15:00:21 <TheJulia> #startmeeting ironic
15:00:21 <TheJulia> o/
15:00:21 <openstack> Meeting started Mon Oct  1 15:00:21 2018 UTC and is due to finish in 60 minutes.  The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:26 <openstack> The meeting name has been set to 'ironic'
15:00:45 <jroll> \o
15:00:47 <cdearborn> o/
15:00:48 <bdodd> o/
15:00:52 <stendulker> o/
15:01:05 <rpioso> o/
15:01:13 <hshiina> o/
15:01:21 <TheJulia> Our agenda can be found on the wiki, as always
15:01:21 <mgoddard> o/
15:01:23 <TheJulia> #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting
15:01:35 <TheJulia> We have a few things on the agenda today, but hopefully they will be fairly quick.
15:01:47 <rloo> o/
15:01:56 <TheJulia> #topic Announcements / Reminders
15:02:55 <TheJulia> #info During the PTG, we discussed Python3, and 3rd party CI.
15:03:38 <TheJulia> #info The basic consensus after consulting with one of the python3 goal champions is that we should ensure that we at least have some 3rd party CI driver coverage for python3.
15:03:43 <rajinir> o/
15:03:55 <mjturek> \o
15:04:09 <rajinir> o/
15:04:26 <sambetts|afk> TheJulia: that is something we can look into once we get our driver's SDKs updated
15:04:29 <rloo> TheJulia: does that mean, each 3rd party CI driver should have some, or does it mean if any 3rd-party CI driver has some?
15:04:29 <TheJulia> #info As such, third party CI operators should switch at least one CI job to be Python3 native. We don't need full coverage for all scenarios, just need to ensure that the driver generally works, and we can slowly phase all CI jobs over to python3
15:04:44 <rloo> (that explains it, thx!)
15:04:48 <TheJulia> \o/
15:04:53 <TheJulia> Sorry, that was a lot of words to type out
15:05:19 <TheJulia> tl;dr: duplication of jobs is not required, just reasonable coverage/conversion for now.
15:05:21 <mjturek> TheJulia: I'm going to talk to our pkvmci guru this week and see if he sees any problems there. I think we'll be able to do it though.
15:05:30 <TheJulia> There is only one case that requires duplication, which is mjturek's case :(
15:05:39 <TheJulia> mjturek: Awesome!
15:05:54 <openstackgerrit> Merged openstack/ironic-specs master: Stein priorites  https://review.openstack.org/603161
15:06:00 <TheJulia> Anyone else have anything for announcement or reminders?
15:06:18 <rloo> did the time for forum proposals pass?
15:06:35 <TheJulia> Oh, I'm likely going to take Thursday off this week, and may take part Wednesday, or tonight. Silly inbound weather :(
15:06:43 <rloo> oh yeah, congrats TheJulia for being on TC. Hope this doesn't take away your time from being our PTL :D
15:06:44 <TheJulia> rloo: It did, four sessions were submitted
15:07:21 <TheJulia> rloo: was there another session you were desiring that was not on https://etherpad.openstack.org/p/BER-stein-forum-ironic-brainstorming ?
15:07:35 <rloo> TheJulia: no, just wanted a FYI thing about it :)
15:07:49 <TheJulia> Okay
15:07:59 <TheJulia> Well then, I think it is time to move on
15:08:19 <TheJulia> #topic Review action items from previous meeting
15:08:27 <TheJulia> I don't think we had any action items last week, double checking
15:08:50 <TheJulia> Nope, no action items
15:09:08 <TheJulia> #topic Review subteam status reports
15:09:41 <TheJulia> I started to put the priorities down in the ironic whiteboard this morning, starting at line 173.
15:09:44 <TheJulia> #link https://etherpad.openstack.org/p/IronicWhiteBoard
15:10:41 <TheJulia> Each group that has volunteered will want to make sure their information is up to date and that they are tracking anything pertinent to that effort under the item.
15:10:53 <TheJulia> Any questions!
15:10:55 <TheJulia> ?
15:11:44 <rloo> TheJulia: are you going to leave it in that format in the etherpad? with the projects/features as items under 'shorter/full' sections?
15:11:57 <TheJulia> rloo: yeah, that was kind of what I was thinking
15:12:09 <TheJulia> I figure we can shuffle items around as we get closer
15:12:19 <TheJulia> to landing their code or remove items from the list
15:12:57 <rloo> TheJulia: i'm concerned that it will make it harder to read, by the time i get to DHCP-less.. one, I am guessing the 'Shorter term' part will have scrolled off my window/screen. And people will get the indents wrong, but whatever works for you/the team.
15:13:23 <TheJulia> I can see that
15:13:29 <TheJulia> I'll just make it one list then
15:13:33 <rloo> it is somewhat an administrivia thing, i'm only concerned with readability :)
15:13:46 <TheJulia> But very valid :)
15:14:38 <rloo> given that we have stories and stories are supposed to be good at tracking stories/features/ ?, we should try to keep the stories updated with what needs to be done, and this etherpad with latest status?
15:14:39 <TheJulia> Since priorities are in the process of merging, is everyone good to proceed, or do we need further discussion or even time to add text to the etherpad?
15:14:52 <rloo> priorities have merged :)
15:14:56 <TheJulia> \o/
15:15:21 <TheJulia> I was going to go through and made sure everything has a story later today
15:15:32 <rloo> so my ask is for folks to please use the stories to list their subtasks etc, to getting things done :)
15:15:39 <TheJulia> ++++
15:15:57 <TheJulia> That has actually helped me a lot for a few times I've added more subtasks. "Oh, I still need to do that one thing"
15:16:49 <jroll> ditto
15:16:54 <rloo> (otherwise, I'm good with this. Unless someone wants to highlight something related to a particular item?)
15:17:31 <rloo> like... exciting/interesting read on devlist about deploy templates, traits and placement ;)
15:17:47 * etingof is wondering is there any difference between a task and a subtask
15:18:01 <TheJulia> I've not gotten to my email but there was a long discussion on friday about that as well
15:18:11 <mgoddard> etingof: in Jira yes, storyboard no AFAIK
15:18:26 <jroll> TheJulia: ohhhh, it continues :)
15:18:29 <openstackgerrit> Dhanuka proposed openstack/sushy master: Add protocol_features_supported property to the root Sushy object  https://review.openstack.org/606974
15:18:30 <etingof> but we are talking about storyboard here?
15:18:38 <mgoddard> etingof: yes
15:18:38 <TheJulia> jroll: oh noes :(
15:18:49 <etingof> thank you, mgoddard
15:19:03 <TheJulia> jroll: I guess I'm jumping into that fray again later today
15:19:18 <TheJulia> Anyway, Shall we move on?
15:19:28 <rloo> ++
15:19:30 <dtantsur> ++
15:19:34 <jroll> please
15:19:51 <TheJulia> #topic Priorities for this week
15:20:21 <TheJulia> So I went through and struck out items that were no longer needed on the list. Does anyone have any preference for things to add?
15:20:43 <TheJulia> #link https://etherpad.openstack.org/p/IronicWhiteBoard
15:20:46 <TheJulia> Line 118
15:20:53 <openstackgerrit> Merged openstack/ironic master: Switch ironic-tempest-dsvm-functional-python2 to zuulv3  https://review.openstack.org/600093
15:20:54 <rloo> i will point out that we got most of our priorities done this past week, yay! :)
15:20:55 <openstackgerrit> Merged openstack/ironic stable/rocky: Add admin documentation for rescue interface  https://review.openstack.org/606905
15:21:22 <TheJulia> rloo: that is very \o/
15:21:49 <TheJulia> I'd kind of like to make the ipxe stuff a priority this week, I should be able to revise the patches again today
15:22:28 * TheJulia wonders how specs are looking
15:23:01 <rloo> is there anything python3 related that needs reviewing?
15:23:11 <TheJulia> not that I'm aware of at the moment
15:23:21 <TheJulia> We could keep the bucket and I could go hunting
15:23:50 <TheJulia> mgoddard: jroll: with the ongoing ml thread re traits, should we wait on trying to get more eyes  on deployo templates?
15:23:58 <TheJulia> s/deployo/deploy/
15:24:05 <jroll> TheJulia: I would say wait at least until jay's proposal is up
15:24:08 <rloo> TheJulia: that's fine, we can wait til next week when etherpad/story is updated, will give us an idea of what is still outstanding wrt py3
15:24:25 <rloo> +1 wait for jay :)
15:24:35 <dtantsur> yeah, his idea looks interesting
15:24:51 <dtantsur> (if nobody comes at say, "No, this is not how you're supposed to use $THING" again :D)
15:24:57 <mgoddard> +1 I'd like to see what he is cooking up before proceding
15:25:33 <mgoddard> I'm a bit concerned that it's not going to tie into deploy steps, but we'll see
15:25:42 <TheJulia> We could do step visibility as a priority for the week?
15:25:54 <rloo> if people dont' have much to do, would a bug bash this week be useful? no, i'm not offering to organize or attend ;)
15:26:19 <rloo> i find it hard to believe we have 256 (out of 303) not triaged :-(
15:26:45 <jroll> mgoddard: I've been considering a blend where the operator publishes the config things in glance (which are basically deploy templates), and the user requests those templates
15:27:48 <TheJulia> rloo: agreed, I guess perhaps we should spend time on that this week
15:28:21 <dtantsur> jroll: yeah, we could store deploy templates there, and pass them to our provisioning API from the virt driver
15:28:33 <TheJulia> I guess we're good to move on to discussion if there are no objections?
15:28:54 <rloo> +1 wrt priorities. fine to have a relaxing week :)
15:29:04 <mgoddard> jroll: that could work
15:29:08 <rloo> (well, ipxe/pxe isnt' that relaxing)
15:29:24 <TheJulia> Eh, but it is a lot of clean up in the end :)
15:29:35 <TheJulia> (And something has to be hard in that list;))
15:29:56 <rloo> TheJulia: right :) (CleaningIsHard TM)
15:30:14 <TheJulia> :)
15:30:20 <TheJulia> Okay, lets move on then
15:30:25 <TheJulia> #topic Discussion
15:30:40 <rloo> we didn't announce it but did i see that we're removing a driver?
15:30:54 <jroll> rloo: yeah, oneview
15:31:11 <TheJulia> rloo: disclosed quite some time ago
15:31:19 <rloo> jroll: ok, worth mentioning :) ^^ done :)
15:31:36 <rloo> TheJulia: right, we said we were going to remove or something like that, but we've removed now?
15:31:37 <TheJulia> I put up two topics for discussion this week. The first because I wanted a wider discussion/audience to look at it.
15:31:44 <jroll> we have not removed yet
15:31:47 <TheJulia> not yet
15:31:52 <jroll> oh, discussion, not open discussion :)
15:32:01 <rloo> Oh, ok.
15:32:16 <jaypipes> jroll: isn't that precisely what my example solution was on the ML thread?
15:32:17 <TheJulia> A contributor proposed an update to IPA to enable parallel erasure of devices (This is a very good thing) - https://review.openstack.org/#/c/346723
15:32:18 <patchbot> patch 346723 - ironic-python-agent - Parallel erase disk devices - 13 patch sets
15:32:23 <openstackgerrit> Dmitry Tantsur proposed openstack/ironic master: [WIP] Completely remove support for Glance V1  https://review.openstack.org/606977
15:32:27 <dtantsur> JFYI ^^
15:32:36 <jaypipes> jroll: "I've been considering a blend where the operator publishes the config things in glance (which are basically deploy templates), and the user requests those templates" <-- pretty much what I was proposing, no?
15:32:52 <jroll> jaypipes: we're mid-meeting right now - it's pretty close, let's loop back :)
15:33:07 <jaypipes> yikes, sorry
15:33:13 <jroll> no worries
15:33:16 <TheJulia> jaypipes: No worries :)
15:33:34 <TheJulia> Anyway, Does anyone have any strong feels regarding parallel device erasure?
15:34:09 * TheJulia hears crickets
15:34:09 <chris_j> only strongly possitive here from the field
15:34:18 <rloo> omg, that was proposed in 2016
15:34:31 <TheJulia> chris_j: Awesome
15:34:51 <etingof> would that cause power consumption spikes and the whole DC lights out... ;)
15:34:52 <mgoddard> +1 from me
15:35:27 <rloo> TheJulia: is it the configuration thing that you wanted to discuss today?
15:35:28 <mgoddard> etingof: how power hungry is erasing?
15:35:35 <TheJulia> rloo: yeah, I think so
15:35:59 <etingof> mgoddard, like sustained writes, no?
15:36:23 <TheJulia> mgoddard: .5A@12v x $NUM_SPINNING_RUST?
15:36:36 <jroll> TheJulia: what needs to be discussed about the config?
15:36:49 <TheJulia> They just wanted to add it as a kernel argument
15:37:13 <jroll> well, it needs a way to be set
15:37:19 <TheJulia> We typically wire in through driver info to allow the conductor configuration to have a flag. I'm generally okay with that, but I think the config file would be better
15:37:22 <TheJulia> yeah
15:37:56 <jroll> yeah, I don't see any reason not to make it a global config
15:38:04 <TheJulia> I concur
15:38:21 <TheJulia> Okay I think that works then, I'll follow-up on my review and we can go from there
15:38:26 <mgoddard> what about if some nodes don't support secure erase?
15:38:49 <rloo> the default is true. should it be false to preserve existing behavior?
15:38:58 <jroll> mgoddard: this is regardless of secure erase or not
15:39:14 <TheJulia> rloo: also true, I agree that is likely best :(
15:39:36 <mgoddard> jroll: yeah, I'm just wondering if you might want different behaviour dependent upon whether nodes support secure erase. Perhaps not
15:39:55 <jroll> mgoddard: well, secure erase is instant, so this only seems useful for drives that don't support it
15:40:26 <TheJulia> Next topic: mgoddard and friends have proposed TENKS https://review.openstack.org/#/c/579583/ and have proposed it to live under ironic's governance. I'm +2 for this because we have a mix of VM creation scripts out there, and this would help drive consistency.
15:40:27 <patchbot> patch 579583 - ironic-specs - Add virtual Bare Metal Clusters spec - 10 patch sets
15:40:42 <jroll> mmm
15:40:53 <jroll> can we punt this one to the ML? I haven't looked at it at all
15:41:20 <jroll> unless you want to wait for me to go read code and code reviews and such :)
15:41:36 <mgoddard> jovial[m] has picked up the work on tenks after w-miller went back to university
15:41:59 <mgoddard> I could write something up for the ML
15:42:17 <TheJulia> Yeah, I think that would be good, I just searched the archives and it had not been mentioned on the lsit
15:42:31 <TheJulia> Okay, well I think we're good to move to Open Discussion
15:42:37 <mgoddard> should have done before now really, was just hoping for a few more spec reviewers
15:42:38 <jovial[m]> \o
15:43:02 <jroll> thanks mgoddard
15:43:18 <TheJulia> mgoddard: I might be looking at it through rose colored glasses thinking of all the spoons I'll get... but I thought the spec reads well.
15:43:38 <TheJulia> anyway, thanks mgoddard and greetings jovial[m]
15:43:49 <TheJulia> With that, Time for Open Discussion
15:43:50 <mgoddard> TheJulia: that's the plan - lure you in with the spoons
15:43:58 <TheJulia> #topic Open Discussion
15:44:07 <TheJulia> mgoddard: That is crafty.... Very craft.
15:44:14 <TheJulia> crafty
15:44:17 * TheJulia can't type today
15:44:28 <TheJulia> Does anyone have anything they wish to discuss?
15:45:47 <TheJulia> Oh, rloo, you might find this interesting https://review.openstack.org/#/c/606199/
15:45:48 <patchbot> patch 606199 - ironic-specs - A fresh way of looking at step retrieval - 1 patch set
15:46:06 <rloo> TheJulia: i put on my list for this week, thx :)
15:46:33 <TheJulia> no problem :)
15:46:48 <TheJulia> Well if nobody has anything else, we can proceed with taking over the world
15:47:06 <dtantsur> ++
15:47:36 <TheJulia> Excellent then!
15:47:52 <TheJulia> Have a wonderful day everyone! Thanks!
15:48:25 <jroll> jaypipes: so yeah, what I said about those templates is pretty much what you described, with some schema differences and the key fact that only the operator can publish templates (I'm not sure if this is true for your ML post)
15:48:34 <jroll> thanks TheJulia :)
15:48:42 <TheJulia> #endmeeting