19:02:17 <mtaylor> #startmeeting
19:02:18 <openstack> Meeting started Tue Jun 12 19:02:17 2012 UTC.  The chair is mtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:02:24 <mtaylor> anybody want to talk about CI?
19:02:35 <clarkb> maybe
19:03:00 <LinuxJedi> you mean that TV show about DNA that plays The Who in the intro?
19:04:39 <mtaylor> yup
19:04:42 <jeblair> Tonight on CI: OpenStack -- the brutal killing of an incubated project.
19:04:47 <mtaylor> hahaha
19:04:53 <LinuxJedi> lol! :)
19:04:56 <mtaylor> first up - I'm deleting all of the melange jobs
19:05:01 <mtaylor> anybody have any problems with that?
19:05:02 <mtaylor> good.
19:05:04 <mtaylor> next?
19:05:10 <mtaylor> clarkb: anything fun to talk about?
19:05:22 <clarkb> sure
19:05:51 <clarkb> I think I am happy with the etherpad lite puppet module now. I need to make it puppet masterable then use the host LinuxJedi gave me
19:06:07 <mtaylor> sweet
19:06:09 <clarkb> so the testing of migrated the DBs should happen RSN
19:06:17 * LinuxJedi upgraded that to 12.04 for clarkb
19:06:22 <mtaylor> excellent
19:06:34 <mtaylor> I'm excited about our new etherpad lite overlords
19:06:34 <LinuxJedi> clarkb: you have SSH access to that now?
19:06:46 <clarkb> LinuxJedi: in theory I do, I haven't tested it yet
19:06:57 <clarkb> jeblair merged a puppet change that added my public key to the host
19:06:58 <LinuxJedi> clarkb: ok, theory is good
19:07:15 <mtaylor> clarkb: so we might have migrated data to test next week some time?
19:07:20 <mtaylor> or this week even (it's tuesday)
19:07:31 <clarkb> mtaylor: hopefully by the end of this week
19:07:34 <mtaylor> baller
19:07:48 <clarkb> I will need to borrow LinuxJedi at some point as he has access to the old stuff
19:08:13 <mtaylor> I betcha that can be arranged
19:08:19 <LinuxJedi> clarkb: in theory... I tried to SSH into it the other day and failed, need to make sure I really do have access...
19:09:00 <mtaylor> LinuxJedi: speaking of machines and access ... it looks like there are still some things using the default tenant in my mordred@inaugust.com and/or my monty.taylor@hp.com accounts on hpcloud
19:09:07 <LinuxJedi> clarkb: worst case I should have shell type access
19:09:14 <mtaylor> jeblair: ^^ devstack is using one of them I think
19:09:20 <clarkb> LinuxJedi: that should work
19:09:36 <jeblair> mtaylor: don't think so.
19:09:47 <LinuxJedi> mtaylor: should only be your servers left there now
19:09:55 <LinuxJedi> mtaylor: I removed all of Stackforge.
19:10:17 <clarkb> I have also spent some time fixing bugs in the important changes and "status:reviewable" patch to gerrit now that it is live
19:10:31 <clarkb> all UI related so nothing horribly broken, but trying to make it useful
19:11:00 <clarkb> Shrews: were you still planning to update review-dev sometime soon?
19:11:00 <mtaylor> all those and Shrews changes have been merged in now I think
19:11:16 <Shrews> clarkb: just submitted the change  :)
19:11:21 <clarkb> woot
19:11:27 <Shrews> https://review.openstack.org/#/c/8450/
19:11:50 <clarkb> the other major time suck over the last week has been standardizing project in tree documentation
19:12:19 <clarkb> the idea is that we can use a templated jenkins job to build in tree sphinx documentation then copy it to docs.openstack.org
19:12:56 <mtaylor> yeah. that's all going to be great
19:13:05 <clarkb> currently many projects build documentation in their own special ways so I have 8 changes in gerrit right now to make them the same (though similar is probably a better description)
19:13:19 <jeblair> that makes me happy
19:13:19 <mtaylor> yeah. _same_ is ...
19:13:25 <mtaylor> ++
19:13:44 <mtaylor> I also like uploading them to docs.openstack.org - makes more sense than $project.openstack.org
19:13:59 <mtaylor> also - I think that means that the wiki server will be down to only being a wiki server
19:14:05 <clarkb> I have also updated jenkins jobs to do that, but haven't been able to test it much beyond spitting out XML and glancing at it
19:14:31 <clarkb> so I may need to fire up a jenkins instance to test that
19:14:36 <annegentle> Yay same.
19:14:36 <mtaylor> I _was_ thinking we should wait until the docs changes land...
19:14:44 <clarkb> mtaylor: yes we should
19:14:47 <mtaylor> but then I realied, screw it - they're not related
19:14:52 <annegentle> did we need to shop the idea with PTLs at all?
19:14:58 <annegentle> or did you already?
19:14:59 <mtaylor> it's not going to break any more than it already is
19:15:01 <mtaylor> annegentle: nope
19:15:05 <annegentle> mtaylor: good point, that
19:15:09 <mtaylor> annegentle: we just submitted the changes
19:15:23 <clarkb> I have seen any screaming. I think I got one "ok..."
19:15:25 <mtaylor> nobody has complained yet - I don't think it's going to bug anyone
19:15:25 <annegentle> mtaylor: ok, normal review process should be sufficient
19:15:30 <mtaylor> yeah
19:15:35 <mtaylor> heckj seemed bemued
19:15:37 <mtaylor> bemused
19:15:39 <clarkb> s/have/haven't
19:15:58 <LinuxJedi> clarkb: ok, my todo this week.  Find out who owns the etherpad server because it isn't on any of our cloud accounts either...
19:16:05 <LinuxJedi> mtaylor: who owned the old eavesdrop?
19:16:32 <mtaylor> nobody
19:16:55 <LinuxJedi> mtaylor: o..k.. how did we get access to it?
19:16:55 <clarkb> so in addition to finishing those things up I will also be attempting to fix the github pull request closing script
19:17:03 <mtaylor> LinuxJedi: ant or rick clark are the usual people to ping
19:17:19 <LinuxJedi> mtaylor: cool, SSH is on the same port so I'm assuming same ownership
19:17:24 <mtaylor> LinuxJedi: antonym
19:17:28 <mtaylor> is the irc nick
19:17:33 <LinuxJedi> mtaylor: thanks
19:17:57 <mtaylor> awesome. that's a bunch of good stuff
19:18:41 <mtaylor> Shrews: updates on gerrit-y things?
19:18:51 <Shrews> mtaylor: a bit
19:19:04 <Shrews> fixed a WIP issue where it was possible to WIP a draft
19:19:09 <Shrews> thus causing "real bad stuff"
19:19:25 <Shrews> and...
19:19:35 <Shrews> i have the python version of gerrit running
19:19:47 <Shrews> well, "running"  (with quotes)
19:20:12 <mtaylor> hehe
19:20:17 * LinuxJedi assigns Shrews the task of re-writing it in Ruby
19:20:31 <clarkb> LinuxJedi: I think that propoganda got to you
19:20:33 * Shrews assigns a big knife to LinuxJedi's neck
19:20:33 * jeblair assigns Shrews the task of re-writing it in go
19:20:44 <mtaylor> for the record - pygerrit uses protocol buffers and protobuf rpc
19:20:46 <LinuxJedi> lol :)
19:20:53 <LinuxJedi> woot!
19:21:07 <LinuxJedi> mtaylor: even more WTF? about ditching it then
19:21:18 <mtaylor> right?
19:21:42 <mtaylor> well- they do seem to have used that partially to talk back and forth between the portions written in java and the portions written in python
19:21:49 <mtaylor> so "pygerrit" might be a bit of a misnomer
19:21:51 <jeblair> ugh
19:22:04 <mtaylor> it's a django app with the git repo stuff done in jgit
19:22:13 <Shrews> it's pyjagerrit
19:22:17 <mtaylor> hahaha
19:22:23 <LinuxJedi> Shrews: bless you
19:22:44 <mtaylor> ok. moving on ...
19:22:52 <mtaylor> LinuxJedi: did you do anything this last week?
19:23:04 <LinuxJedi> mtaylor: I got the puppet master up and running
19:23:09 <mtaylor> ++
19:23:22 <LinuxJedi> we have eavesdrop, planet and paste running from it
19:23:38 <LinuxJedi> and found an issue with file serving which I think we are ready for me to implement a solution for
19:24:21 <LinuxJedi> and I did some minor things that I forgot about
19:24:55 <mtaylor> sweet
19:25:03 <mtaylor> I'm excited about our new puppetmaster overlords
19:25:22 <LinuxJedi> when complete we can call CI the muppet show
19:25:25 <Shrews> how many overlords do we have now?
19:25:38 <mtaylor> many many many
19:25:48 <LinuxJedi> Shrews: we are slaves to the machines
19:26:13 <mtaylor> jeblair: fun in the land of zuul and backups?
19:26:29 <jeblair> well, that's the future.  :)
19:26:55 <jeblair> i've identified a couple of bugs in zuul, but by and large, it has mostly worked.
19:27:10 <jeblair> what has failed should be easy to fix.
19:27:14 <jeblair> i wrote gobs of docs for it
19:27:20 <jeblair> http://ci.openstack.org/zuul/
19:27:30 <mtaylor> I actually just sent someone a link to that, actually
19:27:44 <jeblair> based on clarkb's settings, so we should be able to switch that to use the new job template like the rest of the projects
19:27:49 <mtaylor> oh - speaking of - I got a feedback on one of the lines in the docs:
19:28:08 <LinuxJedi> jeblair: we should totally find a way to make CI docs home link to that
19:28:15 <mtaylor> jeblair: "Zuul queues those changes in the order they were approved, and notes that each subsequent change depends on the one ahead of it merging"
19:28:33 <mtaylor> jeblair: confused the reader and caused him to think it was talking about actual git-level dependent changes
19:29:03 <mtaylor> jeblair: rather than the dependencies created by queue position
19:29:04 <jeblair> i've been looking into server based and swift-based backups.  i'm leaning toward server-based due to the complexity around key and account (ie, cloud account) credentials.
19:29:12 <mtaylor> ++ server-based
19:29:37 <jeblair> mtaylor: okay.  i will take a look at that.  i vomited out the docs in one long session and have hardly even read them.  :)
19:29:55 <jeblair> LinuxJedi: good idea
19:30:02 <mtaylor> jeblair: :)
19:30:07 <mtaylor> jeblair: I'm impressed that they exist
19:30:29 <jeblair> as far as the past -- i switched like all the jenkins slaves to precise
19:30:43 <jeblair> (except we need to keep some oneiric around to run py26)
19:30:59 <mtaylor> excited about that
19:31:03 <LinuxJedi> ah, I was going to write a todo for me to do that in Stackforge
19:31:08 <mtaylor> still haven't observed jclouds bursting yet
19:31:30 <jeblair> and i've been working on figuring out why tempest isn't happy; it turns out at least one problem is that one of our devstack node providers is giving us corrupted images.  no points for guessing which.
19:31:40 <mtaylor> ec2?
19:31:59 <jeblair> mtaylor: maybe take some precise slaves offline?  i did make 8 of them.
19:32:03 <jeblair> mtaylor: close.
19:32:05 <LinuxJedi> Azure?
19:32:46 <jeblair> that's it in broad strokes.
19:33:34 * jeblair hands the speaking stone back to mtaylor
19:33:36 <mtaylor> jeblair: point. I'll do that
19:33:45 <mtaylor> so- I've been poking at a bunch of random things
19:33:56 <mtaylor> tox.ini alignment, got coverage jobs pretty much all green
19:34:00 * jeblair looks worried
19:34:12 <mtaylor> ttx and I are going to talk to the ppb next meeting about client lib release plan
19:34:15 <clarkb> jeblair: cronspam? I wonder if it is related to pypi being slow
19:35:03 <mtaylor> and I'm going to try to get an answer from them too on the global dep list versions that are in conflict
19:35:20 <mtaylor> and I wrote a little code to do tag-based post-release versioning for client libs
19:35:33 <mtaylor> and I tried not to break too many things
19:35:47 <mtaylor> oh - didn't I do the new pypi mirror last week?
19:35:58 <jeblair> yep
19:36:08 <mtaylor> so, wrote a new pypi mirror from scratch :)
19:36:28 <LinuxJedi> yay, fixing what the broken one did was fun :)
19:36:32 <clarkb> decided the existing wheels were too square?
19:36:35 <mtaylor> it grabs all of our projects, pip install downloads all of their depends (and depends of depends) to populate the PIP_DOWNLOAD_CACHE
19:36:50 <mtaylor> and then turns the files in the pip download cache into a pypi style index
19:37:15 <LinuxJedi> clarkb: the existing wheels broke all our jobs overnight ;)
19:37:22 <heckj> o/ (mtaylor: bemused is a good word - fine by me, don't mind being consistent)
19:38:05 <mtaylor> yay!
19:38:10 <mtaylor> we'll consider that a binding opinion
19:38:50 <mtaylor> I think that's about all I've got
19:38:57 <mtaylor> anything from anybody else?
19:39:08 <LinuxJedi> mtaylor: we will be able to do the meeting next week?
19:39:20 <LinuxJedi> (we as in me and you)
19:39:58 <mtaylor> LinuxJedi: good question
19:40:06 <mtaylor> we'll try - but LinuxJedi and I might be offline during meeting time
19:40:28 <mtaylor> even so - I'm sure that jeblair and clarkb and Shrews can soldier on without us :)
19:40:51 <clarkb> but who will Shrews stab if you guys aren't here
19:40:55 <Shrews> mtaylor: you're still here?  heh
19:41:07 * Shrews eyes clarkb
19:41:11 <LinuxJedi> lol :)
19:41:11 <jeblair> we will get so much done... and violence free.
19:41:16 <jeblair> VIOLENCE FREE, SHREWS!
19:41:25 <LinuxJedi> jeblair: nothing will break ;)
19:41:28 <Shrews> pffft
19:41:54 <jeblair> LinuxJedi: was that some kind of curse?   thanks.  :)
19:42:17 <LinuxJedi> clarkb: my availability will be real patchy next week so make sure you get what you need out of me this one :)
19:42:44 <clarkb> just access to the old etherpad db at some point this week
19:42:54 <LinuxJedi> will aim for that at least
19:45:45 <mtaylor> k. well, I think that's about it
19:47:58 <mtaylor> unless anybody else has anything...
19:48:30 <Shrews> mtaylor, jeblair: when can we upgrade gerrit?
19:49:10 <jeblair> asap i think.  if the change looks good on gerrit-dev, propose it for prod, and we can approve it when things seem quietish.
19:49:30 <Shrews> k
19:50:21 <mtaylor> work. k. thanks everybody
19:50:24 <mtaylor> #endmeeting