14:00:31 #startmeeting nova 14:00:36 Meeting started Thu May 14 14:00:31 2015 UTC and is due to finish in 60 minutes. The chair is johnthetubaguy. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:37 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:39 The meeting name has been set to 'nova' 14:00:45 o/ 14:00:53 o/ 14:00:56 o/ 14:00:59 good to you all 14:01:02 o/ 14:01:04 o/ 14:01:16 o/ 14:01:17 o/ 14:01:22 #topic liberty release status 14:01:27 Hello 14:01:39 so we are pushing forward with the spec reviews 14:01:47 hoping to break the back of those before the summit 14:02:00 so those with real issues merging can get a slot during the unconference 14:02:19 anyways, we have a bp to talk about 14:02:22 #link https://blueprints.launchpad.net/nova/+spec/memory-bw 14:02:34 o/ 14:02:40 its adding an extra metric to the existing metric system 14:03:02 it doesn't feel like it needs a spec, in the same way as a scheduler weigher / filter 14:03:12 …but do we want more of this kind of thing in tree? 14:03:14 thoughts? 14:03:20 is this a static value? 14:03:32 * jaypipes looks... 14:03:36 dansmith: I get the idea they poll for it 14:03:39 this is utilization I guess 14:03:41 yeah 14:04:05 is it point in time or averaged? 14:04:13 if it's the former, it's completely useless to us, no? 14:04:51 if the HW provides the info the filter can return either point in time or average so I hope the BP is usgin avg 14:05:07 dansmith, Hi, it's averaged 14:05:10 okay 14:05:12 * mriedem joins late 14:05:34 sounds like any more details are best reviewed in code? are we OK with the concept? 14:05:38 sudipto: this would be adding another value to the compute_node['stats'] thing, yes? 14:05:43 * beagles joins late 14:05:50 sudipto: within the RT... 14:05:54 johnthetubaguy: well, are the units in percent or bps or what? 14:06:03 johnthetubaguy: would want to make sure that this is generic enough 14:06:08 jaypipes, no, this is not tracked by RT... 14:06:23 sudipto, really, why is that? 14:06:23 jaypipes, in the metrics column like the cpu one (as a monitor) 14:06:33 sudipto: ? in the spec it says: "Extend the resource tracker framework to implement a optional 65 14:06:33 monitor for Memory B/W utilization, much in line with the CPU monitor." 14:07:02 so question, are there enough questions that we move this to a spec review? 14:07:05 johnthetubaguy: I think maybe a spec would be good :) 14:07:11 +1 14:07:14 johnthetubaguy: yes. 14:07:16 +1 14:07:22 johnthetubaguy: especially since there's already a spec :) 14:07:26 yeah 14:07:34 #link https://review.openstack.org/#/c/180983/2/specs/liberty/approved/mem-bw.rst 14:07:45 jaypipes, sorry my bad, guess i need to correct it, initially i thought it could be done via RT, but it sounded like it's best fit just as a monitor 14:07:46 cools 14:07:57 lets take this debate to the spec review 14:08:02 thx 14:08:26 #info https://blueprints.launchpad.net/nova/+spec/memory-bw will go to a spec review, too much to talk about for the meeting 14:08:42 #topic Liberty Design Summit 14:09:09 #info Midcycle is: Tuesday 21st July till Thursday 23rd July in Rochester, MN (at IBM) 14:09:16 so the first thing is kinda unrelated 14:09:24 I think the midcylce is official now 14:09:42 the schedule for the design summit has now been "soft" frozen 14:09:45 #link http://libertydesignsummit.sched.org/type/design+summit/Nova 14:10:15 if there are issues with whats written for your session, let me know and I can try get last moment changes 14:10:21 but it would be good to avoid those 14:10:28 #link https://wiki.openstack.org/wiki/Design_Summit/Liberty/Etherpads#Nova 14:10:39 I created an etherpad with a rough agenda for every session 14:10:57 … that took way too long, so the ones at the end were quite slap dash 14:11:09 lol 14:11:12 but anyways, hopefully that helps, and I think we have leaders confirmed for most sessions now 14:11:35 any questions on the summit? 14:12:05 http://libertydesignsummit.sched.org/event/e57451a7a7d69f9e2626fa92862b119e#.VVStNvlVhBc 14:12:05 ha 14:12:08 40 minutes? 14:12:18 3 terribad topics 14:12:22 mriedem: I think they all it time boxing 14:12:23 for the price of 1 14:12:39 * mriedem shuts up 14:12:42 :) 14:12:49 yeah, that's going to be awesome 14:13:04 "it's a trap!" 14:13:12 yeah, if someone has a concrete plan for the resource tracker to present, that would be cool… 14:13:14 * mriedem takes note to tell jogo to be bad cop in that session 14:13:49 cool, so hopefully see lots of you there 14:14:01 #topic Czar and CPL Refresh 14:14:11 so this is a repeat from last week 14:14:25 mostly to say, feel free to catch up with me if you want to help out with things 14:14:34 a few folks have reached out, and thank you for that! 14:14:48 * jaypipes notes that Robert Graham Clark's profile pic has him with some sort of M-16 rifle... 14:14:56 #link https://wiki.openstack.org/wiki/Nova/BugTriage 14:15:10 there are also bug tag owners if you fancy something around stability ^ 14:15:29 #link https://wiki.openstack.org/wiki/Nova#People 14:15:47 anyways advert break over… 14:15:52 #topic Bugs 14:16:05 #link https://bugs.launchpad.net/nova/+bug/1450291 14:16:05 Launchpad bug 1450291 in OpenStack Compute (nova) "nova-manage db sync not committing to database" [High,Incomplete] 14:16:15 oh, that was me who added that 14:16:19 so... I kind of feel like on the bug front, until we get a burndown dashboard that includes "by tag" it's hard to motivate. Burn down dashboards are nice from sense of accomplishment though. 14:16:30 I guess we fixed that 14:16:42 sdague: thats a good idea 14:17:00 #help we could do with per bug tag burn down charts 14:17:08 From an Ironic related perspective. mrda and I have setup this. https://wiki.openstack.org/wiki/Nova-Ironic and https://wiki.openstack.org/wiki/Nova-Ironic-Bugs 14:17:18 dansmith: cherry pick for https://review.openstack.org/#/c/182787/ ? 14:18:24 jlvillal: so someone has to manually update that daily? 14:18:30 johnthetubaguy: that bug is mis-titled 14:18:46 jlvillal: one request, can you maybe recommend what is ready for review on here? https://etherpad.openstack.org/p/liberty-nova-priorities-tracking 14:18:54 mriedem, We were going to do weekly bug scrubs. mrda and I. Open to better suggestions 14:19:04 johnthetubaguy, Will do 14:19:42 jlvillal: it sounds like you are a bug tag owner, could you update yourselves here please? https://wiki.openstack.org/wiki/Nova/BugTriage 14:19:43 whatever happened to tjones' page? 14:19:56 is that in infra now? 14:19:56 dims has a version of it up somewhere I think 14:20:01 johnthetubaguy, Will do 14:20:03 mriedem: I couldn't find the link actually, does anyone have that? 14:20:15 sdague: is it useful? i'd have to think that's better than a wiki 14:20:32 #action johnthetubaguy to talk to dims about getting the bug web page link back up and running 14:20:33 so, non of these tools really did the by-tag pivot 14:20:43 so it's just the whole pile 14:20:50 sdague: right, I like the idea of looking at the smaller pile, that makes sense 14:20:55 the whole pile is usually too big to tackle 14:21:20 we could add a tag filter to tjones' page though probably, right? 14:21:25 if the LP API provides tag filtering 14:21:48 dansmith: can we update the title, did we work out what it was? 14:22:13 johnthetubaguy: it's definitely wrong either way because we know it's committing the changes 14:22:22 johnthetubaguy: something about sqlite 14:22:23 johnthetubaguy: and we know it's just a warning, and we know it's oslo and config related :) 14:22:25 dansmith: agreed 14:22:26 mriedem: there are a lot of iffs there 14:22:56 mriedem: anyway, can take it offline 14:23:03 and we know it seems to only be with ubuntu package-based upgrades ;) 14:23:15 https://bugs.launchpad.net/nova/+bug/1450291 14:23:15 Launchpad bug 1450291 in OpenStack Compute (nova) "nova-manage db sync seems like its pointing at the wrong database for some people" [High,Incomplete] 14:23:19 so I changed it 14:23:32 yeah, any other bugs people want to raise 14:23:42 sdague: did we sort out that python-novaclient patch? 14:23:58 it's here https://review.openstack.org/#/c/182966/ 14:24:19 so there is current debate on where the config file should exist for people to find it 14:24:24 which is the -1s 14:25:15 so *shrug* 14:25:41 sdague: add it to a top level function test config folder maybe? 14:25:46 anyways, more debate on the review 14:25:49 lets keep moving 14:26:03 johnthetubaguy: yeh, that's what jaypipes wants, I would argue that if you can't run a tox command without doing a thing 14:26:09 that thing should be in the top folder 14:26:15 ls 14:26:20 like I said, I'm not strong-willed about it... 14:26:33 jaypipes: you are -1, that's some will :) 14:26:33 sdague: gotcha, I am certainly used to running tox at the top 14:26:52 #topic Review stats 14:26:52 sdague: it's OK to disagree and still be friends :) 14:27:07 jaypipes: I disagree 14:27:11 I vote I drop this topic when there are no items, I forgot to do that 14:27:18 I knew dansmith would say that. 14:27:21 * dansmith stopped reading at "it's okay to disagree" 14:27:23 #topic Open discussion 14:28:06 so we agreed to not dis-agreed about agreeing right? 14:28:21 OK, so I don't have anything sensible on the agenda 14:28:31 we had a good ML debate on images in specs 14:28:43 I think we agreed to try non-ascii art things that are not bitmaps 14:28:54 where that makes sense 14:29:10 any comments? or other topics from folks? 14:29:30 … oh, clearly we are skipping next weeks meeting 14:29:52 do we want the next meeting at this time, to keep the pattern the same, so May 28th 2015 1400 UTC 14:29:54 ? 14:30:08 yes 14:30:13 #info next meeting May 28th 2015 1400 UTC 14:30:17 some of us use calendar reminders 14:30:19 keep the pattern, otherwise, things get confusing 14:30:21 agreed 14:30:22 mriedem: ++ 14:30:30 yes 14:30:38 mriedem: exactly my thinking 14:31:08 cool, save journey for all those folks who have a journey to the summit 14:31:18 s/safe/save/ 14:31:20 oops 14:31:33 thanks all 14:31:39 #endmeeting