19:03:08 <jeblair> #startmeeting infra
19:03:09 <openstack> Meeting started Tue Jan 22 19:03:08 2013 UTC.  The chair is jeblair. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:03:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:03:12 <openstack> The meeting name has been set to 'infra'
19:03:21 <jeblair> #topic agenda
19:03:26 <jeblair> #link http://wiki.openstack.org/Meetings/InfraTeamMeeting
19:03:35 <jeblair> I like, updated the agenda and stuff.
19:03:45 <fungi> oh!
19:03:53 <fungi> i had stopped looking at it
19:03:53 <jeblair> anyone, everyone, feel free to update that at any time
19:04:21 <mordred> dude
19:04:23 <clarkb> woah
19:04:41 <jeblair> since i think of things throughout the week, it's nice to have a place to put it. and we can stop doing the "does anyone have an agenda" thing.  which is just way too exciting.
19:05:04 <pleia2> great
19:05:17 * mordred is in favor of our new agenda wiki overlords
19:05:37 <jeblair> #link last meeting: http://eavesdrop.openstack.org/meetings/infra/2013/infra.2013-01-15-19.04.html
19:05:52 <jeblair> the wiki/calendar are updated (thanks ttx)
19:06:10 <jeblair> and the rest we'll get to with the wiki (which i want to save for last when ttx will hopefully be here)
19:06:23 <jeblair> #topic CLA
19:06:42 <fungi> yeah
19:07:12 <fungi> e-mailed toddmorey about getting the updated appsec key to him, so see if he wanted to do that by phone or what
19:07:21 <fungi> review-dev will no longer break when he changes that
19:07:26 <mordred> neat
19:07:34 <fungi> still poking at doc updates as i get time
19:07:44 <fungi> i think we're probably clear to send the announcement though?
19:07:47 <jeblair> fungi: i gave him a heads up about that on a phone call this morning
19:08:13 <fungi> i've updated the announcement proposed text with the date we agreed upon
19:08:21 <jeblair> fungi: http://wiki.openstack.org/Project_Group_Management  is dead link?
19:08:41 <fungi> jeblair: and should be until we're doing project management in gerrit i think?
19:08:51 <jeblair> fungi: oh, it's in the annourcement
19:09:10 <fungi> double-checking, but thought the announcement said it would appear there once we cut over
19:09:15 <fungi> revisiting
19:09:17 <jeblair> fungi: yeah it does
19:09:38 <jeblair> fungi: how about we make that page exist with the docs, but put a "not implemented" banner at the top
19:09:44 <fungi> didn't want to publish documentation about a process which won't work yet, though i guess i could just slap a big disclaimer on it in that case
19:09:44 * mordred is in favor of our new CLA overlords
19:09:51 <fungi> yeh, that
19:10:01 <fungi> will do
19:10:06 <jeblair> fungi: cool, then i think it's ready to go.
19:10:24 <fungi> need to write that up, but i'll send the announcement as soon as that wiki page exists then
19:10:40 <jeblair> fungi: cool, by end of week reasonable?
19:10:52 <fungi> absolutely reasonable
19:10:55 <jeblair> #action fungi make http://wiki.openstack.org/Project_Group_Management exist
19:11:00 <jeblair> #action fungi send CLA announcement
19:11:03 <mordred> so - for my out-of-touch brain - does that mean we're cutting prod over to new CLA by end of week?
19:11:10 <fungi> ha
19:11:13 <mordred> or putting up docs and announcement for future cutover
19:11:14 <fungi> mordred: february 24
19:11:17 <mordred> k. thanks
19:11:33 <jeblair> mordred: it's right after g3
19:11:50 <mordred> oh right. I remember that now. I'm going to put it on my calendar
19:12:02 <jeblair> anything else about cla?
19:12:04 <fungi> i figure this week is good for announcing, since it gives everyone a one-month warning
19:12:09 <jeblair> fungi: +1
19:12:16 <fungi> that's all there is from me on that
19:12:27 <jeblair> #topic rechecks
19:12:40 <jeblair> So recheckwatch is in production...
19:12:51 <mordred> I'd like to quote markmc:
19:12:51 <clarkb> and people are using it :)
19:12:53 <mordred> "You know what? This is just awesome."
19:12:59 <jeblair> looks like it has data:
19:13:01 <jeblair> #link http://status.openstack.org/rechecks/
19:13:34 <mordred> feature request - how do things get off the list?
19:13:36 <jeblair> so the next/final step is to disable the old syntax.  should we do that now?
19:13:57 <jeblair> mordred: they age out after a month.  they fall down the list as they become less relevant.
19:14:20 <mordred> k
19:14:31 <fungi> i'm good with enforcing
19:14:44 <jeblair> mordred: we could give them additional demerits if the bug is closed (fix committed/released)
19:15:04 <jeblair> mordred: (to move them down the page faster, but probably not remove them)
19:15:10 <fungi> also worth noting, we've gotten an influx of ci bugs, many of which seem to relate to people being encouraged to document their rechecks/reverifies
19:15:21 <jeblair> mordred: (still want to find them easily if the bug is erroneously marked closed)
19:15:31 <mordred> fungi: ++
19:15:36 <jeblair> fungi: yeah, i think we should retarget those to whatever projects seem the most relevant
19:15:56 <jeblair> ie, https://bugs.launchpad.net/openstack-ci/+bug/1101142
19:15:58 <uvirtbot> Launchpad bug 1101142 in openstack-ci "Euca tests failure, cannot ping server with floating ip" [Undecided,New]
19:16:04 <jeblair> should probably be targeted to nova and devstack
19:16:05 <fungi> agreed. none of them were terribly clear to me on first glance, except one i thought was almost certainly nova
19:16:45 <fungi> the devstack ones are always going to be fuzzy, but maybe the devstack bug peeps can make with more refining once they get them
19:17:00 <jeblair> we should probably try to be fairly quick about that too, so that the devs get the benefit of the bugs.
19:17:04 <mordred> ++
19:17:06 <fungi> agreed
19:17:31 <mordred> generalized bug triage on the openstack-ci project is something I fall down at alot
19:17:48 <mordred> but I think that I have email filters that are too broad
19:17:52 <jeblair> so should we go ahead and require the new syntax?
19:18:06 <jeblair> or extend the burn-in period?
19:18:50 <clarkb> I think we can go ahead and turn it on
19:19:10 <clarkb> people seem to understand how it works, and it hasn't had major problems outside the thing jgriffith pointed out
19:19:18 <jeblair> #action jeblair update recheck regex to require bug
19:19:34 <fungi> and discussion about it has more or less died down on the ml with no real detractors
19:19:54 <jeblair> clarkb: yeah, i'm running that in screen now with the fix in place to catch any further errors.  i should also add proper logging to it.
19:20:03 <jeblair> #action jeblair add logging to recheckwatch
19:20:26 <jeblair> #topic jenkins slave operating systems
19:20:53 <jeblair> we should upgrade our precise slaves to quantal, in accordance with the support policy
19:20:59 <mordred> ++
19:21:07 <mordred> also, our cloud providers have quantal now
19:21:22 <mordred> heads up - I believe there is somethign up with running devstack on quantal ...
19:21:42 <jeblair> awesome.
19:21:43 <mordred> devananda: any chance you know what that is? ^^
19:21:48 <clarkb> quantal uses biosname for interfaces
19:21:52 <mordred> that's it
19:21:57 <mordred> devananda: nevermind
19:22:03 <clarkb> and that appeared to cause some problems with devstack networking
19:22:09 * devananda reads scrollback anyway
19:22:09 <clarkb> s/interfaces/network interfaces/
19:22:26 <mordred> yeah. so eth0 is now em0 or something right?
19:22:29 <clarkb> correct
19:22:34 <jeblair> is there a devstack fix?
19:23:03 <jeblair> so i think we need to:
19:23:19 <jeblair> 0) update puppet with a node descriptor for quantal
19:23:20 <devananda> there was an issue with node.js on quantal too: https://bugs.launchpad.net/horizon/+bug/1070083
19:23:22 <uvirtbot> Launchpad bug 1070083 in devstack "devstack fails in quantal - /usr/bin/env node - no such file or directory" [Undecided,Fix released]
19:23:31 <jeblair> 1) launch a long-running quantal slave node
19:23:37 <jeblair> 2) add it to jenkins
19:23:49 <fungi> presumably node needs to be invoked as nodejs on quantal?
19:23:56 <fungi> devananda: ^?
19:24:12 <mordred> jeblair: yah. that way we can try adding some non-voting jobs first
19:24:16 <jeblair> 3) try manually running some unit test jobs on it for various projects
19:24:17 <fungi> oh, already fixed i guess
19:24:21 <jeblair> mordred: or that
19:24:40 <mordred> jeblair: are you just talking about unittest nodes now? or devstack nodes too?
19:24:42 <jeblair> 4) move projects or duplicate as non-voting jobs to the quantal slaves
19:24:45 <jeblair> just unit test
19:24:50 <mordred> k.
19:24:55 <jeblair> 5) add it to the devstack-gate pool
19:25:06 <jeblair> 6) try out devstack on it
19:25:16 <jeblair> [end of rough overview]
19:25:26 <mordred> ++
19:25:26 <clarkb> sounds about right to me
19:25:36 <fungi> seems sane
19:25:37 <jeblair> anyone want to start working down that list?
19:25:54 * mordred is about to get 20 hours of plane flights, so it's probably not me
19:26:05 <jeblair> i have similar availability
19:26:32 <fungi> i can start trying it once i wrap up the oneiric/quantal puppet cleanup stuff
19:26:36 <clarkb> I am hoping to do more logstash related hacking, but if I get stalled on that I can probably spin up a quantal node and get the process going
19:26:47 <clarkb> s/what I said/I can assist fungi/ :)
19:26:50 <pleia2> fungi: let me know if there is anything I can do to help
19:27:06 <fungi> pleia2: i was about to suggest you might want to be involved too
19:27:17 <jeblair> #action fungi start quantal upgrade with help from pleia2 and clarkb
19:27:36 <jeblair> and me too, just at weird hours.  :)
19:27:48 <jeblair> davidkranz: ping
19:27:54 <jeblair> #topic tempest gating
19:28:14 <jeblair> so great news is we're gating on quantum
19:28:40 <jeblair> and the tempest folks are really making headway on getting tempest ready for more intensive gating
19:28:57 <jeblair> (and we're all really excited about testr too!)
19:29:15 <jeblair> davidkranz: sent me this link this morning:
19:29:17 <jeblair> #link https://etherpad.openstack.org/turn-on-tempest-gate
19:29:59 <jeblair> which i believe describes an optimization of the gate where we can avoid running some tests for some projects
19:30:52 <jeblair> that sounds nice, though i'm a little concerned about whether that means we'll miss problems due to unforseen interactions...
19:31:10 <clarkb> which could cause the gate to grind to a halt
19:31:19 <jeblair> from a high level pov, it's also an imbalanced gate, where integrated projects run different code.
19:31:43 <jeblair> the current "tempest runs full" configuration is a form of that.  obviously, this is _designed_ to be better than that
19:31:54 <jeblair> because it's supposed to only test related code
19:32:15 <jeblair> but where reality doesn't match the design, we could see a problem.
19:32:31 <jeblair> clarkb: exactly.
19:33:02 <mordred> so, just in general, when it says  "It is too expensive to run complete functional tests for every project on a commit to any project. So
19:33:04 <mordred> the following is proposed.
19:33:06 <mordred> "
19:33:47 <mordred> I'd love to see what parallelization gets us before we start doing partial tests
19:34:33 <mordred> but, in general the matrix in that doesn't look ridiculous
19:35:25 <fungi> well, parallelization trades expensive time for expensive space (ram, cpu), so it depends on what expensive means there
19:35:27 <jeblair> yeah.  also, we could do something like the xml/json split in the interim before testr.
19:35:51 <jeblair> fungi: indeed.  and we don't know our budget.  :)
19:36:20 <fungi> at least for time, there is a fixed budget. no more than 26 hours a day ;)
19:36:30 <jeblair> heh
19:36:40 <fungi> (or was that letters in the alphabet?)
19:37:01 <mordred> yeah. I've been under the working assumption so far that dev time is the more precious and that our illustrious cloud providers haven't been unhappy with our current usage
19:37:37 <jeblair> mordred: yep.  and you know i'm happy to increase our usage -- i only want to make sure people understand the exponential function.
19:37:43 <mordred> jeblair: ++
19:38:20 <jeblair> mordred: so maybe let's start an email thread with some tempest folks, and work out a med-long range plan?
19:38:31 <mordred> jeblair: ++
19:38:44 <jeblair> #action jeblair start email thread for tempest gate long range planning
19:39:03 <jeblair> #topic wiki migration
19:39:13 <mordred> jeblair: although I will say that the theory of running relevant tests is a nice one
19:39:30 <mikal> jeblair: can you do me a favour?
19:39:35 <mikal> mordred: you too?
19:39:43 <olaph> so, logo/links should be acceptable for v1 of the skin
19:39:58 <mikal> I just sent you guys an email with a PDF attachment. Can you verify the attachment opens and renders ok, and looks correct for me?
19:40:15 <jeblair> mikal: sure.  this is a meeting, btw.  :)
19:40:33 <mikal> Oh bugger. Sorry. Wrong channel.
19:40:46 <jeblair> mikal: it's okay, we'll put you on the agenda for next time.  :)
19:40:53 <mikal> Yay
19:40:56 <jeblair> olaph: cool, updated link?
19:41:16 <clarkb> I think ttx is still having trouble with image uploads
19:41:29 <olaph> same one:
19:41:37 <olaph> #link http://openstack-wiki-instance.instance-proxy.wmflabs.org/wiki/Main_Page
19:42:30 <clarkb> I have no css for that (it must be pulling it from behind the firewall
19:43:35 <jeblair> olaph: that lgtm (the only thing that really stands out is the mediawiki footer
19:44:00 <olaph> jeblair: should it disappear, or just be openstackish?
19:44:23 <clarkb> silly caches, loads fine in chrome
19:44:33 <jeblair> olaph: probably just have blue lines instead of orange.
19:44:39 <jeblair> or grey lines.  or something.
19:45:34 <ttx> o/
19:45:40 <mordred> look. it's ttx
19:45:47 <ttx> and Ryan
19:45:56 <jeblair> Ryan_Lane: your redirects patch is merged, right?
19:46:01 <Ryan_Lane> yep
19:46:02 <jeblair> Ryan_Lane: is that taken care of?
19:46:07 <Ryan_Lane> it is
19:46:15 <Ryan_Lane> it'll work when we switch the domain name
19:46:19 <ttx> yeah, my only standing issue is the inability to upload images
19:46:35 <ttx> or rather, that uploaded images return 404 when served
19:46:45 <Ryan_Lane> ttx: that's also fixed
19:46:55 <ttx> The main page is looking good enough, need image upload to make them look same
19:47:04 <ttx> Ryan_Lane: recently ?
19:47:11 <Ryan_Lane> I emailed the list in response to you ;)
19:47:14 <Ryan_Lane> the same day
19:47:24 <ttx> I tried like 8 hours ago
19:47:27 <Ryan_Lane> crap
19:47:41 <ttx> see my own answer to list :P
19:47:54 <Ryan_Lane> well, maybe my change was wrong
19:48:02 <Ryan_Lane> I tested this before making the change and it worked
19:48:05 <Ryan_Lane> I'll fix that now
19:48:27 <jeblair> annegentle is not here.  :(
19:48:40 <ttx> Main page relooked @ https://wiki-staging.openstack.org/wiki/Main_Page
19:49:15 <ttx> waiting for the theme to do the last relooking steps
19:49:17 <jeblair> ttx: lgtm (modulo images/css)
19:49:44 <ttx> Ryan_Lane: can we have __NOTITLE__ ?
19:50:04 <Ryan_Lane> hm. there's some way of handling this, I think
19:50:13 <ttx> not a big fan of the "Main page" title on that main page, maybe that's just me
19:50:29 <ttx> http://www.mediawiki.org/wiki/Extension:NoTitle
19:50:29 <Ryan_Lane> oh
19:50:35 <Ryan_Lane> for sure on the main page we can turn that off with css
19:50:51 <Ryan_Lane> see wikipedia, as an example
19:51:10 <Ryan_Lane> ugh. I'm dumb. I put the aliases in the wrong order
19:51:30 * mordred points finger at Ryan_Lane and laughs mockingly
19:51:42 * mordred apologizes
19:51:43 <clarkb> >_> I reviewed that change <_<
19:51:49 <ttx> Ryan_Lane: would that be part of olaph theme ? or do I need to do anything ?
19:52:10 <Ryan_Lane> you can modify mediawiki's css through mediawiki
19:52:15 <Ryan_Lane> I can handle that
19:52:19 <ttx> I tried {{DISPLAYTITLE:<span style="display:none">{{FULLPAGENAME}}</span>}} but it left a blank space
19:52:39 <Ryan_Lane> that's one way of doing it
19:52:49 <Ryan_Lane> that can be added anywhere in the page, including the bottom
19:52:57 <Ryan_Lane> and can be included in a template
19:52:58 <ttx> body.page-Main_Page h1.firstHeading { display:none; } in Common.css
19:53:09 <ttx> (following http://www.mediawiki.org/wiki/Manual:FAQ#How_do_I_hide_the_main_page_title.3F)
19:53:13 <Ryan_Lane> yep
19:53:26 <ttx> if I can modify it, will do if nobody beats me to it
19:53:44 <Ryan_Lane> cool
19:53:44 <olaph> either way...
19:53:58 <jeblair> if i read annegentle's email right, she was suggested we schedule the move _after_ the cla change on feb 24...
19:54:07 <jeblair> is that what other people got out of that?
19:54:32 <ttx> jeblair: no, my understanding was BEFORE
19:55:06 <Ryan_Lane> https://review.openstack.org/20268 <— fixes images
19:55:06 <ttx> We just need to avoid concurrency, and I don't see the point in waiting
19:55:21 <fungi> i'm cool either way. it's not like the wiki updates i'm staging will be hard to tweak the markup on
19:55:55 <jeblair> ttx: yeah, i was thinking earlier too.  i just thought her email said "after cla work is complete"
19:56:06 * ttx checks
19:56:18 <clarkb> I am ahppy with earlier as well. I think the wiki change is something that if it breaks we can work through it fairly easily
19:56:31 <Ryan_Lane> if the only changes in the skin are CSS, we may consider just adding it to MediaWiki:Common.css
19:56:39 <Ryan_Lane> or MediaWiki:<skin>.css
19:56:42 <ttx> I'd rather do atht early because end of Feb I'll get busy grizzlying
19:56:49 <clarkb> ++
19:57:00 <jeblair> since we also need to schedule a wiki update sprint, perhaps we should take the scheduling discussion to email to include anne...?
19:57:35 <olaph> Ryan_Lane: will do
19:57:53 <Ryan_Lane> olaph: thanks
19:58:30 <clarkb> jeblair: good idea
19:58:35 <jeblair> that, and we're out of time...
19:59:06 <jeblair> #action jeblair start wiki migration scheduling email thread
19:59:14 <jeblair> thanks all!
19:59:16 <jeblair> #endmeeting