted Chris. 15:08:54 Okay, then I think we can move along then :) 15:08:55 I don't think he's sent sth out. 15:09:08 we also had an implicit action item to fix the CI, and we did it (modulo the non-voting multinode grenade job) 15:09:23 ugh 15:10:46 It does not look like an email went to a mailing list, at least one that shows in my mailing list mailbox inbox 15:10:56 #action TheJulia follow-up on baremetal sig communication 15:11:17 Moving on! 15:11:19 #topic Review subteam status reports 15:11:39 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:11:45 Starting around line 217 15:12:28 Looks like Software raid could use some reviews? 15:12:37 Yes, please :) 15:13:05 Some reviews were already done, thanks! 15:13:05 ++ 15:13:07 I'm going to strike out the bulk of deploy templates, since step decomposition will replace it 15:13:18 I was going to get there, but $errands 15:13:55 Has anyone started ropping the py35 tests? 15:13:58 well, test jobs 15:14:14 yes, I saw at least 2 changes 15:14:18 I think most of our projects don't have them 15:14:24 https://review.opendev.org/652385 15:14:24 note that we shouldn't literally drop them 15:14:25 patch 652385 - networking-generic-switch - Dropping the py35 testing - 2 patch sets 15:14:37 there is a new template in zuul per version 15:14:42 dtantsur: ++ 15:14:54 I don't recall the name, something like python-train-jobs 15:15:16 dtantsur: can you investigate and follow-up for our next meeting for that? 15:15:27 I can try 15:15:32 Thanks 15:16:25 fast track... one day I'll write docs :( 15:16:40 etingof: how is redfish virtual media going? 15:16:58 * dtantsur hides 15:17:02 :) 15:17:04 TheJulia, simulation is up and running 15:17:20 * etingof is debugging redfish hw type against the simulator 15:17:30 awesome 15:17:39 * etingof gets up to the point where VM is [not] booting 15:17:41 etingof: Could you update the whiteboard in that case? 15:17:45 sure 15:17:48 Thanks! 15:18:17 The locking item is not on the proposed priority list, but it is stil a nice to have general improvement that we have some consensus behind. Looks like I need to rebase that. hopefully this week. 15:18:26 ++ 15:18:41 arne_wiebalck: re nova callbacks, is it just power, or is there work in progress for state at this time? 15:19:16 TheJulia: Not sure I understand the question, sorry. 15:19:53 arne_wiebalck: I know you've been invovled in some of the power call back discussikons with nova, does that include state callbacks or just power? 15:20:25 I forgot to announce it, but better late than never: two outreachy interns are joining us for summer! 15:20:25 Only power. 15:20:38 dtantsur: Nice. 15:21:19 dtantsur: \o/ 15:21:21 \o/ 15:21:23 arne_wiebalck: okay 15:21:39 dtantsur: Great news :) 15:21:43 one will work on sushy, the other - on ironic-python-agent-builder 15:21:50 cool 15:23:29 TheJulia: The power stuff is difficult enough ;) 15:23:35 The rest of the items on the whiteboard were dropped from Stein. They are, as far as I'm remembering from typing, not in the proposed patchset for train 15:23:42 arne_wiebalck: Yeah, and it still helps a LOT! 15:24:03 TheJulia: maybe for Train we can have two sections: 1. Priorities, 2. Things people care about that progress well 15:24:16 ++ 15:24:19 that's essentially what we had for Stein, but expressed explicitly 15:24:34 kind of, but I divided that by size of work 15:24:50 Anyway, the change for the train priorities is up in review 15:24:55 great! 15:25:07 #link https://review.opendev.org/#/c/660145/ 15:25:08 patch 660145 - ironic-specs - Add train priorities - 1 patch set 15:26:02 So I guess we're good to proceed onward? 15:27:32 yep 15:27:33 let's! 15:27:40 #topic Deciding on priorities for the coming week 15:27:52 #link https://etherpad.openstack.org/p/IronicWhiteBoard 15:27:55 Starting around line 123 15:28:52 Looks like a few items have merged 15:29:00 A few items are stable-core reviews 15:29:23 Speaking of which, tonyb, let us know if the overall policy has been updated as discussed in Denver 15:29:45 I'm going to remove merged items 15:30:11 Are tehre any other items that people feel need to be on this list? 15:30:31 added one for nova 15:30:36 https://review.opendev.org/659809 15:30:37 patch 659809 - nova - [ironic] Don't remove instance info twice in destroy - 1 patch set 15:31:04 Could use reviews on this sushy patch https://review.opendev.org/#/c/657688/ 15:31:04 patch 657688 - sushy - Enhance Storage models to support RAID config - 3 patch sets 15:31:05 efried has said he will proxy a +1 from an ironic-core 15:31:12 I've not looked at your patch mgoddard, but why just not ignore it? 15:31:28 mgoddard: are you thinking from a backport standpoint? 15:31:48 TheJulia: it requires you to wait for all the retries 15:32:10 I clearly need more coffee 15:32:11 I'd like to have some reviews on the spec https://review.opendev.org/#/c/650630/ 15:32:12 patch 650630 - ironic-inspector-specs - Introduce etcd for service coordination - 4 patch sets 15:32:23 oh, this ^^ is a good one 15:32:28 * dtantsur +1 to adding it 15:32:41 * TheJulia adds to the lists 15:32:46 thanks :) 15:33:01 \o/ 15:33:04 mgoddard: looks like you already got your patch 15:33:34 Okay, if there is nothing else, I guess we can mov on to Baremetal sig 15:35:43 #topic Baremetal Sig 15:35:57 Hmm.. No Chris. :\ 15:36:25 I think the main task for now would be to complete the white paper. 15:36:28 Okay, Well use case documentation is in the list of goals for train. Has anyone had time to look at or discuss the white paper? 15:36:31 ++ 15:37:00 * TheJulia wonders if "review the whitepaper is a thing we could put on the priorities for the week. 15:37:01 " 15:37:09 I have followed up with Lenovo and have a couple of people that would like to get involved. 15:37:15 \o/ 15:37:19 jungleboyj: that is awesome news 15:37:38 Yeah, not sure if they have joined the meeting. May need to get them up to speed. 15:38:05 I got the feeling at the summit though, that we didn't really know what the next steps for the WhitePaper were though. 15:38:26 jungleboyj: The work on the document has started. 15:38:29 Largely it was use cases 15:38:34 and operator cases 15:38:40 jungleboyj: More use cases are probably needed 15:38:58 Ok. 15:39:28 #link https://docs.google.com/document/d/1KBhJcmpCTm8hn0BX-jO5MAyBYf2RHM5Jh5ib7GTZogM/edit#heading=h.1jw85s2vpx7q 15:39:51 Also some context on use, which I think this team can add quite a bit about 15:40:30 Maybe us engineering types should start on the last page and move forward :)_ 15:40:34 s/_// 15:40:45 :-) 15:40:54 Anyhow, if there is nothing else to discuss re the white paper, we can move along to RFE review 15:41:27 I will keep following up with my team. 15:41:44 jungleboyj: Awesome, thanks 15:41:49 Moving on to RFE review 15:41:53 #topic RFE Review 15:42:09 #link https://review.opendev.org/#/c/656799/4 15:42:10 patch 656799 - ironic-specs - Add support for node retirement - 4 patch sets 15:42:39 I've not had a chance to read this sadly 15:43:00 This will probably need to be superseded by a more general proposal. 15:43:36 I'll try to gather more use cases and make a suggestion. 15:43:37 I think we're generally open to general proposals and the overall theme is covered by the train priorities... It hink 15:43:39 think 15:44:07 I guess the key point is to avoid inventing the same feature under two slightly different flavors 15:44:15 ++ 15:44:15 i.e. retirement and quarantine 15:44:24 I suspect we'll need separate workflows 15:44:25 dtantsur: totally agree 15:44:33 I remember discussing it last time and... I don't remember anything from then :( 15:44:35 but overall mechanics under the hood should be similar 15:44:44 Sounds like we all need coffeee 15:44:46 \o/ coffee 15:45:04 I guess we can talk about the next item then. A deployment API 15:45:05 #link https://storyboard.openstack.org/#!/story/2005737 15:45:17 I know we don't want to overload "maintenance", but quarantine sounds like an extended maintenance ? :) 15:45:32 One point will be whether or not this need to be a state (or a flag). 15:45:40 rpittau: but it is still a state of being for for the node.. :\ 15:45:44 * arne_wiebalck was surprised to see how ops voted at the ptg 15:45:49 yeah, right 15:45:58 * TheJulia barely remembers that now 15:46:09 state:4, flag: 0 15:46:36 flag seems more natural to me tbh 15:46:52 (and easier :-D) 15:47:06 a flag it's definitely easier, but also less visible 15:47:13 dtantsur: true 15:47:23 dtantsur: re your deployment api.... I think your missing instance_uuid since that needs to match nova's already created DB entry 15:47:53 a flag also gets us into lots of corner case code that is less visible when we go through the code :\ 15:47:54 TheJulia: instance_uuid is implicit there. it's either set in advance or can be set to deployment UUID 15:48:11 I will mention it in the spec, if we decide it needs a spec 15:48:40 dtantsur: I'm not sure a spec is really required, it is fairly clearly a pass-through (in flask maybe?!?) 15:48:53 but a flag can be carried around while walking the state machine 15:49:01 POST /v1/deployments/instance_uuid ? 15:49:07 yeah, it's largely a pass-thru to Node, although it will probably have to have a short record in the database 15:49:24 * arne_wiebalck wonders if we have two discussions at the same time :) 15:49:27 TheJulia: you mean, make Deployment UUID == Allocation UUID? 15:49:29 arne_wiebalck: yeah... And I think that is where the two features begin to differ 15:49:30 arne_wiebalck: we do :) 15:49:50 ok, let's stop the quarantine one for now 15:49:56 dtantsur: well, if one was chosen externally, one has to be able to define it 15:50:09 oh, but I just had a thought :) 15:50:16 re quarentine :) 15:50:18 * TheJulia stops now 15:50:31 I seriously need more coffee and we only ahve 10 minutes left 15:50:57 I'm basically good with the rfe for the API. I do think a quick spec is going to be useful since the proposal is a tl;dr :) 15:51:11 yeah, I'll probably write something this week 15:52:11 k 15:52:21 I guess that leaves us to open discussion! 15:52:26 #topic Open Discussion 15:52:34 Does anyone have anything to discuss during the meeting? 15:53:52 * TheJulia hears the army of crickets appear 15:54:42 :) 15:58:34 Well everyone, thanks! 15:58:37 Have a wonderful week! 15:58:55 Thank you. 15:58:57 thanks! you too! 15:59:11 Riccardo Pittau proposed openstack/ironic-python-agent stable/rocky: [WIP|DNM] testing coreos jobs https://review.opendev.org/656507 15:59:14 Thank you 15:59:29 Thanks! 15:59:42 Thank you, and welcome back TheJulia. 15:59:54 thank you TheJulia 16:00:11 #endmeeting