17:00:17 #startmeeting ironic 17:00:17 Meeting started Mon Apr 4 17:00:17 2016 UTC and is due to finish in 60 minutes. The chair is jroll. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:18 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:21 hi everyone :) 17:00:21 The meeting name has been set to 'ironic' 17:00:21 o/ 17:00:21 o/ 17:00:22 o/ 17:00:23 o/ 17:00:32 #chair NobodyCam devananda 17:00:33 Current chairs: NobodyCam devananda jroll 17:00:46 o/ 17:00:46 o/ 17:00:50 o/ 17:00:51 I may have to run off for a few minutes during this, so you two be ready to take over :) 17:00:52 o/ 17:00:58 o/ 17:01:22 o/ 17:01:24 that... is apparently now 17:01:27 NobodyCam: mind running this? 17:01:53 I can Just wrapping up a internal meeting 17:02:10 * jlvillal thinks dtantsur also loves to run meetings ;) 17:02:11 morning all 17:02:23 * jroll apologizes, bbi 10-15 17:02:32 o/ 17:02:41 jlvillal runs the worlds fastest meetings :) 17:02:46 lol 17:02:52 as always the agenda can be found at : 17:02:56 #link https://wiki.openstack.org/wiki/Meetings/Ironic 17:02:59 * jlvillal just doesn't want it to be him. 17:03:37 o/ 17:03:46 lol 17:03:54 \o 17:04:10 #topic announcements 17:04:29 Summit is just around the cornor 17:04:52 lots of good Ideas on the etherpad: 17:04:59 #link https://wiki.openstack.org/wiki/Meetings/Ironic 17:05:01 gah 17:05:11 #link https://etherpad.openstack.org/p/ironic-newton-summit 17:05:25 o/ 17:05:49 looks like the 15th is our cut off 17:06:27 please take a minute to review and add your thoughts and priorites 17:07:07 I've just added a new topic 17:07:07 Some of the topics are without proposer. If you were the proposer, please, put you name into it :) 17:07:10 line 82 17:07:16 NobodyCam, what does the cut off mean? That people can comment up til then, or that ironic needs to provide the summit info. 17:07:30 do we have a final session count? 17:07:56 we do, but I don't remember it :) should be on the ML 17:07:58 NobodyCam: I think jroll sent a PDF with all sessions 17:07:58 rloo: I believe we will need to finilize our sessions by then 17:08:13 session rooms* 17:08:16 I think it is about scheduling the room and sessions. 17:08:18 and time slots 17:08:27 NobodyCam: ok, so maybe we want people's feedback by the 14th? 17:08:40 thiagop: yes and if I recall Wednesday was our BIG day 17:08:48 yep 17:08:49 rloo: ++ yes! 17:09:46 any other Announcements I am missing? 17:10:25 we have 14 topics already, don't we? 17:11:05 about that 13 /14 17:12:11 ok then lets move on to SUbteam status reports 17:12:22 #topic SubTeam Status 17:12:43 of course status can be found on hte white board: 17:12:47 #link https://etherpad.openstack.org/p/IronicWhiteBoard 17:13:21 around line 105 17:14:27 while folks take look, seems the gate is stable again :) 17:14:43 Thanks TheJulia for fixing the latest issue :) 17:15:18 TheJulia++ 17:15:29 dtantsur, wrt bugs. I haven't been looking. are the new bugs related to new features? 17:15:29 no problem :) 17:16:05 dtantsur, and are the bugs coming in faster than they are being resolved? 17:16:09 jlvillal: I see the note about the grenade testing, any more thoughts on how thats comming along? 17:16:28 rloo: https://bugs.launchpad.net/ironic/+bug/1555725 17:16:29 Launchpad bug 1555725 in Ironic "Node will crash if another OneView user removes the Server Profile of the node" [Undecided,New] - Assigned to Sinval (sinvalneto01) 17:16:35 rloo, 1. no, 2. yes 17:16:53 dtantsur: yuck. 17:16:59 NobodyCam: I can add a little more detail. 17:17:00 we have a separate counter for wishlist things. some untriaged rfe's might get to the list of bugs, but many of them are actual bugs 17:17:11 mat128: ? 17:17:34 jlvillal: awesome thank you :) 17:17:36 jroll, when you're back: we should probably start giving updates to things where "newton" is written currently :) 17:17:44 you were asking for bugs related to new features, I saw some OneView stuff recently 17:17:55 dtantsur: the 26 high bugs concern me. 17:17:57 I might be totally off, just went off by the title 17:18:04 rloo, they should.. 17:18:17 * devananda runs in late, wondering why his calendar notifications didn't go off for this meeting 17:18:24 dtantsur: is the increased bug reporting rate due to the fact that we cut a release and folks are testing that? 17:18:33 :) 17:18:39 NobodyCam, I'd say, due to increasing adoption of ironic 17:18:41 devananda, DST change? 17:18:41 mat128: oh. yeah, i am not surprised if there are bugs related to new features, just wondering how many were due to new :) 17:18:50 e.g. a lot of things reported by Mirantis folks 17:18:53 Ahhh :P my bad 17:18:59 possibly? 17:19:00 dtantsur: :) thats actually really awesome them :) 17:19:05 then* 17:19:12 o/ 17:19:16 mat128: no worries. 17:20:01 any other questions on the subteam status 17:20:42 yes 17:20:50 :) 17:20:52 several sections are empty right now, any updates on them? 17:21:10 I saw several no-update ones 17:21:12 actually nearly all are empty, counting also ones stating "newton" :) 17:21:34 is lintan here? 17:21:56 * dtantsur removed everything mentioning netwon, cause newton is going on right now 17:21:58 Shanghai time is 1:21am 17:22:29 devananda: jroll had to run this morning would you have a update for Multi compute hosts? 17:22:40 NobodyCam: none here 17:23:07 I think jroll is coordinating with the nova team to arrange for a joint session in Austin, but I don't have details yet 17:23:14 I dont see lucas or lintan for live upgrade 17:23:15 (that session also might be for the network discussion) 17:23:53 lucas is on vacation, IIRC 17:24:11 wasn't there some email discussion about docs? 17:24:11 ah thats right he is 17:24:33 rloo: do you have ref's to add there? 17:24:40 devananda: there is a couple of sessions we're disucssing for the networking intergration, a cross project ironic/neutron/nova session to disucss the future of networking in Ironic and then we've got a neutron/ironic session planned for the neutron track for disucssing vlan aware instances 17:24:56 krotscheck: any updates for web Ui? 17:24:59 NobodyCam, no, i saw it in passing. hmm, let me see if i can find it although it may take many minutes. 17:25:13 rloo: ack thank you :) 17:25:20 L32 -> https://etherpad.openstack.org/p/newton-neutron-summit-ideas 17:25:35 Number 12 -> https://etherpad.openstack.org/p/newton-cross-project-sessions 17:27:15 sambetts: sounds good 17:27:29 I'm back, so sorry 17:27:40 hey hey jroll 17:27:45 NobodyCam: here's part of the thread where jroll commented on doc plugin: http://lists.openstack.org/pipermail/openstack-dev/2016-March/090274.html 17:28:02 multi compute stuff - no session at the summit, we know what we need to do 17:28:04 any update on for the doc section of subteam status 17:28:11 :) 17:28:18 I mean, jroll would like a doc plugin for our install guide to be included with the 'official' documentation. 17:28:44 I've got no (coherent) update on docs right now 17:28:57 s/coherent/substantial/ maybe :) 17:29:08 jroll: just added the link rloo posted tothe doc for folks to look at 17:29:11 rloo: Thank you :) 17:29:17 cool 17:29:20 thank jroll :) 17:29:54 jroll: that was prob about the end of subteam 17:29:59 perfect 17:30:07 want to jump in for REF stuff? 17:30:17 thanks for the help NobodyCam :) 17:30:18 yep 17:30:19 RFE* :p 17:30:24 #topic RFE review 17:30:28 just one today 17:30:30 #link https://bugs.launchpad.net/ironic/+bug/1561137 17:30:32 Launchpad bug 1561137 in Ironic "[RFE] Add support for operators on root device hints" [Wishlist,New] - Assigned to Lucas Alvares Gomes (lucasagomes) 17:30:48 lucas is out today, but I figure we could look at this anyway 17:31:05 I think I'm okay without a spec on this, does anyone disagree? 17:31:21 +1 for just doing it 17:31:29 +1 Looks fairly trivial 17:31:30 I'd prefer "notin" over "ni" 17:31:33 +1 17:31:36 besides that... :) 17:31:37 +1 17:31:51 rloo devananda NobodyCam thoughts? :) 17:32:13 jlvillal: nothing should by default be "eq", like ComputeHostCapabilities filter 17:32:22 and backwards compat 17:32:42 i think it is fine w/o a spec. my only question is whether the abbreviations he is proposing is 'standard'. 17:32:45 mat128: I'm not quite understanding that. 17:33:02 http://docs.openstack.org/developer/nova/filter_scheduler.html 17:33:11 maybe we should use that syntax instead? 17:33:15 rloo: I think all of them are standard up to but not including 'in' 17:33:32 good question, https://github.com/openstack/nova/blob/master/nova/scheduler/filters/extra_specs_ops.py 17:33:43 ok, we should propose a cross-project spec to propose those abbreviations as openstack standards then :D 17:33:47 (not saying it's more readable, but it's already in "openstack" so yeah) 17:33:50 Lol 17:33:51 lol :| 17:34:01 * thiagop thinks that it scalated quickly 17:34:13 that is a good point though 17:34:13 rloo, things like eq, ne, lt, etc are used in ceilometer and got cargo-culted to inspector 17:34:21 aha 17:34:48 in that case I'm okay with it, especially if it's inspector already 17:34:53 I'm more or less fine with bashisms 17:34:55 I prefer this syntax to nova's 17:35:14 same here, was just pointing out a difference 17:35:20 will it be confusing to people/operators for ironic syntax to be different from nova's? 17:35:26 I'm fine with the "test" version / ironic-inspector 17:35:27 https://docs.python.org/2.7/library/functools.html 17:35:28 (in case of inspector I'm referring to http://docs.openstack.org/developer/ironic-inspector/usage.html#conditions ) 17:35:29 rloo: it already is :) 17:35:35 rloo: was just about to ask that 17:35:39 rloo: will it be confusing if it's different between ironic-* projects? 17:35:41 I guess I think of the "Python" names... 17:35:45 mat128: ++ 17:35:59 to end the argument, we can support both 17:36:01 I'd prefer matching with inspector over matching with nova, I think 17:36:06 it's not terribly hard to be honest 17:36:13 dtantsur, how hard would it be to change inspector to match with nova? 17:36:19 or do we even want it to match with nova? 17:36:30 well, then ceilometer is misaligned 17:36:31 tbh I was surprised by nova's syntax at first 17:36:39 so i think we should match with nova then. unless nova is an outlier? 17:36:46 rloo, this is part of introspection rules API, changing API is meh.. 17:37:04 geez, maybe we really need a cross-project spec :-( 17:37:12 rloo, if we're only talking about operators, not the whole root device hint syntax, then we can easily support both 17:37:20 oh god no 17:37:21 jroll: I believe aligning our API with Nova would be more valuable than aligning with Ceilometer, if that is a choice 17:37:25 can we support both 17:37:36 NobodyCam, YES 17:37:36 devananda: yeah, agree, but inspector throws a wrench in it 17:37:41 Do the nova ones require escaping of the URL that is generated? Or is that in the POST section. 17:37:50 I'm okay with supporting both, honestly 17:37:55 jlvillal, it's not in URL 17:37:55 ++ 17:38:03 dtantsur: thanks 17:38:08 NobodyCam: Sorry, was talking with wife. Yes, lots updates, much reviews needed. Wow ui. 17:38:16 Both seems reasonable compromise to me. 17:38:27 krotscheck: please update the whiteboard 17:38:38 :) 17:39:18 okay, so let's approve this, and we can go forward from there as to whether we support both or not 17:39:36 folks okay with that? 17:39:41 lets record in the comment, that we might need to support both variants 17:39:53 dtantsur: good Idea 17:39:53 yep, mind doing that dtantsur? 17:39:55 to not confuse people who are used to nova syntax 17:39:57 sure 17:39:59 jroll. not sure. shouldn't we agree on what it should support first? 17:40:00 +1 for approve and adding comment. 17:40:33 rloo: I feel like it should support the syntax proposed, at a minimum, because it matches inspector 17:40:49 jroll: +1 on this 17:41:00 we can have a separate item for nova syntax 17:41:02 jroll, but if inspector can change to support nova syntax, do you think ironic should support both too? 17:41:12 yes because the change will have made it to master 17:41:15 imo 17:41:27 rloo: I don't have strong opinions on supporting nova syntax or not 17:41:29 rloo: Not sure inspector should "change". maybe add additional support. But to change would break the API wouldn't it? 17:41:45 rloo, I can't guarantee inspector will change it in the observable future 17:41:48 jlvillal, that's what i meant. 17:41:49 yeah, api break is bad 17:42:02 i just don't like the idea of ironic supporting two syntaxes if we don't need it to do so. 17:42:02 we would have to go through a deperaction to change 17:42:11 so he alternative to this discussion is writing a small spec on this 17:42:16 which lucas said he doesn't mind doing 17:42:23 but I don't want to bikeshed for weeks in gerrit 17:42:27 me too 17:42:36 * mat128 thinks I should have kept the nova link to myself... lol 17:42:40 I don't see any problems with having a simple mapping == is the same as eq 17:42:50 mat128: I was going to say the same thing anyway, don't worry :) 17:43:17 i think we'd be doing ourselves a disservice by not mentioning/thinking of compatibility issues outside of ironic 17:43:21 going off into implementation land, but it is a dict with mappings. '==' and 'eq' can both map to the same thing. So seems easy to support both. 17:43:26 okay, why don't we just approve with the caveat we must support both? 17:43:49 I'm okay with aproval and comment that both format should be supported 17:43:57 +1 17:44:01 * dtantsur has actually written such comment already 17:44:06 :) 17:44:07 and the ni -> notin 17:44:10 sorry for being a bit too fast :) 17:44:24 lol 17:44:30 If 'ni' is how inspector does it, then I'm okay with it being 'ni' :) 17:44:33 dtantsur: is 'ni' an inspector thing today? 17:44:43 +1 to both being implemented. 17:44:44 jroll, "not-in" 17:44:51 not-in then 17:45:09 dtantsur: okay, can you actually list out what inspector does, we can map to nova, and that'll be the things we support? 17:45:27 jroll, http://docs.openstack.org/developer/ironic-inspector/usage.html#conditions 17:45:44 oops, I lied about not-in 17:45:49 heh 17:45:50 confused with a different one, sorry folks 17:45:54 * dtantsur is probably tired 17:46:01 not not-in then 17:46:09 not-not-in 17:46:10 so maybe this needs some work to align with inspector and nova, then 17:46:23 jroll, we can start with a simpler subset 17:46:27 let's ask lucas to re-propose the syntax that works well for both 17:46:31 and go from there? 17:46:40 ++, thx jroll 17:47:03 * jroll comments 17:47:06 thnx 17:47:53 k 17:47:58 #topic open discussion 17:48:01 so before people shout things 17:48:03 I have a thing 17:48:18 this week and next, I'll be pretty intermittent availability on irc 17:48:36 I'm moving from california to michigan, about 2500 mile drive (and packing, and dealing with movers, etc) 17:48:39 oh thats right your moving 17:48:43 wow 17:48:48 so I'll be checking irc once/twice a day, I'll be keeping up on email 17:48:49 :) 17:48:54 driving? Ouch! 17:48:57 jroll: awesome :) 17:49:13 quickest way to reach me will be via google hangouts (jim@jimrollenhagen.com) 17:49:23 but in the car, may still be a couple hours response :) 17:49:37 Safe travels and congrats 17:49:38 jroll, good luck! 17:49:39 just wanted everyone to be aware 17:49:45 thanks, it'll be a ride :P 17:50:01 jroll, that's a better time zone too :) 17:50:05 jroll: should I plan on covering hte meeting next week just to be sure? 17:50:08 rloo: :D 17:50:14 NobodyCam: oh yeah, if you or deva could cover 17:50:18 jroll: enjoy the drive! and congrats :) 17:50:19 ++ 17:50:21 I'm going to try to jump in, because session planning 17:50:27 but cell service may be spotty 17:50:37 have a safe trip 17:50:41 ++ 17:50:49 good luck with the move :) 17:51:00 thank you all 17:51:07 * mgould checks: that's about 3x the length of my longest ever road trip :-) 17:51:10 now, anything more interesting for the last 9 minutes? 17:51:15 Neutronic! ;) 17:51:15 krtaylor and I have an announcement. As you may know, we've been working on an ironic CI system for KVM on Power that uses actual baremetal nodes as targets (rather than guests). It's been running well and wanted to know if there were any objections to us tuning on commenting on ironic patches? 17:51:17 mgould: about double for me 17:51:27 mjturek1: please do! 17:51:30 jroll, i was going to ask you about session planning. we had mentioned it earlier, with a deadline for feedback by 13th apr. 17:51:32 jroll: awesome :) 17:51:38 \o/ 17:51:41 and congrats, safe travels 17:51:47 jroll: did you want to actually discuss sessions next monday at th emeeeting? 17:51:55 * jlvillal drove from Biloxi, Missippi to Santa Rosa, California round trip when he was younger. Never again. 17:51:56 maurosr: \o/ 17:51:56 mjturek1: nice :) w00t 17:51:59 rloo: I mentioned in channel this morning (and would have mentioned in announcements if I was here), I'm going to rough draft this week, then discuss and finalize on monday 17:52:09 maybe final tuesday/wednesday 17:52:21 jroll: good to know. thx. 17:52:26 mjturek1, you heard 'em, lets turn it on! 17:52:52 krtaylor: btw, can you compile me a list of third party folks and CI status for wednesday's qa meeting? 17:53:04 and I can send out some emails and whatnot 17:53:10 BTW, we have a bare metal pool management that we have written, that we are looking at contributing also 17:53:33 yep, it's called molteniron and it's pretty integral to our CI 17:53:41 jroll, will do, thats on my list, I'll move it up 17:53:43 aren't we like, building one of those? :) 17:53:57 more seriously, that would be great 17:53:59 yes, thats why I wanted to mention it 17:54:21 other projects have shown interest also 17:54:23 heh, yeah we're not sure where it might overlap, but figured it'll be useful (at least initially :)) 17:54:38 hey, it works :) 17:54:40 On that topic, we have totally virtual baremetal lab that we still have to contribute 17:54:47 PDUs, IPMI and everything 17:54:53 sweet! 17:55:01 mat128: wow 17:55:09 even network equipment 17:55:15 very cool! 17:55:25 yeah, we're very test-driven and we had to make everything work in a reliable fashion 17:55:25 * jlvillal isn't quite sure what "totally virtual baremetal lab" means 17:55:41 similar to the devstack nodes that we manage with the ssh power driver 17:55:45 but, totally independent 17:55:49 you dont need any special driver 17:55:57 you use the regular SNMP driver, etc. 17:56:02 mat128: using lucas virtual bmc thing? 17:56:27 I think we are still using pyghmi for the virtual BMC, but we've had nothing but problems with that 17:56:40 which kinda segways to a critical bug of ours 17:56:40 https://bugs.launchpad.net/ironic/+bug/1531981 17:56:41 Launchpad bug 1531981 in Ironic "ipminative.NativeIPMIPower fails on reboot" [High,In progress] - Assigned to Jonathan Provost (jprovost-sh) 17:56:42 mat128: awesome \o/ 17:56:50 FYI, anything CI related, for quick contribution just to get it out in the open, we can push it to the openstack third party repo I created, ping me for details 17:56:50 pyghmi/ipminative doesnt work at all 17:57:04 krtaylor: will do 17:57:14 is there anyone using ipminative our there? 17:57:22 apparently not :P 17:57:30 we initially prefered a pure python approch, but it doesnt even work :P 17:57:41 should we suggest a spec to remove it? 17:57:48 well... 17:57:52 why not fix the bug? 17:57:59 *two minutes* 17:58:06 krtaylor: might you send the link of the repo? 17:58:09 we fixed it, but pyghmi doesnt even work 17:58:11 lol 17:58:15 maybe it's just on our side 17:58:16 (though, to be fair, if nobody puts up third party CI this cycle it will be dropped) 17:58:18 mat128: ^ 17:58:32 ok interesting 17:58:40 we can test all of ipmi* with virtualdatacenter 17:58:41 mat128: if you look at pyghmi commits, there's basically one guy working on it, could email him 17:58:46 or file a bug against pyghmi 17:58:48 gabriel-bezerra, https://github.com/openstack/third-party-ci-tools 17:58:57 krtaylor: thank you 17:59:06 jroll: dropped or moved to https://github.com/openstack/ironic-staging-drivers/ ??? 17:59:21 NobodyCam: if someone wants to propose it there, they are welcome to 17:59:27 ack 17:59:34 I only have time for working on ironic >.> 17:59:48 mat128: fwiw, pyghmi was written by IBM initially 17:59:55 alright, we're at time. thanks everyone 18:00:00 #endmeeting