15:01:41 #startmeeting ironic 15:01:42 Meeting started Mon Oct 29 15:01:41 2018 UTC and is due to finish in 60 minutes. The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:43 o/ 15:01:44 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:46 The meeting name has been set to 'ironic' 15:01:48 o/ 15:01:50 o/ 15:01:56 o/ 15:02:01 o/ 15:02:11 o/ 15:02:19 \o 15:02:41 o/ 15:02:47 o/ 15:02:50 o/ 15:02:53 * dtantsur is here but will have to go afk for a while really soon 15:02:55 Good morning everyone! And welcome to another ironic meeting! 15:03:17 dtantsur: then we should postpone the clean step discussion? 15:03:30 We have a few things on our agenda today, one of which is the just mentioned discussion 15:03:32 #link https://wiki.openstack.org/wiki/Meetings/Ironic#Agenda_for_next_meeting 15:03:43 #topic Announcements/Reminders 15:03:47 rloo: I nearly feel like it deserves a meeting of its own 15:04:06 but feel free to go ahead without me. my head does not work properly after all these news anyway :) 15:04:22 dtantsur: oh, right. go celebrate... ? :) 15:04:22 * TheJulia is in the same space as dtantsur at the moment 15:04:29 * iurygregory too 15:04:42 I have no announcements or reminders for today. Does anyone have anything they wish to raise? 15:04:50 sigh. maybe lets just do a quick meeting then. 15:04:55 o/ 15:05:19 #info Looks like we're going to try and do a team gathering in Berlin during the summit. 15:05:22 #link https://doodle.com/poll/iw5992px765nthde 15:05:23 o/ 15:06:02 That just so happens to be the one action item from last week, so I believe we can jump directly to reviewing status? 15:06:19 jump + 15:06:20 yeah, please vote 15:06:33 and yes, please vote before Friday 15:06:51 Okay, if nobody has anything else, we shall jump to status review 15:07:05 #topic Review subteam status reports 15:08:12 I updated a few things on Friday that I'm aware of 15:08:41 It looks like kaifeng and I have started proposing CI job updates to clean up our templates for zuulv3 and also get integration testing with python3 going 15:08:48 #link https://github.com/dtantsur/ironic-proxy how a federating ironic API proxy could look (early WIP) 15:09:10 Of which, I think it would be good to tag them all with a unique topic and put them up for review this week (which I can do after the meeting) 15:10:39 * kaifeng used zuul as topic, but actually contains python3-first 15:11:00 Yeah, since we kind of merged the efforts together (which does actually make a l ot of sense) 15:11:37 I'm intending to continue that this week, and I'm going to keep working on ipv6'ifying one of our CI jobs 15:11:45 I'm close. So close. 15:12:27 wrt deploy templates -- is the next step to review/approve the spec? 15:12:42 It looks like a lot of our smaller items just need some patch reviews this week 15:12:47 rloo: I believe so yes 15:13:41 and i'm not sure about the comment about jay's proposal. is there more info as to his proposal. i thought he was going to write a spec. is it an email thread? is it captured in deploy templates spec? 15:14:12 by 'not sure', i don't know what the action item is for that... 15:14:34 rloo: I think we partially captured that in the spec, but we could likely write more. We should make that contingent upon the existing spec but I can do a follow-up 15:14:47 TheJulia: ok, thx for explaining. 15:15:15 rloo: I think my comments on the spec reflect that as well, so I think reviews are the more important thing right now 15:15:40 etingof: Quick question, is it just the two patches your working with regarding ipmitool execution? 15:15:53 etingof: I'm kind of feeling like larger discussion may be required so punting to next week may be best 15:15:59 TheJulia: thx, i'll try to look at that spec. 15:15:59 but we could encourage reviews for feedback? 15:16:04 rloo: thanks 15:16:51 dtantsur|brb: wrt federation capabilities, did we approve a spec about it? 15:17:11 rloo: we're kind of doing Proofs of concept to see what actually sticks 15:17:30 TheJulia, there are many related patches, but the one I currently propose to make ironic more predictable is this -- https://review.openstack.org/#/c/610007/\ 15:17:30 patch 610007 - ironic - Kill misbehaving `ipmitool` process - 2 patch sets 15:17:33 TheJulia: ok, i just realized I mixed that up with 'Conductor role splitting'... 15:18:04 TheJulia, +1 for a large discussion 15:18:05 rloo: I suspect that is going to take a back seat until later in the cycle. Depends on Jim's time I guess 15:18:29 etingof: Okay, if there are no objections, I can put them in the list for feedback and we can discuss them next week as well. 15:18:43 +1 from me 15:19:09 TheJulia: I'm going to delete 'Zuul v3 playbook refactoring' (around L295). I think it is captured under the python3 work. OK? 15:20:10 rloo: wfm 15:20:35 done 15:20:57 Okay, I'm good to proceed to priorities for the upcoming week. How about everyone else? 15:21:18 ++proceed 15:21:58 #topic Priorities for the coming week 15:22:24 I need to do some quick clean-up of merged items on the list, and I think we have two additions based on the discussion so far. Is there anything else people feel should be on the list for this week? 15:23:54 TheJulia: could we add the deploy template spec under 'specs needed for this cycle' ? 15:24:43 rloo: it is already there 15:25:05 TheJulia: i only see clean step retrieval 15:25:17 TheJulia: oh, there are two specs for the cycle. 15:25:29 yup, let me fix 15:25:31 one is 'needed' the other isn't... 15:25:32 :) 15:26:17 Well, they are both kind of needed in the grand scheme of the universe 15:26:35 just different points along the continuum of interaction 15:27:13 I'm good with the priorities, I'm just curious if anyone has anything they wish to add. 15:27:55 for cisco, it sez 'patches in works...'. so no actual patches they want reviewed just yet? 15:28:10 bump: review https://review.openstack.org/#/c/602132/ for idrac vendor priority 15:28:11 patch 602132 - ironic-specs - Add firmware update spec for Dell EMC hardware - 8 patch sets 15:28:22 I can hunt the patches down, they are somewhere in our list 15:28:54 s/list/ironic review/ 15:29:01 thx. otherwise, I'm good with the list of stuff! 15:29:11 cdearborn: already in the list under vendor stuffs. 15:29:17 vendor driver stuffs at least 15:29:40 Okay then, if there are no objections we can go to open discussion and go from there. 15:29:48 TheJulia: y - not getting much attention, hence the bump 15:30:53 cdearborn: what does 'bump' mean? that you're mentioning here? :) 15:30:56 cdearborn: I can review again this week, I'll ask around :) 15:31:29 rloo: just a request for more attention 15:31:53 cdearborn: gotcha :) 15:32:28 Moving to Open Discussion 15:32:30 #topic Open Discussion 15:32:55 TheJulia: wrt the job migration, do you know if there is any project has done? 15:33:09 i saw this one https://review.openstack.org/#/c/588496/, but it seems doesn't work from the comment 15:33:10 patch 588496 - ironic - Switch ironic-grenade-dsvm too zuulv3 - 3 patch sets 15:33:29 I has one discussion topic, and it was largely an observation... that many of our gate failures are related to the rescue test failing due to port plugging. 15:33:53 kaifeng: Any project like in the openstack community, or do you mean CI job that has been migrated? 15:34:31 Grenade is a special case because there is a lot of scripting that has to be invoked a certain way at this time. So migrating it doesn't seem entirely feasible. 15:34:34 yeah, because it seems no project under ironic has completed grenade 15:35:38 emm, if that so, we just hold that until there is a solution 15:39:46 TheJulia: wrt rescue/port plugging, any idea why there are issues? is it something random that we have no control over? 15:39:48 Yeah, I suspect it might be good to sync up with derekh about that and maybe try and start a larger discussion on the mailing list regarding grenade 15:40:06 rloo: it comes back to the lack of callbacks and timing I think. 15:40:43 TheJulia: do we wait? Maybe increase the wait? 15:41:18 rloo: we do wait, 15 seconds hard coded in the conductor and I think the job retries for like 5 minutes 15:41:39 and it takes longer than 5 minutes for the port plugging? 15:42:03 or err, 5 minutes in rescue_wait or whatever state? 15:42:05 the job times out, there does seem to be more than one thing going on, I can sometimes see the port plug fail 15:43:11 Maybe if we could keep an eye out and start making a list and possibly trying to corrilate the failures... we might be able to target the general area to work on 15:43:39 it "feels" super sporadic to me... Just enough to get on my radar with rechecks. 15:45:37 TheJulia: any way to use that elastic recheck or whatever thing, to track this? 15:46:11 General timeout in remote jobs and I'm fairly sure the logs we would need to peer into are not indexed 15:47:27 s/remote/rescue/ 15:47:53 I'll start keeping a list and maybe we'll have enough data in a couple weeks to begin to see a trend 15:48:11 ok. not sure who else has been looking at ironic ci failures. 15:48:11 Does anyone have anything else to discuss? 15:48:24 Merged openstack/ironic master: Migration step to update objects to latest version https://review.openstack.org/613181 15:48:36 Ohh, well that should be in the priority list for stable cores 15:49:00 gah, dtantsur|brb has WF-1 on them 15:49:01 wrt rescue -- maybe put it under the Gate Status stuff... 15:49:04 in our whiteboard 15:49:07 rloo: will do now 15:49:11 thx TheJulia 15:50:31 done 15:50:37 Do we have anything else to discuss? 15:51:29 crickets 15:51:52 Well then, Thanks everyone! 15:52:11 tks =D 15:53:27 Have a wonderful week! 15:53:35 #endmeeting