21:00:54 #startmeeting networking 21:00:54 Meeting started Mon Apr 18 21:00:54 2016 UTC and is due to finish in 60 minutes. The chair is armax. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:54 armax: Just tired. So very, very tired. 21:00:54 armax: are you trying to invite everyone for a coffee in Austin? 21:00:55 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:56 hello 21:00:57 The meeting name has been set to 'networking' 21:01:00 hey mlavalle! 21:01:11 hi 21:01:15 hello 21:01:15 hi 21:01:15 o/ 21:01:16 thanks everybody for joining this meeting 21:01:17 hello 21:01:21 o/ 21:01:23 armax: yw 21:01:25 :) 21:01:28 y helo 21:01:28 o/ 21:01:30 this is the week ahead of the summit and I know that many of view are superbusy 21:01:31 It's getting warm here, so inverse hibernation 21:01:34 o/ 21:01:46 johnsom: I am having this meeting with my speedo on 21:01:51 * armax tmi 21:01:58 eh 21:01:59 Things you can't unsee 21:02:01 * Sam-I-Am leaves channel 21:02:11 is this a dev brothel or a meeting? 21:02:14 that’s my strategy to keep the meeting short! 21:02:23 aloha 21:02:27 I guess it’s working 21:02:29 anyhoo 21:02:46 #link https://wiki.openstack.org/wiki/Network/Meetings 21:02:57 #topic Announcements 21:03:04 armax: TMI! 21:03:12 some of you must be aware already 21:03:21 the schedule for the neutron sessions is available 21:03:24 #link v 21:03:27 #undo 21:03:29 Removing item from minutes: 21:03:42 for those you may wonder, yes, I haven’t got a new keyboard yet 21:03:45 #link https://www.openstack.org/summit/austin-2016/summit-schedule/global-search?t=Neutron%3A 21:04:25 so please do make sure you spot glaring errors about conflicts etc 21:04:53 if you are a session chair or not, come and talk to me if you want to get any update reflected 21:05:02 Wed and Thu in the same spot. neat 21:05:17 who is providing the beverages? 21:05:24 There are more sessions that are related to Neutron 21:05:30 that link doesn’t show them all 21:05:49 for instance, on Tuesday, there are a bunch of cross project ones 21:05:57 like the one Sukhdev promoted 21:06:17 and for instance the one we’re having in the Nova’s time on Wed morning 21:06:30 yes - please look for Ironic-neutron session at 4:40pm on Tuesday 21:06:35 so be on the lookout 21:07:13 #info searching with the schedule app: http://lists.openstack.org/pipermail/openstack-dev/2016-April/091186.html 21:07:15 armax: the nova-neutron joint meeting is Wednesday 11:00 - 11:40, Hilton hotel 21:07:32 anteaya, mlavalle thanks 21:07:37 the magicial colon is important when searching with the schedule app 21:07:48 armax : Can I take a moment to make the annocement about the Ironic-neutron session? 21:07:59 Sukhdev: didn’t you take it already? 21:08:11 ha ha - 21:08:16 Sukhdev: I can’t stop you, now can I? 21:08:25 wanted to share an etherpad 21:08:30 #link: https://etherpad.openstack.org/p/newton-baremetal-networking 21:08:36 Sukhdev: you got my permission 21:08:48 We need some critical mass from neturon side to attend this session 21:09:18 such as armax kevinbenton akihiro rossela 21:09:35 Sukhdev: are you saying that I am fat? 21:09:44 Sukhdev: link to that session ? 21:10:02 #link: https://www.openstack.org/summit/austin-2016/summit-schedule/#day=2016-04-26&summit_types=2&tags=3601 21:10:07 look for 4:40pm session 21:10:30 #link https://www.openstack.org/summit/austin-2016/summit-schedule/events/9491?goback=1 21:10:34 jckasper: ^ 21:10:35 Ideally, we want an API to support Multiple VLAN netoworks on a single port 21:10:40 thanks 21:10:53 Sukhdev: you have it already 21:11:03 kind of like Vlan aware VMs - but, we would like a general purpose API from neutron 21:11:18 so that it can be used for BM as well 21:11:38 Anybody who has interest in this topic, please join that session 21:11:48 I would add, come prepared 21:12:06 yes thanks armax 21:12:18 I am done thanks armax 21:12:26 there’s substantial work done in this arew with http://specs.openstack.org/openstack/neutron-specs/specs/mitaka/vlan-aware-vms.html 21:12:37 if there are gaps, these should be documented somewhere 21:12:51 there’s a review out: https://review.openstack.org/#/c/277853/ 21:13:36 and that somewhere should be neutron’s RFEs 21:13:38 Yes, that is why we want all the relevant players in one place so that we can identify any overlaps and conflicts and come up with a unified solution/API 21:14:20 Sukhdev: let me absolutely clear about this 21:15:13 I would rather give http://specs.openstack.org/openstack/neutron-specs/specs/mitaka/vlan-aware-vms.html time to mature in the context of virtual workloads 21:15:35 this means that this should be our objective for Newton 21:15:50 anyhow, we’ve taken more than a moment, let’s move on 21:16:11 last of the reminders 21:16:31 we have an entire day to have more focussed discussions on Friday 21:16:35 the etherpad is 21:16:38 #link https://etherpad.openstack.org/p/newton-neutron-unplugged-track 21:17:25 I appreciate that most people end up flying out sometime early in the last day of the summit, but if you are indeed around 21:17:37 can i bring my acoustic guitar? 21:19:22 Sam-I-Am: I suppose the silence means, yes you can 21:19:38 Ok moving on, unless someone else has an announcement to make 21:19:42 cool. we can sing together. 21:19:54 :) 21:20:26 What is the format of the sessions? 21:20:37 On Wed and Thursday 21:20:40 Big room? 21:20:53 last time the room was not nearly big enough. 21:20:56 kevinbenton: big rooms 21:21:23 neutron sessions make me feel like a sardine 21:21:29 and most of those people dont provide useful input 21:21:34 I hope it's one of those "everything is bigger in Texas" moments 21:21:38 they just exist to complain 21:21:40 i meant the contributor meetup. 21:21:43 so maybe we need our secret meeting :) 21:21:43 Sam-I-Am: things are getting better summit after summit 21:21:57 *aaS was always meeting somewhere else 21:21:58 i know of some back alleys in austin 21:21:58 the weather is bigger in texas atm 21:22:15 #topic Blueprints 21:22:22 Sam-I-Am is that a good thing or bad :-) 21:22:30 Sam-I-Am: usually everyone gets into this community with a complaint ;) 21:22:38 as long as they server beer in those alleys 21:22:52 that's a way of saying "hello I am here to help, because everything is a mess" 21:22:54 last two meetings ihrachys and I went through the current workload to see if assignee/approver were still valid 21:23:18 we went through the entire list 21:23:26 xgerman: its austin, so yeah... except you'll have to wait 2 hours in line 21:24:09 a few others got kicked out relatively early from mitaka and appear as deferred in this list 21:24:10 #link https://blueprints.launchpad.net/neutron 21:24:42 we’re figuring out what to do with those, but if you do know something, please reach out 21:25:14 one last thing re: blueprints 21:25:41 I have seen people going on https://launchpad.net/neutron and registering blueprints ignoring the RFE process altogether 21:25:48 I mark those BP’s as obsolete 21:26:11 until a while ago I was giving an explaination, but I have lost the will to scold and I simply mark them obsolete 21:26:42 armax: Let your iron fist rain down fire on those who ignore the process! :) 21:26:57 how dare they 21:27:03 What should I do if I see a bug that says '[RFE]' in the title but is not tagged with the 'rfe' tag? Refer them to the drivers team? 21:27:05 Honestly, the fact that people can even do that with Launchpad means that Launchpad is not the right tool 21:27:06 the https://launchpad.net/neutron landing page has clearly stated that for submitting blueprints/RFE and defect reports instructions are available at: http://docs.openstack.org/developer/neutron/policies/blueprints.html 21:27:33 njohnston: when in doubt add the tag, and we’ll clean it up 21:27:41 armax: Thanks, will do 21:27:55 mestery: we’re working within the constraints of the tool we have 21:28:10 armax: Yup, I'm just pulling a salv-orlando and providing insight from the sidelines 21:28:11 mestery: give me a level and I will lift the world said a genius 21:28:21 mestery: but I am no genius so I don’t know what else to do 21:28:26 besides as the latin said "repetita iuvant" 21:28:59 salv-orlando: lol 21:29:04 +1 for lingua latina 21:29:04 armax: he said lever 21:29:06 armax: wasn't it a lever? 21:29:17 whats that ? 21:29:20 I'm all for working it over again with level 21:29:27 according to armax it was a level 21:29:35 much better 21:29:46 #link http://www.brainyquote.com/quotes/quotes/a/archimedes101761.html 21:29:50 no it’s lever 21:30:00 No one is questioning the lever armax :) 21:30:01 #link http://dictionary.cambridge.org/dictionary/english/lever 21:30:02 I think yours is better 21:30:17 not sure I understand the source of confusion 21:30:23 ha ha ha 21:30:24 the quote or the object? 21:30:30 give me a level and I will beat the monster at the end, said the genius of arcade videogames 21:30:39 :D 21:30:43 I've always heard it as 'give me a fulcrum', but I think that is just because fulcrum sounds cooler. 21:30:46 :-) 21:30:53 anyhoo we’re clearly digressing 21:31:00 * mestery can tell we're all ready for summit 21:31:00 njohnston: fulcrum inplies lever 21:31:11 this is how you bike shed even regular words 21:31:16 njohnston: it isn't a fulcrum without a lever 21:31:16 good jorb, neutron 21:31:35 njohnston: yeah like the MIG-29 fulctrum (ok I will stop it now) 21:31:35 Looks like this crowd is ready for a beer 21:31:36 I appreciate people’s effort to keep the conversation alive, btw 21:31:45 the shed is a rainbow 21:31:48 it means people are reading! 21:31:54 #topic Bugs 21:31:55 the shed is black from all the colors. 21:31:56 dougwig: with many levels 21:31:58 armax: Your words are like poetry 21:31:58 paint it black 21:32:12 mestery: you’re welcome 21:32:16 Sam-I-Am: ha ha ha 21:32:27 * anteaya changes the playlist 21:32:46 ok, I think last week deputy was reedip 21:32:49 this week is blogan 21:32:56 armax: nope last week was me 21:33:10 sorry 21:33:12 you’re right 21:33:29 sounds like blogan is doing it again! 21:33:32 I misread the last entry in the table 21:33:40 sounds like dougwig just volunteered himself! 21:33:44 so we need a volunteer for this week 21:33:47 and possibly next week 21:33:59 any taker? 21:34:08 Note for any potential takes: These are the light weeks 21:34:11 i'll do this week. 21:34:15 dougwig: lol 21:34:15 :) 21:34:17 i could have sworn someone volunteered last week for this week 21:34:18 the job comes with a signed picture from me 21:34:21 as a gift 21:34:48 I remember someone taking this week as well 21:34:57 #action dougwig deputy for the week of Apr 18 21:35:00 * ihrachys already have the photo to pick up 21:35:13 no more amazon gift cards? 21:35:13 johnsom: both haleyb and john-davidge offered 21:35:37 ihrachys: since you were one of the earliest you also got a frame with it 21:35:51 dougwig: those if for code contributions 21:35:58 s/if/are 21:36:01 armax: sigh. 21:36:06 armax: do you think i could get a strand of your hair? 21:36:09 "the job comes with a personally signed email containing my LinkedIn profile, with the heartfelt sentiments 'I'd like to add you to my professional network' engraved." 21:36:19 blogan: what hair? like I had anything left 21:36:36 who needs hair -- waste of time! 21:36:38 armax: exactly, its even more valuable 21:36:40 Creepiest neutron meeting ever.... 21:36:50 blogan: ok, let me grow some I just shaved last week 21:36:53 mestery: tmi? 21:37:01 lol 21:37:26 ok, before moving on 21:37:31 let’s look at our dear grafana 21:37:47 which reminds me that last week we had a bump in our api job strategy 21:38:16 we put it back to non-voting temporarily to switch to using tempest-plugin 21:38:23 #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate 21:39:02 I saw an interesting failure. Default security group creation hit a db deadlock. Then the retry mechanism kicked in as expected 21:39:13 kevinbenton: is there a bug report? 21:39:24 And every attempt after failed with a "no such save point error" 21:39:27 kevinbenton: did that show up in the gate or check queue? 21:39:27 Not yet 21:39:37 Showed up in a rally run 21:39:57 kevinbenton: ack 21:40:00 we should also monitor 21:40:02 #link http://status.openstack.org//elastic-recheck/ 21:40:13 and 21:40:14 #link http://status.openstack.org//elastic-recheck/data/uncategorized.html 21:40:27 in particular to see if there are signatures for new failure modes 21:40:48 * njohnston bookmarks 21:41:07 armax: thank you 21:41:41 summit would be a great place for those folks looking to learn how to catagorize those uncatagorized failures 21:41:42 for some reason, something is unsettled with fullstack 21:41:58 http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=5&fullscreen 21:42:01 #link v 21:42:04 #undo 21:42:05 Removing item from minutes: 21:42:07 #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=5&fullscreen 21:42:31 then there’s this very worrying upward trend with grenade multi 21:42:34 #link http://grafana.openstack.org/dashboard/db/neutron-failure-rate?panelId=6&fullscreen 21:42:41 has anyone looked into it? 21:43:09 Is that check or gate? 21:43:12 multi with dvr 21:43:15 that’s check 21:43:24 armax: Looks like fullstack jinked upwards yesterday, in a couple of increments over the day: http://grafana.openstack.org/dashboard/db/neutron-failure-rate?from=1460865600000&to=1460951999999 21:43:37 I did not know about grenade one. I will take a look tomorrow. 21:45:02 ihrachys: to be fair 21:45:04 ihrachys: https://review.openstack.org/#/c/299844/ 21:45:15 this got merged last nite 21:45:25 so it may be a simple reflection of more data points 21:45:58 ack. if no one gets to it till my morning, I will take a look. 21:46:16 ... and if you do, please drop something my way so that I am aware. 21:46:34 ihrachys: no sorry 21:46:38 I am getting confused 21:46:55 ihrachys: I take back what I said, the grenade multi that’s spiking up is not multi 21:48:35 armax: how come 21:49:02 ihrachys: I mean is not dvr, 21:49:24 * armax is having a hard time tracking all the multiple ice cream flavors we have 21:49:53 ok. what I see on the graph is: grenade multinode is below usual multinode jobs. so it probably just reflects some issues with the latter. 21:50:13 since it runs less tests, it fails less frequently than -full 21:51:26 ihrachys: let’s take this offline 21:51:38 but we definitely want to understand what’s going on before we go ahead with 21:51:40 #link https://review.openstack.org/#/c/299843/ 21:51:49 sc68cal: ^ 21:52:18 ok, moving on 21:52:29 #topic Docs 21:52:39 moo. 21:52:39 Sam-I-Am: do you have anything for us? 21:53:03 I have a couple of docs patches I am waiting to review 21:53:18 nothing in particular. i've been busy working on a summit pres. trying to work through updating the scenarios for mitaka, but its been hard to find time. 21:53:25 nudge to carl_baldwin, tidwellr 21:53:39 i dropped working on the install guide for N, so hopefully i'll have more time to focus on the networking guide (including reviews) 21:53:55 armax: ack. :) 21:53:59 i've seen a few reviews. just cant focus on them right now. 21:54:07 Sam-I-Am: do you have an update on the api docs progress? 21:54:10 and apparently we dont have a lot of other people reviewing 21:54:17 no, i dont. that'd be an anne gentle thing. 21:55:09 this was the last I saw on that 21:55:12 #link http://lists.openstack.org/pipermail/openstack-dev/2016-April/091157.html 21:55:43 thats about where i'm at now 21:55:49 Sam-I-Am: ack 21:55:51 I had an interesting conversation with agentle earlier this week. I noticed that while create network/update network API calls get extra QoS-related arguments when the QoS plugin is installed, those are not reflected in the API guide on the page that talks about those API calls. 21:55:51 Sam-I-Am: thanks 21:56:04 i dont follow api much, just too much to track 21:56:27 we probably should have someone here who tracks api docs, patches, etc. 21:56:34 I asked if the extra arguments should be grouped with the API calls or with QoS, and she said to add an additional bit plus example under the generic case for the API calls. Just FYI for all. 21:56:34 one person can't do all the docs stuff 21:56:39 sorry meant to link http://lists.openstack.org/pipermail/openstack-dev/2016-March/090659.html 21:56:43 #link http://lists.openstack.org/pipermail/openstack-dev/2016-March/090659.html 21:57:46 Sam-I-Am: I can’t say I disagree 21:58:10 ok, 2 mins left 21:58:36 before we leave I wanted to share this link with you folks 21:58:40 #link https://www.youtube.com/watch?v=M4cDyM2s2bc&feature=youtu.be 21:58:50 OpenStack Design Summit - Tips for session moderatorsOpen 21:59:17 whether or not you’re a session moderator at this summit, it’s good to keep those notes in mind 21:59:27 arrive drunk 21:59:31 thats it 21:59:32 thanks armax 21:59:34 and help for a smooth summit 21:59:42 thanks armax 21:59:44 bye folks 21:59:46 #endmeeting