19:03:18 <mtaylor> #startmeeting
19:03:19 <openstack> Meeting started Tue Mar 13 19:03:18 2012 UTC.  The chair is mtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:03:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:04:02 <mtaylor> how are people doing today?
19:04:16 <LinuxJedi_cell> Good thanks
19:04:22 <jeblair> great.  i would like to give a status update on things i'm doing.
19:04:34 <mtaylor> I would love that
19:04:52 <jeblair> should i do that now?
19:06:22 <jeblair> <mtaylor> yes, jim go ahead
19:06:24 <jeblair> mtaylor: thanks!
19:06:32 <LinuxJedi_cell> Lol :)
19:06:44 <jeblair> so i've changed the tarball jobs to trigger on ref-updated
19:07:19 <jeblair> and they no longer use the versions bzr repository, so we should have fewer failures due to the idiosynchrosies of that system
19:07:29 <jeblair> this necessitated changing the ppa jobs as well
19:07:32 <LinuxJedi_cell> Woot!
19:07:43 <mtaylor> woohoo
19:07:46 <jeblair> that's been completed for all projects.  here's a page with the listing of them:
19:07:51 <jeblair> https://jenkins.openstack.org/user/corvus/my-views/view/tarball/
19:08:08 <jeblair> there's one failure there because the packaging for keystone is currently broken
19:08:29 <jeblair> otherwise, of the jobs that have run since the change, they're working
19:08:57 <jeblair> the other thing i'm starting on today is a re-vamp of the way we do images for the devstack gate test
19:09:19 <jeblair> i'm fleshing out the database a bit more to better support multiple providers, and multiple images per-provider
19:09:55 <jeblair> i'm also planning on creating a new snapshot image from scratch each day, rather than trying to keep a long running machine updated and snapshotting from that
19:10:22 <jeblair> it turns out that strategy, which we're using now, fails a lot, and seems to cause problems with launching nodes when it does
19:10:29 <soren> ref-updated is an event sent out by gerrit?
19:10:33 <jeblair> soren: yes
19:10:37 <soren> (Catching up)
19:10:43 <soren> Ok, cool.
19:11:00 <jeblair> it's when the tip of a branch, or tag, or several other git-ish things change
19:11:32 <jeblair> so actually those jobs look for a ref updated for any branch, but ignore tags, etc.
19:11:59 <jeblair> but if we wanted, we can use that to trigger jobs when a tag is set, which could come in handy
19:12:05 <mtaylor> ++
19:12:23 <jeblair> so back to the devstack thing --
19:12:44 <mtaylor> yes
19:12:50 <jeblair> the main reason i'm working on this now is to make sure we're prepared to run devstack gate tests on precise when it comes out
19:13:16 <jeblair> so i want us to be able to easily build new images, and have the tests support multiple image types as needed as we change targets
19:14:07 <jeblair> so i'm taking a general approach in solving that so we're prepared for whatever we need to do. (hopefully)
19:14:11 <jeblair> that's about it for me.
19:14:15 <mtaylor> excellent
19:14:25 <mtaylor> btw - all awesome
19:14:25 <jeblair> oh, one more thing:
19:14:46 <jeblair> choumel added screen logging to devstack, so the gate tests now copy those over to jenkins
19:14:53 <mtaylor> AWESOME
19:14:55 <jeblair> that's a big win for developers being able to debug those tests.
19:15:13 <mtaylor> yes. that makes me very happy
19:15:51 <mtaylor> LinuxJedi_cell: you want to talk about what you've been hacking on, or do you want me to since you're on a cell phone
19:16:01 <mtaylor> (and/or - anybody have any questions for jeblair)
19:16:08 <heckj> yeah for chmouel
19:16:13 <heckj> yeah - had a quick question
19:16:21 <heckj> (may not be jeblair specific)
19:16:28 <LinuxJedi_cell> Mtaylor: would be good if you could or give me 5 minutes
19:17:08 <heckj> who are folks packaging downstream? Annegentle was poking me about install for keystone, and wanted to reference the packages for a common use case. I know the ubuntu folks, but didn't know who else was providing those packages to get some details around their setups/choices.
19:17:25 <mtaylor> LinuxJedi_cell: cool
19:17:28 <heckj> (or should I just poke the mailing list and see what comes out)?
19:18:00 <mtaylor> heckj: I'm not sure other than the stuff the folks at Canonical are doing
19:18:56 <heckj> mtaylor: kk - thanks.
19:19:04 <heckj> I'll probably poke the list and see what comes out
19:19:12 <mtaylor> cool. I think that's the best choice
19:19:48 <mtaylor> so, in my world, I've gottten more of the projects on to tox-based virtualenvs for testing
19:20:09 <mtaylor> most notably, worked with Maru on getting his patch landed to swift to add tox support
19:20:26 <jeblair> heckj: there are some fedora packages too
19:20:30 <mtaylor> which is great - because we weren't running any of swift on virtualevns before
19:22:19 <mtaylor> also, and possibly more importantly - we have a new CI team member at HP this week
19:22:30 <mtaylor> devananda has joined us
19:22:41 <LinuxJedi_cell> Yay :)
19:22:42 <mtaylor> and we'll be torturing him properly real soon now
19:22:44 * devananda waves
19:22:47 <LinuxJedi_cell> Welcome!
19:22:52 * soren tips his hat
19:23:05 <mtaylor> if things go well, I may have two more people coming on board by the design summit
19:23:11 <jeblair> hi devananda
19:23:31 <mtaylor> of course, stefano is going to kill me when I ask him for more entry tickets :)
19:27:16 <jeblair> while we're in a lull: i want to bring this SF area meetup to everyone's attention:
19:27:17 <jeblair> http://www.meetup.com/openstack/events/55240842/
19:27:34 <jeblair> wikimedia folks are particularly interested in talking jenkins and gerrit
19:27:41 <jeblair> i'm planning on attending
19:29:17 <LinuxJedi_cell> I am one minute from home then can cover my stuff
19:30:50 <LinuxJedi> ok, mtaylor, ready whenever you want
19:31:39 <LinuxJedi> <mtaylor> LinuxJedi: cool, shoot
19:31:41 <LinuxJedi> mtaylor: thanks
19:32:05 <LinuxJedi> so, yesterday I was working on making Gerrit look more like an Openstack site
19:32:24 <LinuxJedi> those who watch planet openstack have probably seen a quick preview
19:32:46 <LinuxJedi> there are a couple of hiccups I'm working though.  But we are very close to a release
19:33:05 <LinuxJedi> I'm also working on a Jenkins prototype which will hopefully be ready this week
19:33:28 <LinuxJedi> also I have been working on Stackforge which is in a betaish release state
19:34:16 <LinuxJedi> for those who don't know Stackforge is a launchpad/github/gerrit/jenkins setup much like Openstack but for projects related to openstack that don't fit in the normal umbrella
19:34:40 <LinuxJedi> we have one project starting on there at the moment and probably another by the end of this week
19:35:09 <LinuxJedi> I'm hoping to also create a homepage for it soon (I knocked up a quick logo for it Friday night)
19:35:14 <LinuxJedi> EOF
19:37:13 <jeblair> LinuxJedi: cool; any updates on the groups/db issue in gerrit 2.2.2 that was showing up on stackforge?
19:37:25 <LinuxJedi> oh, good point, so...:
19:37:53 <LinuxJedi> Gerrit 2.2.2 is being testing on Stackforge.  This is mostly down to it being InnoDB compatible and we would really prefer our data not to go to hell
19:38:27 <LinuxJedi> But, we found out on stackforge that listing groups takes a really long time
19:39:03 <LinuxJedi> This is down to the group rendering code now iterating through all users for all groups when rendering that page just to get a couple of details (most data is thrown away)
19:39:44 <LinuxJedi> I've filed a bug with gerrit and it has been accepted as a bug, but no progress there
19:40:04 <LinuxJedi> as a workaround we can set Gerrit's caches really high and that takes the rendering time down from over a minute to about 3 seconds
19:40:14 <LinuxJedi> EOF
19:42:12 <mtaylor> yay for caches!
19:42:28 <LinuxJedi> yay for eating the entire VM's RAM with them ;)
19:42:32 <mtaylor> heh
19:42:43 <mtaylor> oh - I have one more piece of announcy stuff...
19:42:55 <mtaylor> the first cut of the jenkins-jclouds plugin is up and ready for testing
19:43:14 <soren> Ooooh!
19:43:19 <soren> How?
19:43:26 <LinuxJedi> we are full of gifts this week ;)
19:43:37 <mtaylor> https://github.com/jenkinsci/jclouds-plugin
19:43:54 <mtaylor> we don't have it installed  on any of our jenkinses yet - vijay just pused it up yesterday
19:44:07 <mtaylor> but he can apparently launch slaves on rackspace cloud at the very least
19:44:23 <mtaylor> and adrian cole is going to finish up nova support in jclouds this week
19:44:24 <soren> *drool*
19:44:47 <mtaylor> so we should be able to use it then to spin up nodes on a trunk openstack install, or trystack, or whatnot
19:44:51 <mtaylor> so - WIN
19:44:57 * mtaylor could not be more excited
19:45:16 <LinuxJedi> awesome stuff :)
19:46:02 <mtaylor> jeblair: I guess we should figure out how to start testing that
19:47:22 <jeblair> mtaylor: yes, if i work on it, i'd like to prioritize it after the devstack work -- i don't want to be caught unable to run devstack tests the way we need to.  but that should be done by the end of the week (at the latest), i think.
19:47:57 <jeblair> if someone else has time before then to start poking on it, that's cool.
19:48:01 <mtaylor> jeblair: ++
19:48:20 * mtaylor will try to carve out some time to poke at it
19:50:28 <notmyname> this is the CI meeting, right? /topic says keystone meeting (I have questions for CI)
19:50:41 <LinuxJedi> notmyname: yes it is CI
19:51:05 <LinuxJedi> notmyname: mtaylor forgot to set topic :)
19:52:05 <notmyname> I've got 2 core devs for swift that can't contribute (aren't seen as core by gerrit)
19:52:09 <notmyname> any progress on that?
19:52:39 <LinuxJedi> notmyname: which one of us is working on that?
19:52:52 * LinuxJedi fixed the only one I know of
19:53:11 <notmyname> not sure. ust trying to follow up (it may be all LP now)
19:53:23 <notmyname> I didn't know if it was really fixed or a one-time fix sort of thing
19:53:29 <notmyname> letterj and redbo
19:53:40 <notmyname> both had the issues of merged accounts
19:53:44 <LinuxJedi> notmyname: I was working on letterj and can update you there:
19:54:28 <LinuxJedi> notmyname: letterj's OpenID was broken due to an account merge.  A very wll known bug on Launchpad.  Unfortunately Launchpad team made it worse when they tried to fix it
19:54:37 <notmyname> yay
19:55:03 <LinuxJedi> notmyname: so I have temporarily made both OpenIDs work for him.  But pointed out this is a very temporary workaround.  In the mean time he is working with Launchpad to fix it
19:55:11 <LinuxJedi> notmyname: I don't know how far he has got after that
19:55:12 <notmyname> ok, thanks
19:56:06 * LinuxJedi thinks jeblair was working on redbo.  As far as I am aware that was a different problem, but I don't have logs going back that far
19:56:58 <notmyname> so he just submitted code for review (as we've been chatting), and it looks like it's working for him now
19:57:07 <LinuxJedi> awesome :)
19:57:30 <notmyname> also, one other minor issue I'd like to ping status on
19:57:45 <LinuxJedi> notmyname: shoot
19:58:14 <LinuxJedi> oh, for those who want to know about the Launchpad account merge bug, it is at: https://bugs.launchpad.net/bugs/881019
19:58:15 <uvirtbot> Launchpad bug 881019 in summit "Lp login is broken after account merge" [Critical,Confirmed]
19:58:17 <notmyname> getting commit messages to be simply "+1", "+2", etc instead of the "I'd prefer you didn't submit this..."
19:58:52 <notmyname> a minor issue (but it's my pet issue), so I wanted to see if there has been any movement on it :-)
19:59:01 <LinuxJedi> notmyname: is there a bug open for it?
19:59:05 <notmyname> no idea
19:59:13 <notmyname> mtaylor: ^ ?
19:59:27 <LinuxJedi> notmyname: we have goldfish brains so we need ways of tracking ;)
19:59:49 <notmyname> I was talking to mtaylor about it, so I'd guess there is one somewhere
19:59:58 <notmyname> he's generally good about that sort of thing
20:00:12 <jeblair> we have 87 open bugs.  :/
20:00:15 <LinuxJedi> ah, probably
20:00:31 <notmyname> jeblair: I know the feeling. just ignore them and they go away (right?)
20:00:33 <LinuxJedi> jeblair: yay, new achievement unlocked!
20:00:41 <jeblair> hehe
20:01:11 <LinuxJedi> jeblair: to be fair, several are actually for other companies/teams/projects to fix
20:01:26 <LinuxJedi> such as the LP bug and Gerrit 2.2.2 bug
20:01:59 <LinuxJedi> anyway, this meeting is overrunning
20:02:07 <LinuxJedi> mtaylor: time to close up?
20:02:10 <mtaylor> yeah - I think so
20:02:17 <mtaylor> sorry, my goldfish brain got distracted
20:02:21 <mtaylor> Thanks guys!
20:02:23 <mtaylor> #endmeeting