19:02:04 <jeblair> #startmeeting infra
19:02:05 <openstack> Meeting started Tue May 21 19:02:04 2013 UTC.  The chair is jeblair. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:02:08 <openstack> The meeting name has been set to 'infra'
19:02:20 <reed> you're all revealing your age with silly references to sillier cartoons
19:02:58 <fungi> reed: yes, i'll be 6 years old for the rest of my life probably
19:03:00 <jeblair> #link https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting
19:03:07 <jeblair> agenda ^
19:03:45 <jeblair> #topic actions from last meeting
19:03:49 <jeblair> there are no actions from last meeting
19:03:55 <jeblair> #link http://eavesdrop.openstack.org/meetings/infra/2013/infra.2013-05-14-19.02.html
19:04:09 <jeblair> fungi: is there anything to talk about re jenkins slave os?
19:04:32 <fungi> jeblair: should probably talk about taking it off the agenda so you don't ask again next week ;)
19:04:43 <fungi> i'll take care of that
19:04:48 <jeblair> fungi: done
19:04:56 <jeblair> #topic testr in project progress
19:04:56 <jeblair> clarkb: ?
19:05:22 <reed> I vote to abolish passwords from websites, there has to be a better way to authenticate
19:05:32 <clarkb> hi
19:05:45 <clarkb> ceilometer has been testr'd according to lp
19:05:58 <clarkb> jgriffith plans to try and get cinder testr'd during H-2
19:07:06 <clarkb> still need to convince the keystone folks that this is worth doing. I think swift will be on board if I poke them.
19:07:15 <clarkb> glance is going to be the fun one because their test suite is interesting
19:07:46 <clarkb> things are moving, just slowly
19:08:40 <jeblair> is there a particular strong objection from keystone, or just no one willing to pitch in?
19:09:25 <clarkb> it seemed to more a matter of motivating effort and not a technical objection
19:09:29 <dolphm> for testr support? no objections here
19:09:45 <clarkb> yeah it was called busy work at one point. But I don't think anyone is against testr itself
19:10:39 <dolphm> keystoneclient is already done
19:11:08 <jeblair> okay, we'll just need to promise everlasting fame and glory to whoever does keystone
19:12:05 <zaro> o/
19:12:10 <jeblair> i think the bug calendar agenda item is stale, so i'm converting it to a reminder:
19:12:16 <jeblair> Next bug day: June 4, 2013 at 1700UTC
19:12:35 <jeblair> #topic Publishing of old documentation now that diablo and essex are EOL.
19:12:42 <jeblair> #link https://bugs.launchpad.net/openstack-ci/+bug/972724
19:12:44 <uvirtbot> Launchpad bug 972724 in openstack-ci "Dev docs sites for past releases of nova (2011.1) now 404 error" [Medium,Triaged]
19:12:57 <jeblair> i think we needed input from annegentle on that
19:13:00 <annegentle> there's one for swift too
19:13:56 <jeblair> clarkb: do you want to speak?
19:13:57 <annegentle> #link https://bugs.launchpad.net/openstack-manuals/+bug/1025099
19:13:59 <uvirtbot> Launchpad bug 1025099 in openstack-org "Redirects needed for swift 1.4.3, 1.4.4, 1.4.5, etc doc builds" [Undecided,Confirmed]
19:14:05 <annegentle> I marked it won't fix
19:14:30 <clarkb> oh yes
19:14:44 <clarkb> annegentle: so you are happy letting those die on the vine?
19:15:04 <clarkb> annegentle: anything folsom or greater will have the correct files in place. its just a matter of linking to them where you want
19:15:17 <annegentle> clarkb: that sounds like a great idea
19:15:34 <annegentle> clarkb: so there's still a task to link to the "correct" versions?
19:15:40 <annegentle> clarkb: but it's folsom and up
19:15:56 <clarkb> annegentle: that depends. We are copying them to specific places
19:16:08 <clarkb> annegentle: it will be under the project/tag_name dir for that project
19:16:15 <annegentle> clarkb: oh, ok
19:16:36 <clarkb> annegentle: I don't think we are currently automagically editing links within the aggregator sections of the docs
19:16:42 <annegentle> clarkb: that sounds like a marvelous solution, is it just a matter of assigning someone to edit?
19:16:49 <annegentle> clarkb: automagic is even better
19:17:10 <clarkb> My worry with automagic is I am not sure how friendly docbook will be to that
19:17:35 <clarkb> yes, if you want to link to things today you can just do so by hand
19:17:49 <annegentle> clarkb: these aren't docbook, theyre sphinx sites
19:18:01 <clarkb> http://docs.openstack.org/developer/python-novaclient/2.9.0/ for example
19:18:22 <clarkb> and http://docs.openstack.org/developer/nova/2013.1/
19:18:24 <annegentle> clarkb: but tell me why it matters?
19:19:04 <clarkb> annegentle: because making a thing to automagically add links when nova et al tag a release is something new that would need to be written
19:19:10 <clarkb> we don't currently have anything like that today
19:19:54 <clarkb> and it would probably propose a change rather than merge it automatically so the human element remains
19:20:27 <jeblair> clarkb: where would you add links?
19:20:45 <annegentle> jeblair: the links existed in the sidebars of the nova and swift sphinx sites
19:21:05 <annegentle> jeblair: with the new templating for nova for example, the links aren't there
19:21:33 <clarkb> oh I was thinking the links would be at http://docs.openstack.org/developer/ or similar
19:21:42 <annegentle> clarkb: we manually maintain http://docs.openstack.org/developer/openstack-projects.html
19:22:13 <annegentle> for example, jeblair, cinder uses the "old" template at http://docs.openstack.org/developer/cinder/
19:22:25 <clarkb> annegentle: the problem with putting links internal to nova itself is that if you compile those sphinx docs without having the old stuff you will get broken links
19:23:16 <annegentle> clarkb: yeah I'm fine with removing the links then
19:24:25 <clarkb> annegentle: does that mean we need to go and clean some of the sphinx docs up? is that something the doc team would do?
19:24:38 <annegentle> clarkb: honestly we don't have the time/resources for sphinx docs
19:24:58 <annegentle> clarkb: I see a new doc bug for cinder to not link to "versioned" sites that are broken
19:25:04 <annegentle> clarkb: so I guess just log bugs?
19:25:13 <clarkb> in that case probably best file bugs against the projects themselves
19:25:13 <annegentle> clarkb: and indicate the tagging mechanism?
19:25:16 <clarkb> annegentle: yup
19:25:23 <annegentle> clarkb: makes sense to me
19:25:57 <dprince> clarkb: log question for you. What is the best way to handle the logs for this jenkins job: https://review.openstack.org/#/c/29978/
19:26:45 <jeblair> dprince: this is the meeting channel.  :)
19:26:57 <dprince> jeblair: thanks!
19:27:18 <jeblair> clarkb: any other questions about old docs?
19:27:35 <clarkb> jeblair: nope I think that covers it
19:27:49 <jeblair> clarkb: clouddocs-maven-plugin internalization?  do we need to talk about that, or is that now an off-line task?
19:28:11 <clarkb> jeblair: I think that is now an offline task
19:28:15 <annegentle> jeblair: I really really really want forward motion on that though.
19:28:32 <annegentle> where's it at? Waiting on dwcramer?
19:28:36 <clarkb> annegentle: I started looking at it last week. zaro and I discussed the jenkins jobs
19:28:38 <jeblair> annegentle: i agree, i only meant "is clarkb working on that".  :)
19:28:50 <annegentle> ah ok
19:28:53 <clarkb> annegentle: the biggest thing right now is having working jenkins jobs to do releases to maven
19:29:12 <clarkb> once we have that we can do the switch without impacting your release process
19:29:15 <annegentle> oh ok I wasn't sure what the next steps were after the summit, I'm behind :)
19:29:34 <annegentle> clarkb: NICE in all caps for sure then.
19:29:52 <zaro> i entered bug on that: https://bugs.launchpad.net/openstack-ci/+bug/1182154
19:29:53 <uvirtbot> Launchpad bug 1182154 in openstack-ci "Automatic deploy to jenkins-ci.org repository" [Undecided,New]
19:30:15 <jeblair> lets rearrange the agendo to continue talking about annegentle's topics:
19:30:18 <jeblair> #topic  OpenStack Operations Guide - tooling, licensing, publishing discussion
19:30:24 <annegentle> ok thanks jeblair
19:30:31 <annegentle> do you want an overview?
19:30:37 <jeblair> annegentle: would love one
19:30:43 <annegentle> ok, here goes
19:30:45 <dwcramer> I'll sync up with  you guys on the clouddocs-maven-plugin soon.
19:30:54 <clarkb> dwcramer: awesome
19:30:57 <dwcramer> I learned a bit more about how Maven works.
19:31:05 <dwcramer> That I didn't know at the summit.
19:31:25 <annegentle> At the Summit, I got to talk to O'Reilly community manager and a strategic accounts manager.
19:31:40 <annegentle> They mentioned that they've had difficulty getting an author to finish an OpenStack book
19:31:47 <annegentle> I had heard this previously from an O'Reilly editor.
19:32:20 <annegentle> Since we have a "completed" book more or less, they wanted to know if the Foundation wanted pro-editing on it and more publishing channels and epub-ish outputs.
19:32:58 <annegentle> Currently we have non-pro editing, a Lulu publishing channel, and no epub-like output that is automated.
19:33:30 <annegentle> so we started the conversation and asked for a few meetings and a proposal for a contract with O'Reilly between them and the Foundation
19:34:21 <annegentle> the O'Reilly workflow involves a tool called Atlas. Authors collaborate in a github repo using docbook or asciidoc with a web front editor, O'Reilly editors interact with the content and authors online.
19:34:54 <annegentle> O'Reilly uses Github to gather public feedback about Atlas
19:34:56 <annegentle> #link https://github.com/MakerPress/atlas-public-feedback/issues?labels=&page=3&state=open
19:35:16 <annegentle> but as far as I can tell, Atlas is not open sourced
19:35:48 <annegentle> since it's git backed, my (likely mistaken) hope was that we could use the same repo to publish through our current Gerrit/Jenkins process and to O'Reilly's channels.
19:36:10 <annegentle> however we're very early in the days of exploration and learning
19:36:30 <annegentle> the latest today is that Mark Collier at the Foundation has a copy of O'Reillys proposal and is reviewing it
19:36:42 <annegentle> I'm not sure it's quite what we had in mind but we're going to discuss.
19:36:53 <annegentle> whew.
19:37:11 <annegentle> I'd like to connect team CI to O'Reilly technical contacts.
19:37:44 <annegentle> Since I have questions from one perspective, but you're going to have questions from another.
19:37:46 <jeblair> annegentle: that sounds like a great idea.  i'd be happy to talk to them
19:38:08 <jeblair> annegentle: i'm really curious about the contract with the foundation...
19:38:11 <annegentle> The other part is that the authors have to agree to do this, so far all six of them are onboard, for finding out more details.
19:38:25 <annegentle> All of this is not a done deal at all. I know you all know this but emphasizing for anyone reading the logs.
19:38:47 <jeblair> a contract around marketing, trademarks, etc makes sense
19:39:11 <jeblair> if the contract involves tooling, that seems weird.  and i don't see how there could/should be a need for a contract regarding rights to the content itself
19:39:14 <annegentle> Another consideration is translation for this particular book.
19:39:16 <jeblair> since it's CC-by licensed
19:39:23 <annegentle> we have a Chinese and Japanese version already
19:39:44 <annegentle> jeblair: yeah the contract I saw today isn't what I expected in terms of keeping CC-By.
19:39:50 <annegentle> jeblair: so we have a ways to go
19:40:25 <annegentle> jeblair: does it make sense to loop you in to the contract info with Mark Collier?
19:40:32 <jeblair> so with the caveat that we're sill exploring and learning about this, i'll just say some general thoughts based on what i've heard so far
19:40:37 <annegentle> jeblair: or not yet? I don't want to waste your time either.
19:40:52 <annegentle> jeblair: sure, want to hear more thoughts
19:41:23 <jeblair> i don't know anything at all about their tool and how it works, which is about as far from open as you can get
19:41:25 * mordred also wants to talk with them
19:41:29 <jeblair> and open is one of our core community values
19:41:38 * mordred is lurking over phone tethering waiting for plane to take off
19:42:42 <jeblair> i understand that the git workflow isn't always great for documentors, but i'm really concerned about whether their tool is a right fit for our project
19:43:04 <annegentle> mordred: thanks for that
19:43:46 <jeblair> then there's the technical issue, about whether it would work with anything else we're doing at all.
19:44:19 <annegentle> jeblair: so for the use case of accessing an established publishing brand, I don't know how you'd accomplish that with a completely open source tool
19:44:31 <mordred> annegentle: I reject that premise
19:44:31 <jeblair> annegentle: i don't know what that means.  :(
19:44:34 <annegentle> jeblair: it's like the training/certification tooling, will we be able to have all-open for that?
19:44:39 <mordred> annegentle: yes
19:44:43 <annegentle> mordred: fair 'nough :)
19:44:54 <mordred> annegentle: o'reilly has made MILLIONS on open source
19:44:57 * ttx waves
19:45:03 <annegentle> hey ttx
19:45:05 <mordred> I believe it's not too much to ask for tools to interact with them to be open
19:45:16 <mordred> or else I'm going to have to call Tim out as a hypocrite
19:45:21 <annegentle> mordred: I certainly have made that request. I'm just saying, what if it's not possible.
19:45:42 <mordred> if it's not possible, then I think we'll need to have a very deep though about what's important to us
19:45:55 <mordred> because if all we want is world domination and we don't also want openness
19:45:58 <mordred> that's one path forward
19:45:58 <annegentle> we use open source fonts, we are open source as far as we can for the use case of publishing documentation. For publishing books, do you need to have Lulu's source for their ecommerce site?
19:46:28 <annegentle> is it possible that certain use cases don't lend themselves to all-open, that's all I'm asking here.
19:46:44 <mordred> I'm not passing judgement on any specific case
19:46:50 <mordred> I'm saying that so far al of our tools are open
19:46:54 <mordred> and it's an important community value
19:46:56 <jeblair> annegentle: i don't think the contribution workflow for our documentation should be proprietary
19:47:10 <mordred> so if we're going to undercut that, I want us to fully understand the ramifications of that
19:47:25 <mordred> and I _personally_ do not believe there is any valid reason why we should need a non-free tool
19:47:35 <jeblair> annegentle: i think anyone should be free to publish our documentation using whatever systems they want.  but i don't think we should start embracing propretary access to contributing.
19:47:43 <mordred> ++
19:47:44 <clarkb> ++
19:47:49 <annegentle> jeblair: ah, okay, that helps me understand
19:48:05 <fungi> well said
19:48:05 <annegentle> jeblair: I don't really want to stop contributors from using our current workflow either.
19:48:48 <mordred> right.
19:48:55 <jeblair> annegentle: yeah, i like to think that being able to contribute to docs in the same way as the rest of the project is a worthwhile system.
19:49:02 <mordred> which means that we'll need to have some deep conversations with oreilly about things
19:49:11 <mordred> because our workflow is likely different than what they're planning for
19:49:19 <annegentle> so to me, when I look at it, I get git-backing. But yesterday you pointed out that OpenStack source is not technically housed on github.
19:49:25 <mordred> if their system is _a_ way to get at things, then I'm less opposed to it
19:49:34 <mordred> annegentle: that is correct
19:49:37 <annegentle> So I saw Atlas as a second view into the source.
19:49:51 <mordred> if it is a second view, I will have many less objections
19:50:02 <mordred> however, oreilly is going to have to grok how our system works
19:50:12 <mordred> because so far most of the 3rd party folks who have tried ot "integrate" with us
19:50:18 <mordred> don't seem to understand that we don't use github
19:50:21 <annegentle> mordred: ok, good -- I think we have the same core value, but I had a bit of a bridge to build from github to openstack that I didn't have conceptually yet.
19:50:27 <mordred> yah
19:50:38 <mordred> it's definitely a conceptual bridge
19:50:38 <fungi> annegentle: if it just consumes our git history and displays it, that's one thing... but if they have editors working on t we'll want a feedback loop of some kind
19:50:55 <clarkb> yeah this is similar to the transifex issue
19:50:56 <mordred> right. which means figuring out how that interacts with a review pipeline
19:50:58 <annegentle> I think I have it conceptually now, but we still need a technical bridge. Are you all willing to explore the technical framework a bit? I can really use the help.
19:51:04 <mordred> absolutely
19:51:07 <jeblair> yep
19:51:08 <clarkb> we are basically forcing everything to be unidirectional for sanity
19:51:26 <mordred> although - it would be easier for us to have that discussion with them if their tool wasn't proprietary junk :)
19:51:28 <annegentle> clarkb: ooo that's a good one too, Transifex as a second source.
19:52:04 <annegentle> mordred: I honestly didn't get the vibe that it would never be open source, more like "we're a tad bit disorganized to actually show everyone the source"
19:52:17 <jeblair> annegentle: maybe we can help motivate them.  :)
19:52:19 <clarkb> annegentle: not anymore. transifex is now the only source
19:52:23 <fungi> if they wait until it
19:52:39 <clarkb> if you consider automagic process
19:52:43 <annegentle> Ok, so I think the agreement is to connect the ci team to O'Reilly technical contact(s) (I'm pretty sure I have at least one). How should I do that? Email?
19:52:46 <clarkb> you can still propose a fix to things by hand
19:53:11 <fungi> if they wait until it's bug-free to release, they'll be the only bug-free open source project in history i think
19:53:11 <mordred> annegentle: email is probably a great intro - I'm guessing it'll take a phone call too, since there is a LOT of ground to cover
19:53:11 <jeblair> annegentle: my guess is email to set up a phone call
19:53:20 <annegentle> jeblair: yeah ok
19:53:21 <mordred> I'd be happy to provide a conference bridge for that purpose
19:53:27 <annegentle> mordred: ok I have one too.
19:53:30 <mordred> awesome
19:53:51 <annegentle> Ok, thanks for exploring... the crazy world of dead trees and ISBNs.
19:54:08 <annegentle> sorry to take up so much tim
19:54:09 <annegentle> time
19:54:15 <jeblair> annegentle: thanks for chatting with us here about this; i think it's important
19:54:18 <mordred> ++
19:54:24 <jeblair> there's a few minutes left to talk about remote participation....
19:54:30 <jeblair> #topic What can infra do to enable remote participation at the next Design Summit?
19:54:34 <jeblair> reed: around?
19:54:35 <fungi> i have academic/technology librarian friends who are big into free/open publication. i can solicit some suggestions there too
19:54:52 <annegentle> fungi: cool
19:56:00 <jeblair> #topic open discussion
19:57:06 <annegentle> I could swear I had something for open discussion but it has slipped my mind.
19:57:10 <annegentle> Ahwell.
19:57:14 <mordred> jeblair: just was chatting with zul about pymox
19:57:23 <mordred> jeblair: apparently there is a python3 compat fork of it
19:57:30 <mordred> jeblair: but the upstream will not take the patch
19:57:33 <mordred> sigh
19:57:35 <jeblair> (!)
19:57:56 <mordred> this brought up the idea of making a pymox3 fork, hosting it ourselves and publishing to pypi ourselves
19:57:57 <jeblair> mordred: fork all the things
19:58:00 <mordred> yeah
19:58:03 <mordred> I mean
19:58:06 <mordred> not my first choice, but also, jeez
19:58:34 <mordred> so, zul was going to look at adding it as a stackforge project, but I was thinking it might should be an openstack-dev/pymox3 instead?
19:58:36 <clarkb> mordred: any response from holger?
19:58:37 <mordred> thoughts?
19:58:46 <jeblair> annegentle: did you want volunteers to convert all the manuals to TeX?
19:58:50 <fungi> perhaps this python 3 support craze is just a passing fad?
19:58:54 <annegentle> jeblair: LOL
19:58:58 <mordred> clarkb: yes. he and I are going to have a phone call the next couple of days
19:59:07 <clarkb> mordred: ! yay
19:59:14 <mordred> zul is putting in a lot of effort to python3-ify our stuff, and I'd like to be supportive of that
19:59:21 <jeblair> mordred: openstack-dev seems right
19:59:23 <clarkb> mordred: I am more than happy to be the hg guinea pig if that helps
19:59:30 <mordred> jeblair: kk
19:59:46 <mordred> clarkb: I will get confirmation from him on acceptability of the idea, and then point you at it if he says ok
19:59:46 <fungi> i am honestly thrilled to see more attention paid to python 3 support
19:59:57 <mordred> zul: ^^ openstack-dev/pymox3
20:00:06 <jeblair> and on that bombshell
20:00:08 <jeblair> #endmeeting