17:00:55 <NobodyCam> #startmeeting Ironic 17:00:55 <NobodyCam> #chair devananda 17:00:55 <NobodyCam> Welcome everyone to the Ironic meeting. 17:00:55 <openstack> Meeting started Mon May 4 17:00:55 2015 UTC and is due to finish in 60 minutes. The chair is NobodyCam. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:58 <openstack> The meeting name has been set to 'ironic' 17:01:00 <openstack> Current chairs: NobodyCam devananda 17:01:08 <devananda> g'morning/afternoon/evening! 17:01:14 <NobodyCam> Of course the agenda can be found at: 17:01:15 <NobodyCam> #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 17:01:26 <NobodyCam> #topic Greetings, roll-call and announcements 17:01:26 <NobodyCam> Roll-call: Who's here for the Ironic Meeting? 17:01:30 <zzzeek> o/ 17:01:32 <pshige> o/ 17:01:38 <jlvillal> \o/ 17:01:38 <Shrews> o/ 17:01:42 <rloo> o/ 17:01:45 <mariojv> \o 17:01:48 <devananda> \o 17:02:06 <NobodyCam> howdy y'all 17:02:20 <NobodyCam> I believe we have several folks on holiday today 17:02:40 <NobodyCam> and even a few in other meetings so we may be a little light 17:02:42 <kan__> \o 17:02:58 <JoshNang> o/ 17:03:05 <NobodyCam> #topic announcements: 17:03:14 <Nisha> o/ 17:03:22 <NobodyCam> I believe devananda will be traveling this week 17:03:39 <devananda> indeed I will 17:03:52 <devananda> that means I'll actually do some reviews 17:04:04 <devananda> because gertty rocks and planes have no internet 17:04:05 <jroll> \o 17:04:09 <NobodyCam> DiscoverD and Bifrost are up for project review in tomorrows TC meeting 17:04:11 <devananda> (well, planes over the pacific) 17:04:11 * jroll is only half here right now 17:04:32 <jroll> NobodyCam: what kind of project review? 17:04:54 <NobodyCam> to be included in the OpenStack repo tree 17:05:06 <devananda> i proposed them to infra/project-config to move into openstack/ namespace 17:05:18 <jroll> word 17:05:28 <pshige> oh 17:05:31 <devananda> er, rather, I mean I proposed them to the governance tree 17:05:36 <NobodyCam> :) so major congrats to all! 17:05:47 <NobodyCam> and good luck! 17:05:48 <devananda> I still need to create the project/config changes to go along with that ... 17:06:02 <devananda> (unless anyone beats me to it) 17:06:24 <NobodyCam> devananda: will you have time today for that? as I know your leaving to fly around the plant tonight 17:07:00 <devananda> NobodyCam: at least for a first pass, yes 17:07:06 * rloo wonders where 'Bifrost' name came from 17:07:25 <NobodyCam> thor? 17:07:37 <devananda> rloo: the recesses of the collective unconscious? 17:07:57 <NobodyCam> was that not name of the bridge thing 17:08:02 <devananda> anyone else have announcements, or shall we move on? 17:08:08 <Shrews> devananda named it, so i suspect he was solidly drunk at the time :) 17:08:11 <rloo> Bifrost (pronounced roughly “BEEF-roast;” Old Norse Bifröst) is the rainbow bridge that connects Asgard, the world of the Aesir tribe of gods, with Midgard, the world of humanity. 17:08:27 <devananda> rloo: that's the one 17:08:31 <NobodyCam> :) 17:08:48 <NobodyCam> status report ? 17:08:50 <devananda> Shrews: I neither confirm nor deny that 17:09:24 <devananda> NobodyCam: yah. status report 17:09:30 <NobodyCam> #topic SubTeam: status report 17:10:00 <NobodyCam> of course details may be posted on the white board 17:10:04 <NobodyCam> #link https://etherpad.openstack.org/p/IronicWhiteBoard 17:10:43 <rloo> wrt oslo, Ghe updated a few patches :-) 17:10:47 <NobodyCam> looks light today.. new Bug numbers 17:11:09 <devananda> yea, I dont see any updates from the driver folks 17:11:24 <devananda> and dtantsur is on PTO this and next week 17:11:41 <NobodyCam> any comments or questions on status releated things? 17:12:05 <devananda> given the usually non-contentious nature of oslo changes, it would be great to get some of thoes changes landed inthe next week or two 17:12:11 <rloo> is it still worth sending out weekly emails about this? 17:12:17 <devananda> like things that were oslo-lib-ified in kilo but didn't get merged into ironic yet 17:12:29 <devananda> rloo: right now, probably not. however... 17:12:47 <rloo> +1, esp the logging changes that touches many files 17:12:57 <BadCub> sorry I am a bit late 17:13:06 <devananda> i expect we'll be talking about drivers a bunch at the summit, so post-summit, I hope we'll have more activity from driver authors/maintainers/testers 17:13:30 <rloo> devananda: ok, will put it on hold til after the summit. 17:13:30 <devananda> wanyen: any news on iLO CI? 17:13:37 <NobodyCam> devananda: ++ 17:14:14 <wanyen> devananda, all patches for 3rd-party CI were submitted for review and test run works. 17:14:15 <devananda> I know we have had some specs for drivers recently approved, eg. iRMC vmedia deploy using NFS/CIFS 17:14:57 <devananda> wanyen: do you have links available to share? 17:15:10 <rloo> the iRMC vmedia deploy was mentioned last week: http://lists.openstack.org/pipermail/openstack-dev/2015-April/062866.html 17:15:21 <NobodyCam> we really will need to talk about the Driver composition matrix at the summit. 17:15:25 <devananda> NobodyCam: yah. 17:15:49 <devananda> wanyen: or rather, is there anything we can do // has anything changed since last week? 17:16:41 <devananda> rloo: oh, right :) 17:16:55 <devananda> well, let's move on then. sounds like nothing happened last week 17:17:01 <NobodyCam> ack 17:17:03 <NobodyCam> #topic Summit planning 17:17:05 <NobodyCam> #link https://etherpad.openstack.org/p/liberty-ironic-design-summit-ideas 17:17:49 <NobodyCam> we've got some great ideas up so far 17:17:57 <devananda> we discussed this at the last weekly meeting, but since then I've talked with the ML2 subcommittee chair (sukhdev) 17:18:14 <devananda> and he's investigated the networking issues more deeply -- there's a POC up, even 17:18:29 <NobodyCam> there are also several cross - project things (neutron / ml2 as example) 17:18:34 <devananda> tldr; it looks like most of the work required is in ironic and nova, so they've asked for us to host the session 17:18:49 <wanyen> devanada, https://etherpad.openstack.org/p/IloDriverGateTesting 17:19:07 <NobodyCam> #link https://etherpad.openstack.org/p/IloDriverGateTesting 17:19:40 <devananda> wanyen: ty. I've added to the etherpad for future reference 17:19:54 <rloo> wrt Friday: why RAID and Zapping spec reviews, as opposed to other specs that have been proposed? 17:19:55 <wanyen> devanana, np 17:20:18 <NobodyCam> devananda: do we still need to sit down with hte nova folk RE: Gate testing stuff? 17:20:47 <devananda> rloo: in part because those two are central to finishing the state machine, and in part because they've been contentious in the past 17:21:16 <devananda> rloo: are there other specs that you think would be a better use of time? 17:21:19 <rloo> devananda: ahh, ok. I thought I was the only one that was negative on the RAID one. 17:21:37 <rloo> devananda: I was thinking that it would be good to slot some time to review specs. 17:21:46 <devananda> rloo: ++ 17:22:36 <rloo> devananda: not mention specific ones now, but see what specs are outstanding/contentious at the time. 17:23:11 * zzzeek has to brb, sorry 17:23:12 <devananda> for the six working sessions, how do folks feel about that allocation? 17:23:23 <devananda> I'd like to swap in networking, but not sure which one to remove 17:23:45 <rloo> devananda: does Scaling the Team Structure require 45 minutes? 17:23:49 <NobodyCam> 3.9 17:23:52 <devananda> NobodyCam: I dont think that's worth a whole session, and I dont have the sense from Nova that they do either 17:24:27 <rloo> Is 2.6 where we discuss more frequent releases? 17:25:11 <devananda> NobodyCam: sdague and I may go find a corner to discuss it, but basically we need to continue improving our testing, and just clearly communicate which test(s) are exercising portions of Nova that no other nova-specific tests cover 17:25:12 <NobodyCam> devananda: ack. do we need a session on Logging Standards cleanup ? or an event? 17:25:41 <devananda> NobodyCam: there is a cross-project spec on logging standards, but I have the impression folks may not be aware of it, so I think it's worth a discussion 17:25:57 <devananda> NobodyCam: but not a whole session, which is why I put it on the friday agenda 17:26:03 <NobodyCam> devananda: ahh okay!!!! 17:26:26 <devananda> rloo: hmm. yes 17:26:30 <BadCub> devananda: do you think we can combine 2.5 & 2.6 into one session? 17:26:32 <devananda> rloo: think we can merge those two? 17:26:34 <devananda> :) 17:26:37 <devananda> done 17:27:21 <rloo> devananda: what's done? if 2.6 includes frequent releases, no, can't merge. i think that will take time. 17:27:41 <rloo> darn, he merged it 17:27:58 <NobodyCam> lol 17:28:21 <devananda> BadCub: didn't you want to talk about the release process itself? 17:28:32 <NobodyCam> rloo: is that something we need to talk to external folk about or can we discuss over say dinner? 17:28:33 <rloo> can't we do functional testing 2.2 on Friday? 17:28:50 <BadCub> devananda: yeah. How we can make it less "crunchy" and stressful for reviewers 17:28:51 <rloo> NobodyCam: if we're talking about more frequent releases, doesn't that affect the community? 17:29:19 <rloo> NobodyCam: I'm not quite sure what that means though (more freq releases). did jroll propose something yet? 17:29:31 <devananda> rloo: it affects developers (us) but not users or operators (except those that want to use more frequent releases) 17:29:46 <devananda> it is an internal process discussion for ironic developer community 17:29:48 * zzzeek back 17:30:03 <devananda> so it's a thing we should put in the "working sessions" area 17:30:06 <BadCub> But I think that will include spec review cut-off's (e.g. when we stop approving/at what # do we stop approving/etc) 17:30:12 <rloo> devananda: oh. so maybe we can move that to friday. or maybe we can have that session as is and continue on fri if need be. 17:30:18 <devananda> BadCub: indeed. it's related 17:30:25 <Sukhdev> devananda: sorry - just signed in…I see item 2 says need to add networking session. 17:30:39 <devananda> Sukhdev: hi! yes - what time slot on thursday can you join us for that? 17:31:28 <Sukhdev> devananda: I have to check what other sessions are planned that that, but, this is priority - so, I will put this on top of my list 17:31:51 <Sukhdev> devananda: I will put together an etherpad sometime this week and send out the link 17:32:28 <devananda> BadCub, rloo: on the team structure & release cycle, there's a couple different points here: how often we release, how we coordinate designs (specs, BPs, etc), and how we structure review subteams (IPA, discoverd, specs, bifrost, ...) 17:32:58 <NobodyCam> awesome Thank you Sukhdev, could you also post on the liberty-ironic-design-summit-ideas etherpad 17:33:01 <devananda> Sukhdev: ack. It's a priority for us, too. so if I need to move working sessions around to accomodate, lmk. 17:33:13 <BadCub> devananda: indeed 17:33:27 <rloo> devananda: I'm thinking that might take more than 45 min. The more general something is, the more people have opinions. that's my opinion anyway ;) 17:33:50 <devananda> Sukhdev: we have 6 slots, 11am - 5pm. 17:34:16 <devananda> rloo: heh, right. we can spill that over to friday, but I think keeping that discussion narrow on thrusday will be helpful 17:34:28 <rloo> devananda: +1 17:34:39 <Sukhdev> devananda: checking schedule 17:34:48 <devananda> of those 3 things, the first 2 are most tightly coupled and most relevant 17:34:52 <devananda> to the largest # of people 17:36:22 <Sukhdev> devananda: I can make anytime - as of now... 17:36:43 <devananda> jroll: huh. I dont actually see a section on the 'pad with notes on release cadence 17:37:04 <devananda> Sukhdev: ok. I'll put it just after lunch, but can move if needed. let me know 17:37:09 <rloo> according to the schedule, sessions are 40 min, not 45. if we don't include the breaks. 17:37:18 <BadCub> devananda: If memory serves, you, me and jroll were supposed to put aside some time to discuss ^ 17:37:27 <Sukhdev> devananda: Sounds good - thanks. 17:37:40 <devananda> BadCub: you're probably right 17:37:54 * BadCub checks notes to be sure 17:38:15 <devananda> rloo: breaks are just there for us to use them, right? ;) 17:38:30 <NobodyCam> lol 17:38:48 * NobodyCam need his smoke breaks 17:39:02 <BadCub> devananda: yep.. We were.. I got sidetracked with downstream stuff though, so my bad for not staying on top of that 17:39:08 <rloo> anyone know what 2.4 Scheduling Capabilities, is going to involve? 17:39:23 <BadCub> NobodyCam: ++ on ^^ for sure! 17:39:29 <devananda> rloo: "Capabilities" 17:39:49 <devananda> it's pretty light onthe 'pad, though 17:39:52 <devananda> line 202 17:39:53 <rloo> devananda: wrt nova? 17:40:17 <rloo> devananda: oh, ok, thx. looking. 17:40:53 <devananda> yea, wrt Nova and our API for exposing them to the scheduler 17:41:02 <devananda> we need more concrete proposals there, though 17:41:18 * rloo seems to recall a session before about capabilities. can't recall what we decided. hope this isn't a rehash of that. 17:41:46 <devananda> rloo: ++ 17:42:34 <wanyen> capabilitites spec was propsoed inJuno but bumped to Liberty 17:42:48 <rloo> wanyen: does it need to be discussed in a session? 17:43:00 <rloo> wanyen: there's time set aside on Fri to discuss specs. 17:43:28 <wanyen> rloo, I did not propose that session. 17:43:41 <NobodyCam> oh devananda see e0ne's comment in ironic's channel 17:43:53 <rloo> wanyen: thx for clarifying. 17:44:06 <e0ne> hi 17:44:09 <devananda> hi! 17:44:30 <devananda> e0ne: is there any way to put the cinder-ironic discussion on wednesday? 17:44:42 <e0ne> thursday is also all of cinder's sessions :( 17:45:02 <e0ne> i'll ask thingee 17:45:09 <NobodyCam> FYI * there is one Open discussion Item on the agenda. and 15 minutes left in the meeting * 17:45:16 <wanyen> rloo: however, capabilities can be very helpful to provide better cotnrol of workload placement and ironic can do work to support capabilities taht needs actions e.g., setting correct firmware setting adn RAID level 17:45:40 <rloo> wanyen: I doubt that anyone thinks that capabilities are not useful. 17:45:59 <devananda> e0ne: ah, I see. then I think we need to find a slot where we can split one of the teams :-/ 17:46:01 <rloo> wanyen: I am just wondering what the Capabilities session is for, and whether we need to have a session on it. 17:46:34 <devananda> wanyen: if it's not a contentious topic -- if folks already reviewed and approved the spec, for instance -- then there's probably no need for a sessoin 17:46:52 <rloo> devananda: we don't know who proposed that session 17:47:08 * devananda removes it 17:47:19 <Nisha> rloo, i think it would be better to have a session as if we need any changes in nova(which i think will be required for act upon capabilities), then we can propose them early rather than later 17:47:24 <wanyen> rloo: one potential topic is how to define a few common capabilities, e.g., capabilities for compute node, control node, etc 17:47:49 <e0ne> devananda: we could move it to wednesday 17:47:57 <e0ne> devananda: " <thingee> preferred time?"? 17:48:03 * rloo just doesn't want to have another capabilities session like the last summit 17:48:11 <devananda> e0ne: ironic doesn't have any slots on weds. 17:48:23 <wanyen> The topic that I am interested in is to support of capabilities with multiplepossible values 17:48:40 <devananda> wanyen: that's more of a nova scheduler question, isn't it? 17:49:04 <wanyen> devananda, yes, it's related to Nova but ironic virt driver may need to change 17:49:09 <Nisha> devananda, it should be in ironic virt driver 17:49:37 <devananda> right -- so that should have been proposed to Nova, since it changes the nova virt driver & scheduler APIs 17:50:07 <Nisha> devananda, but the approach has to be first discussed in ironic so that we know what we are proposing to nova...isnt it 17:50:24 * devananda thought we had previously agreed on the way that that data would be stored in Ironic 17:50:26 <Nisha> because that will/may require further changes in ironic 17:50:35 <devananda> oh 17:50:44 <e0ne> devananda: i'll notify you about time of session once i'll know it 17:50:46 <devananda> rloo: so that's why it needs a discussion, apparently 17:50:55 <NobodyCam> * 10 minutes * 17:51:09 <wanyen> devananda, JayF subitted a capabilities spec for Juno but that spec was not under much reivewed or approved. JayF voluntarily bump the spec to Liberty 17:51:16 <devananda> ok, folks, so please take some time to comment on the etherpad 17:51:22 <wanyen> s/subitted/submitted 17:51:26 <devananda> all of this should be discussed up there throughout the week 17:51:33 <devananda> so we dont have to wait until next Monday to discuss it again 17:51:39 <Nisha> yes as of now we store all the capabilities while actually only those capabilities should be copied to ironic node whic require any action from driver 17:51:44 <pshige> e0ne: thx 17:51:53 * BadCub wont be in attendance next Monday meeting 17:51:54 <rloo> devananda: by 'discuss', we just add our comments to the etherpad? 17:51:56 <Nisha> if it is just scheduling property we dont need to clutter instance_info 17:51:59 <devananda> rloo: yes 17:52:06 <rloo> devananda: okey dokey 17:52:18 <devananda> I will "finalize" the schedule next week 17:52:34 <devananda> which really means deciding what we talk about thursday 17:52:41 <NobodyCam> good to move on? 17:52:57 <NobodyCam> is ken__ here? 17:53:01 <devananda> as friday is free-form and we'll no doubt think of more things to talk about then :) 17:53:05 <devananda> NobodyCam: ++ 17:53:09 <NobodyCam> gah kan__ 17:53:34 <kan__> hi 17:53:36 <NobodyCam> #topic Open Discussion 17:53:44 <devananda> #info I will "finalize" the Thursday summit schedule next week. Please *PLEASE* take some time to discuss and add comments on the etherpad: https://etherpad.openstack.org/p/liberty-ironic-design-summit-ideas 17:53:50 <e0ne> devananda, pshige: Wednesday, May 20 • 5:20pm - 6:00pm 17:53:58 <e0ne> http://libertydesignsummit.sched.org/event/57b44975517778e6450613b5c0ca2681#.VUex7608aHo 17:54:00 <NobodyCam> Hi kan__ you have a text thing on the agenda 17:54:06 <devananda> e0ne: awesome, thanks. adding a note to our 'pad 17:54:13 <e0ne> thanks! 17:54:31 <NobodyCam> #link https://bugs.launchpad.net/ironic/+bug/1446508,https://github.com/openstack/oslo.db,https://github.com/zzzeek/sqlalchemy 17:54:31 <openstack> Launchpad bug 1446508 in Ironic "Use 'last_error' and 'maintenance_reason' as the sort key in node list will error with db2" [Low,Triaged] 17:54:31 <kan__> yes, I have a bug im ironic :https://bugs.launchpad.net/ironic/+bug/1446508 17:55:35 <kan__> when using some sort keys with db2, there will error for the 'text' data type is transfered to 'clob', and it can not be used in 'order by' sql command 17:56:21 <rloo> kan__: there are a lot of comments in this bug. which comments should we be looking at? (What is the question?) 17:56:23 <kan__> talked with mike bayer, try to find some ways to fix it 17:56:40 <devananda> e0ne: fyi, i believe cheddar allows you to put that inthe Ironic track, which might help folks find it 17:56:52 <devananda> e0ne: by using the "Also appears in..." box 17:57:03 <kan__> comment 16 and 19 17:57:21 <zzzeek> the simplest solution involves prodiucing an attribute that is safely sortable, there just needs to be a translation of this name when paginate_query() is called 17:57:43 <devananda> kan__: first, why is anything sorting by those fields? 17:57:45 <kan__> mike gives some advices about how to fix in comment 16 and 20-22 17:57:48 <devananda> kan__: second, db2? 17:58:27 <rloo> devananda: when you do a node-list (or chassis-list), you can specify the sort key. 17:58:44 <NobodyCam> * two minuts * 17:58:47 <rloo> devananda: I guess we allow users to specify last_error and maintenance_reason 17:59:20 <devananda> rloo: from a UX perspective, those are useless columns to sort by. why don't we not allow it? 17:59:22 <zzzeek> devananda: db2 is a backend we seem to be supporting in oslo.db, I would assume some are using it 17:59:34 <NobodyCam> devananda: ++ 17:59:40 <rloo> devananda: that's what i was wondering. would be an easy way to solve this! 17:59:53 <kan__> I have another bug https://bugs.launchpad.net/python-ironicclient/+bug/1446146 to disable some invalid sort keys, but the last_error and maintenance_reason is able 17:59:53 <openstack> Launchpad bug 1446146 in python-ironicclient "Meaningless sort key should disabled in the list command" [Wishlist,In progress] - Assigned to Kan (kansks) 18:00:03 <zzzeek> devananda: as far as that particular col being a sort field, I agree it would be easier for it to not be sortable. the discussion on the bug report seems to feel that sorting by error code is very useful and I can see how it might be, though if i were searching for a particiular error I’d just use grep 18:00:23 <NobodyCam> * thats time... can we contine in the Ironic channel 18:00:34 * zzzeek has given all the info he has :) 18:00:38 * BadCub wanders off 18:00:41 <devananda> intersting. yea, I dont mean t otroll, just catching up on the bug 18:00:58 <NobodyCam> thank you all... great meeting 18:00:59 <zzzeek> devananda: feel free to jump on the bug :) 18:01:07 <devananda> thanks all! 18:01:27 <NobodyCam> !endmeeting 18:01:27 <openstack> NobodyCam: Error: "endmeeting" is not a valid command. 18:01:34 <pshige> thanks 18:01:42 <NobodyCam> #endmeeting