13:00:17 #startmeeting Docteammeeting 13:00:17 Meeting started Tue Oct 15 13:00:17 2013 UTC and is due to finish in 60 minutes. The chair is annegentle. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:18 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:21 The meeting name has been set to 'docteammeeting' 13:00:21 Agenda is here 13:00:23 #link https://wiki.openstack.org/wiki/Meetings/DocTeamMeeting#Agenda_for_next_meeting 13:00:37 * ttx lurks 13:00:45 I know we're all itching to talk install but let's circle back on the action items from last time 13:01:15 heh first one was "Merge in install guide as-is" 13:01:30 done :D 13:01:31 all the action items might be install related anyway 13:01:38 annegentle to work on glance conceptual intro, and swift 13:01:38 install" 13:01:43 also done 13:01:59 sgordon to share patterns for nova <--- for release notes? 13:02:14 i wrote the vast majority of the nova release notes 13:02:20 nice work sgordon 13:02:22 sgordon: nice. 13:02:34 #link https://wiki.openstack.org/wiki/ReleaseNotes/Havana#OpenStack_Compute_.28Nova.29 13:02:40 annegentle to dive into glance new features for release notes 13:02:46 oh I haven't started on this ^^ 13:02:56 then there's NickChase deep dive into Neutron new features 13:02:58 SMEs for vmware, powervm, vmwareapi and xen wrote those specific bits 13:03:06 I don't think Nick's online 13:03:08 ? annegentle, image notes are done 13:03:17 summerl: awesome 13:03:35 guess that's why I took that one, ha ha 13:03:45 nermina to study ceilometer features for rel notes <--- how's that one? 13:03:55 annegentle, I volunteered at the last bit 13:04:01 sgordon: that's a thing of beauty 13:04:23 percale did the review 13:04:31 koolhead17 to ask Swift PTL notmyname for feature set for release notes 13:04:34 and then last was ^^ 13:04:46 summerl: I don't see you in the action items from the meeting minutes but that's ok 13:05:04 summerl: oh did you do the image notes? 13:05:06 annegentle: yes will get it 13:05:07 oh well, annegentle, did it anyway :) 13:05:15 summerl: thank you! 13:05:19 annegentle: notmyname is in Austin only :) 13:05:46 koolhead17: yeah I saw that, but you did ask him right? So you can call the action item done 13:06:05 annegentle: sure 13:06:16 Ok I think we call all those action items a wrap 13:06:21 On to the favorite topic 13:06:27 #topic Install guides 13:06:36 #link https://etherpad.openstack.org/p/havanainstall 13:06:49 and 13:06:51 #link https://wiki.openstack.org/wiki/HavanaDocTesting 13:06:58 fifieldt: you want to walk us through latest? 13:07:17 ok, sure 13:07:26 so in recent days we've added quite a few new sections to the install guide 13:07:57 it seems like: overview, basic OS, identity, image, dashboard are done and known good 13:08:03 object storage is likely good to go as well 13:08:24 we also have added sections on networking for compute, neutron, heat and ceilometer that might be passable 13:08:29 block storage too 13:08:46 but the testing matrix (the wiki link) shows the current working status 13:08:56 basically if it doesn't have a green tick, we don't know if it works 13:09:06 if it's a warning sign, it means we haven't written it yet 13:09:15 as far as I can tell we still do not have a working compute install 13:09:18 does anyone know if others are successfully installing havana compute? 13:09:50 hi everyone 13:09:54 hi nermina 13:09:56 I'm still working through the nova networking stuff, but otherwise yes 13:09:58 annegentle: we have 100 options are we talking about compute with KVM 13:10:33 the nova conroller stuff appears to work 13:10:40 also am going through the swift/ubuntu install guide will comment/finish reviewing soon 13:11:04 shaunm_: ok so Fedora is done for Compute? 13:11:06 which is enough to test the dashboard .... but without compute nodes with networking, it's not so useful for other things :) 13:11:13 it's ubuntu we can't quite get yet? 13:11:38 fifieldt: you have to change 100 options from the default? That seems bad 13:11:50 that's not true, annegentle .... 13:11:59 fifieldt: ok 13:12:27 is a VM a good-enough install test? 13:12:47 it's how I "grew up" on openstack, so I'd say yes :) 13:13:04 running tempest is the best option 13:13:05 I don't believe networking is working under any distribution right now 13:13:06 fifieldt: ok good 13:13:11 annegentle: excepting networking, yes. I wanted to briefly look back over the sections leading up to compute before putting the check mark on the wiki, because I know there were changes 13:13:12 EmilienM: oh, how do you mean? 13:13:38 shaunm_: ok got it 13:13:43 annegentle: we are deploying havana, against several os, to test if all was right, we run tempest 13:14:01 not useful in this case EmilienM - we're mainly testing instructions 13:14:03 shaunm_, but network is the hardest part of nova :) 13:14:09 that ensures you functionnality coverage 13:14:09 EmilienM: ah ok, but yeah, we need to manually install 13:14:16 fifieldt: make sense :) 13:14:33 fifieldt, and hence it's the thing most in need of good docs :) 13:14:46 my sense is that we're in a similar position as we were for Grizzly - networking is far, far behind in docs. Suggestions? 13:15:01 yes, but I don't like to hear things are "almost done" when they still need networking ... it's just not true :) 13:15:06 how to communicate this to the rest of OpenStack? And when? 13:15:13 fifieldt: oh I'm with ya :) 13:15:19 annegentle: we need to get some core devs help to fix it :) 13:15:55 fifieldt, did you just write that section from memory? you didn't run the commands? 13:16:14 I wrote the section derived from a random assortment of sources 13:16:18 koolhead17: well, we have some neutron folks talking about ml2 writing, but no patches yet 13:16:31 the commands will run, but they are likely to be unsuccesful in practice 13:16:37 it actually has the user configuring quite a bit more stuff than the grizzly install docs did 13:16:54 annegentle: we need to run after them :D 13:16:58 shaunm_: the networking stuff that got added last night has use cases 13:17:06 indeed shaunm_ but we're going for multi-host 13:17:11 annegentle, that's for neutron 13:17:12 koolhead17: well sure, but in 2 days I don't expect miracles... so what do we need to do 13:17:16 I think we're still talking about compute 13:17:21 compute networking* section 13:17:33 fifieldt: yeah are you talking nova-network or neutron? 13:17:37 nova-network 13:17:41 fifieldt: ok 13:18:20 fifieldt: via nova-network it should be easy 13:18:22 fifieldt, annegentle, has anyone looked at ch_installcompute? 13:18:27 as it's written, it's a bit more magic than I'd like. but first priority is for it to be functional 13:18:48 shaunm_: ah magic. sigh. 13:18:48 I just filled out the section with what I think has the best chance of working 13:19:01 nermina, yes, but as you saw, I fell. Hard. 13:19:08 koolhead17, if you have a working set of config for nova-network ... 13:19:16 yes I think the main goal is testing, and we need instructions to test 13:19:22 oh, i didn't see. sorry, summerl 13:19:31 summerl: ouch! 13:19:39 yes, it wasn't pretty. 13:19:48 that was the aim of the patch, to hopefully get some people debugging the stuff and make it work 13:19:57 sgordon, have you looked at the compute chapter? 13:20:00 ok, here's the question, who has time and access to systems to test? 13:20:02 the compute network patch that is 13:20:13 I know shaunm_ has fedora systems to test 13:20:30 I'm working through fifieldt's nova-network stuff now 13:20:31 I think fifieldt has ubuntu, and I can do ubuntu (but networking is NOT my strong suit) 13:20:36 shaunm_: ok good 13:20:42 i will ask our folks but don't want to promise 13:20:43 #action shaunm_ testing nova-network on Fedora 13:20:47 fifieldt: nopes 13:20:49 nermina: ok, fair enough 13:20:57 summerl, i've looked at it yes, but as far as networking for compute i havent really touched it 13:21:06 I think part of our communication back to the community is that the testing is very difficult and requires hardware 13:21:07 as 99% of what i've looked at with networking is neutron 13:21:13 tell me exactly what needs testing, annegentle, so i can be specific 13:21:42 specific testing instructions: follow the manuals for a given section, dumbly, to the letter, and see if it works as advertised 13:21:56 distro/section combos requiring this can be found at https://wiki.openstack.org/wiki/HavanaDocTesting 13:22:09 thanks, fifieldt 13:22:11 nermina: sure, looking at https://wiki.openstack.org/wiki/HavanaDocTesting, ask if they can test ubuntu or RHEL for Compute 13:22:20 i will follow up today 13:22:27 you can hold me to that 13:22:28 nermina: thank you 13:22:36 Ok, any other discussion on install? 13:22:52 on that list, CentOS is probably a low priority 13:23:18 as long as RHEL tests OK 13:23:22 shaunm_: yeah I'd agree 13:23:27 good to know 13:23:47 indeed two complete distros are better than many incomplete 13:23:48 I'll ask a ton of questions as I test 13:24:00 fifieldt: that would be fantastic 13:24:13 oh, I do want to say, I'm creating a branch tomorrow that is our release candidate 13:24:41 Ok, so no updates now, annegentle, in other guides? 13:24:47 so we'll need to do the whole backporting thing to update the install guide? 13:24:49 and I'll rebase with whatever gets in by Thursday morning my time 13:24:58 (really it's cut on ttx time) 13:25:16 actually, let's talk about fifieldt's question 13:25:30 I think we cut a branch and have to backport 13:25:35 I can't think of another way 13:25:56 my thinking was more along the lines of: with the current state of the install guide, is it worth cutting? 13:25:57 fifieldt: i just asked pranav/aptira for the networking help 13:26:14 koolhead17: thanks 13:26:26 fifieldt: that's certainly up for discussion 13:27:02 personally, I think we need to see how the compute stuff ends up 13:27:06 and what we need to communicate to the openstack community if its different from our planning all along 13:27:15 if we can't get a tested compute install going 13:27:18 backporting is not all that terrible 13:27:20 then a "come back in 2 days" message 13:27:22 are there people wanting to add stuff to master that isn't for havana? 13:27:31 shaunm_: nope 13:27:37 is better than releasing a guide that doesn't work 13:28:07 fifieldt: yeah and my sense is that the chances of the guide working in 24 hours isn't great 13:28:32 ok, so the "other way" is to not cut a release branch 13:28:39 we need to set some deadline though 13:28:48 we really can't let this slip 13:28:53 Yes, I was just going to say, I refuse to go to the summit again with non-released docs 13:28:58 if I have to release without install I will 13:29:04 fifieldt: right. 13:29:15 I don't think releasing without install is a good option either 13:29:19 fifieldt: ok 13:29:25 other ideas? 13:29:27 i will respond on compute within the next hour 13:29:33 cheers nermina! 13:29:42 nermina: ok thanks! 13:29:45 we need to call in all favours and just get it done 13:29:57 Ok, let's all report in about 12 hours on the mailing list 13:29:59 we can drop later chapters if they're not ready 13:30:12 We'll keep triaging this guy every 12 hours 13:30:28 And do a go-no-go vote via the mailing list 13:30:31 I think we should be making phone calls to people at this stage 13:30:31 sound ok? 13:30:35 rather than just emails 13:30:49 fifieldt: fair 13:30:56 if we don't have the skill to write the nova-network section 13:30:59 we need to find someone who does 13:31:05 and do that within the next 12 hours 13:31:13 get it tested 13:31:16 and then release 13:31:20 fifieldt: any ideas who? 13:31:44 all my guys are already asleep :) sans people like our ops guide co-authors 13:31:56 joe, jon etc 13:32:09 oh yeah Joe saved our butts last time :) 13:32:20 blast it on the mailing lists as well? 13:32:27 nermina: yep, that sounds right 13:32:28 these guys are pro at debugging systems - if we can get a testbed going based on our install that they can ssh into and just "fix" 13:32:28 dguitarbite1: is going to help us with the networking part annegentle 13:32:32 that might be a good way to do it 13:32:36 hi 13:32:44 #action AnneGentle to email the openstack list asking for help 13:32:55 might be good to send the structure 13:32:56 dguitarbite1, fix this: http://docs.openstack.org/trunk/install-guide/install/apt/content/nova-network.html so it works 13:32:57 :) 13:33:00 dguitarbite1: awesome, do you have what you need? 13:33:15 :) 13:33:22 fifieldt: kool 13:33:26 annegentle: yes 13:33:33 dguitarbite1: thanks much 13:33:43 i got quantum working with OVS 13:33:55 dguitarbite1: ok definitely update then! 13:33:57 even inside Virtual Box it runs perfect :) 13:34:16 dguitarbite1: ^^ 13:34:25 ok sorry gotta keep moving 13:34:27 #topic Doc tools updates 13:34:43 We do need to go to 1.11.1 for Havana for the API site for certain, for other docs they can stay at 1.11.0 13:34:54 #link https://github.com/rackerlabs/clouddocs-maven-plugin#release-notes 13:35:12 we needed this one: Updated apipage-main.xsl to include metering API 13:35:24 the log a bug link has been working well! 13:35:31 loves it! 13:35:41 indeed 13:35:56 we also need Rename "Template" parameters as "URI" parameters in output for API docs, hence the 1.11.1 release. 13:36:11 In case you're wondering how releases work, check out https://wiki.openstack.org/wiki/Documentation/Release#Actions_on_Final_Branch 13:36:16 those are all the actions on the final branch 13:36:37 one thing I'm not sure about, is whether I have to cut a second final branch to make sure the sitemap.xml is accurate. 13:36:49 since the branch has to be built on to get docs.openstack.org/havana 13:37:00 and without the /havana, the sitemap won't be accurate 13:37:28 I am also seeing reports of people not getting redirected as expected, most recent is from today https://bugs.launchpad.net/openstack-manuals/+bug/1240059 13:37:29 Launchpad bug 1240059 in openstack-manuals "Redirects aren't working right in grizzly docs to new documents" [High,Confirmed] 13:37:34 any questions on doc tools? 13:37:53 #topic Bug report, DocImpact state 13:38:11 Ok, as fifieldt noted on the list, DocImact wasn't working for about the month of September/half of October 13:38:21 apologies all :( 13:38:28 it got fixed by infra yesterday, and I wrangled the rest into the system 13:38:35 I think the net add was just a dozen or so, WHEW 13:38:58 #link https://bugs.launchpad.net/openstack-manuals/+milestone/havana 13:39:31 #link https://bugs.launchpad.net/openstack-api-site/+milestone/havana 13:39:58 I'm pretty excited about 82 confirmed and 146 Fix Released though 13:40:13 We can have a LOT of pride in this release overall. A LOT. 13:40:25 Even if install is kicking our asses. 13:40:26 :) 13:40:52 we'll get it done 13:40:53 ok any questions on doc bugs? 13:41:01 nermina: +1000 13:41:12 #topic Design Summit sessions 13:41:30 #link http://summit.openstack.org 13:41:41 You can click the Topic column to sort 13:42:12 There are four proposed for Documentation and we have four slots, but can move things around as needed, so do propose by end of this week or next 13:42:48 also create blueprints 13:42:55 fifieldt: yes 13:43:17 fifieldt: Diane has a blueprint but isn't going to the Summit, about API docs, that I'll probably discuss during the restructure one 13:43:28 fifieldt: if you don't mind using the restructure as a catch-all 13:43:44 sure, I don't even recall what the original idea I had in mind was 13:44:04 "Following from the session at the previous summit[1][2], let's take stock. Assessing how we did in our restructure efforts over Havana, and plan what structural work is needed over the next cycle or two." 13:44:26 sounds reasonable 13:44:36 ok 13:44:47 #topic Open Discussion 13:44:58 I do want to talk about meeting times summerl :) 13:45:08 :) 13:45:12 zzzz 13:45:29 fifieldt, i would like to work on this, Tool to get statistic data in Transifex 13:45:33 well let's take one step back, is this meeting nice for status? 13:45:45 work division? 13:45:48 news-gathering? 13:45:52 what are the goals for this meeting? 13:46:04 sounds great chandankumar, let's discus offline? I need to sleep soon 13:46:13 yes 13:46:29 annegentle, I think I'm just a wuss, everyone else on the team is all like, yeah, nighttime. 13:46:51 random data point: I am beginning to like the idea of little mini-meets during the timezone overlaps AU->EU, EU->US, US->AU 13:46:52 summerl: heh I'm a morning wuss (and missed too many early ones I'll admit)! I'm also a nightime wuss. 13:46:57 Double wuss status! 13:47:07 !! 13:47:08 summerl: Error: "!" is not a valid command. 13:47:14 :D 13:47:24 We have weekly-ish newsletters What's Up Doc and weekly office hours, what can the meeting fill? 13:47:25 lol 13:47:42 (I'm really just asking, not asking leading questions) 13:48:14 You'll also notice my fellow Rackers aren't at this one and it's only 8 here 13:48:17 so yeah 13:48:25 want to figure out the best goals for meetings 13:48:34 strawman: forming fast consensus on difficult aspects that have previously been raised on the mailing list 13:48:40 it's just as needed, right, if someone is on 13:49:12 fifieldt: yeah I want to talk more about the positives of meetings and how to fill the needs 13:49:31 fifieldt: we know the difficulties, so what can we do to meet needs? 13:49:40 summerl: do you really like a check-in time? 13:49:43 Definitely good for critical situations, annegentle 13:49:56 summerl: yep totally agree 13:50:07 and sgordon came up with 2x a month nearing the release 13:50:11 what's best for post release? 13:50:18 Not trying to over complicate either! :) 13:50:40 if we should meet weekly because all the other programs do, there's that too. I was hoping office hours filled that need 13:50:52 but office hours attendance has dropped 13:51:25 maybe office hours could go to my evening/ Aus day? 13:51:31 and we find a new meeting time? 13:51:54 in terms of regularity, I like it when I login of a morning and annegentle tells me what's going on, then I can pass the messages on to AJaeger in the afternoon when he logs in. I find that useful for status. 13:51:59 think on that, APAC folks, I'd love office hours on your side of the world 13:52:10 fifieldt: yeah that's valueable 13:52:23 about to speak with mirantis qa, annegentle, anything else to test besides compute 13:52:32 lets' think about an evening time then 13:52:43 nermina: networking! nova-network and neutron 13:52:48 annegentle, sounds good. 13:52:52 * annegentle loves QA/QE 13:52:55 nermina, object storage, heat and ceilometer as well 13:53:02 the latter two may just not work at all 13:53:08 so tell them to fail fast 13:53:12 fifieldt: nermina: but definitely top priority networking :) 13:53:15 heh yes fail fast 13:53:19 sure 13:53:28 ok, anything else? 13:53:46 I haven't seen any proof of concept on a redesign from Todd Morey, I keep checking in with hi 13:53:48 him 13:54:02 oh and Nick Chase is working with Alice King, the Foundation lawyer, on the licensing details 13:54:13 they're drafting a memo outlining the issues to go to the board 13:54:29 annegentle, what are the basic issues? 13:54:32 Oh and please do vote for me for Technical Committee. I don't usually campaign hard but I do feel it's important 13:54:44 voted! 13:54:44 done 13:54:56 oh, I'm sure that's in my inbox somewhere 13:54:59 is it closing soon? 13:55:15 summerl: in October 2012 the board voted to use CCBY for the docs, but we never did the detail work of how to ensure contributors know the difference, and what to do with legacy licensed docs 13:55:32 fifieldt: probably this week, right? 13:55:40 hmmk 13:55:41 oh well 13:55:44 bed time for me 13:55:47 annegentle, and that will happen for havana? 13:55:52 #link https://wiki.openstack.org/wiki/TC_Elections_Fall_2013 13:56:22 summerl: honestly, hard to say. Since it's been a year since the licensing changed we could face another year of getting the tooling in place :) 13:56:32 ah, ok. 13:56:34 fifieldt: good night, and thanks for all the hard work! 13:57:08 alrighty! I can't think of anything else! 13:58:00 Thanks everyone for all the hard work, and pass that to your colleagues as well. I tell Diane and David thanks all the time! 13:58:03 #endmeeting