19:02:17 <jeblair> #startmeeting infra
19:02:17 <openstack> Meeting started Tue Feb 26 19:02:17 2013 UTC.  The chair is jeblair. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:02:20 <openstack> The meeting name has been set to 'infra'
19:02:57 <jeblair> #topic previous meeting
19:03:01 <jeblair> #link http://eavesdrop.openstack.org/meetings/infra/2013/infra.2013-02-19-19.02.html
19:03:29 <fungi> disabling the old wiki...
19:03:35 <pleia2> shiny new wiki
19:03:40 <fungi> notified the openstack and openstack-dev lists
19:03:48 <fungi> #link https://lists.launchpad.net/openstack/msg21102.html
19:03:56 <fungi> #link http://lists.openstack.org/pipermail/openstack-dev/2013-February/005993.html
19:03:57 <jeblair> fungi: cool, any change to the plan?
19:04:19 <fungi> jeblair: not as of yet. i'll comment those aliases out on friday as it stands now
19:04:24 <jeblair> #action fungi disable old-wiki sometime after mar 1
19:04:35 <fungi> or monday. whenever
19:04:44 <jeblair> how bout the quantal images?
19:05:16 <fungi> i finally decided formal contact was probably in order, and opened a ticket under our openstackjenkins account
19:05:39 <fungi> they've since confirmed something's not right with their image metadata, but not fixed it yet
19:05:53 <fungi> at least not as of a couple hours ago when i tested again
19:06:03 <jeblair> cool.  progress!
19:06:13 <fungi> seemingly so, anyway
19:06:30 <fungi> if/when that works, i'll propose changes to all the projects
19:06:45 <mordred> o/
19:07:18 <jeblair> clarkb, mordred: we have good ram limits for hpcloud now...
19:07:30 <jeblair> but i think az3 is still broken for us
19:07:35 <mordred> yay
19:07:37 <mordred> boo
19:07:38 <jeblair> https://jenkins.openstack.org/job/devstack-update-vm-image-hpcloud-az3/13/console
19:08:04 <jeblair> clarkb: would it be a good idea for mordred to yell at more people?
19:08:17 <mordred> ugh. do I need to yell at MORE people?
19:08:18 <fungi> or maybe buy him a megaphone
19:09:07 <jeblair> #action mordred yell at hpcloud to fix the openstackjenkins account (az3 doesn't work)
19:09:08 <clarkb> jeblair: maybe. The support folks claim it is enabled for us
19:09:20 <clarkb> I am guessing that the account is special enough that some extra bit needs flipping
19:09:42 <jeblair> great.  :)
19:10:11 <jeblair> #topic New CLA
19:10:25 <jeblair> so the exciting news is what we _didn't_ do this weekend!
19:10:28 <fungi> that didn't happen... *sigh*
19:10:29 <mordred> yay!
19:10:31 <mordred> wait
19:10:32 <fungi> #link http://lists.openstack.org/pipermail/openstack-dev/2013-February/005960.html
19:10:48 <jeblair> so the suggested change has been merged
19:11:02 <fungi> #link https://review.openstack.org/22910
19:11:24 <fungi> was that the only pending alteration? if so should i send a new announcement to the dev ml pointing people at it?
19:11:40 <jeblair> i chatted with the foundation staff this morning, and they figure that if we send out an announcement today, that should be enough review time (since it's a relatively minor change)
19:11:45 <fungi> and do we want to pick a new date yet, or is that premature?
19:12:01 <jeblair> so i think we're ready to pick a new date at our convenience, then send an announcement
19:12:08 <pleia2> great
19:12:19 <fungi> works for me. when's convenient for everyone?
19:13:05 <fungi> the previous date was chosen in part based on input from ttx, to accommodate the development cycle
19:13:11 <jeblair> Sunday (Mar 3) works for me
19:13:30 <clarkb> I can probably spend a few hours on the third
19:13:30 <jeblair> i imagine earlier is better than later as far as that goes
19:13:35 <clarkb> (I can't do saturday)
19:13:36 <mordred> ++
19:13:37 <pleia2> Mar 3 would be great for me too
19:13:47 <mordred> I'm free the 3rd
19:14:16 <fungi> okay, let's tentatively say the 3rd. i need to check with my meatspace event coordinator to confirm my calendar
19:14:38 <fungi> assuming it checks out, i'll send the notification and announcement this afternoon
19:14:45 <pleia2> thanks fungi
19:14:45 <jeblair> fungi: great
19:14:54 <jeblair> #action fungi announce new CLA cutover
19:15:33 <fungi> the plan is all set, for anyone who wants to look it over and didn't get a chance last weekend
19:15:48 <fungi> #link https://etherpad.openstack.org/cla-maintenance-2012
19:16:07 <pleia2> oh drat, I actually have a class until noon pacific
19:16:14 <pleia2> but I'll be around all afternoon :)
19:16:32 <fungi> pleia2: you can take the "re-re-reexplain to people who didn't read" shift ;)
19:16:39 <pleia2> right!
19:16:44 <jeblair> oh, anyone want to move it one hour earlier?
19:16:45 <anteaya> I will probably be onto other things by the 3rd but I can make some time to help out
19:17:05 <fungi> i'm cool with 17:00 utc if everyone else is
19:17:22 <fungi> i'll be onto my second pot of coffee by then regardless
19:17:48 <jeblair> i think this time we found out we're pretty much staged and ready to go, so we don't need any prep time
19:18:17 <jeblair> mordred, clarkb: 1700 (9am pacific) work for you, or would you prefer 10?
19:18:24 <clarkb> 10 is better but 9 works
19:18:25 <fungi> yeah, as long as everyone who wants to +1/+2 the changes goes ahead, i'll leave my -2 in place until 1700
19:18:49 <fungi> or 1800. whatever. you lot decide
19:19:33 <jeblair> #startvote When should we start the CLA cutover? 1700, 1800
19:19:34 <openstack> Begin voting on: When should we start the CLA cutover? Valid vote options are 1700, 1800.
19:19:35 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
19:19:46 <jeblair> #vote 1700
19:19:55 * fungi abstains
19:20:17 <mordred> either is fine with me ...
19:20:43 <fungi> clarkb vs jeblair... cage match to decide the outcome
19:21:05 <jeblair> "abtains" and "either is fine with me" were not options.  :)
19:21:12 <fungi> ha
19:21:17 <mordred> fine...
19:21:19 <mordred> #vote 1700
19:21:27 <fungi> #vote 1700
19:21:51 <clarkb> #vote 1700
19:21:53 <clarkb> I will play along
19:22:02 <jeblair> #endvote
19:22:03 <openstack> Voted on "When should we start the CLA cutover?" Results are
19:22:04 <openstack> 1700 (4): mordred, jeblair, clarkb, fungi
19:22:43 <fungi> honestly, i'll cut and paste the stuff from that etherpad and the rest of you can just pitch in once i break stuff
19:23:03 <pleia2> :)
19:23:04 <fungi> shouldn't be difficult nor time consuming
19:23:09 <jeblair> next on the agenda is 'jenkins slave operating systems' which i think we more or less covered...
19:23:28 <jeblair> #topic grenade
19:23:46 <jeblair> dean troyer has merged some changes to bring the grenade running code up to date...
19:24:01 <jeblair> so i think it's on us (me, probably) to add a test job that actually runs it...
19:25:05 <jeblair> #topic gearman
19:25:34 <jeblair> zaro: are you around?
19:27:02 <jeblair> zaro: let us know when you're back
19:27:06 <jeblair> #topic reviewday
19:27:24 <jeblair> pleia2: how's that going?
19:27:32 <pleia2> it's ongoing
19:27:43 <pleia2> fungi did a review this morning which I hope will get me to the end of things very soon
19:27:56 <pleia2> we do need to set up a gerrit user for it once the puppet configs are done though
19:27:57 <zaro> sorry, i'm back
19:28:16 <fungi> #link https://review.openstack.org/21158
19:28:19 <fungi> pleia2: that?
19:28:23 <jeblair> pleia2: ok, that shouldn't be too hard.  does anything about reviewday or the repo need to change?
19:28:42 <pleia2> fungi: thanks
19:29:05 <pleia2> jeblair: I've already patched reviewday itself for everything we needed (most notably command line option for output directory)
19:29:16 <pleia2> so we should be good
19:29:34 <jeblair> pleia2: groovy
19:30:09 <jeblair> #topic gearman
19:30:34 <jeblair> zaro: so things are coming along with the gearman plugin
19:30:40 <zaro> yes, last checkin exposed a bug in jenkins.
19:30:52 <zaro> due to switch to using maven3
19:31:10 <zaro> found the problem.  clarkb will help me fix it after meeting.
19:31:23 <zaro> but things are in good shape.
19:31:33 <jeblair> awesome!  what's the short version of the problem?
19:32:22 <zaro> Jenkins fails build because one of it's core components attempts to download hpi-plugin from a bad location.
19:32:42 <zaro> workaround is to override the location in maven settings.xml
19:33:33 <jeblair> zaro: you mentioned that we still need to get queue canceling added to gearman.  have you talked to brian about it?
19:33:34 <zaro> gearman plugin is going well.  only 2 more items to do
19:33:53 <jeblair> mordred could probably lobby for us too.
19:34:14 <zaro> yes, cancel jobs from gearman queue and updating gearman functions as nodes come online/offline.
19:34:37 <zaro> that would be swell
19:34:43 <jeblair> mordred: thoughts?
19:35:10 <mordred> jeblair: I'll ping him about it
19:35:11 <jeblair> zaro: i think there are already methods to de-register functions, right?
19:35:34 <zaro> i've talked to brian about 2 weeks ago but have not gotten to check with hime again.
19:35:42 <mordred> I think he said someone had written a dequeue function
19:35:49 <zaro> yes, there are methods to de-register.
19:36:08 <jeblair> zaro: oh, i see, that was one of your '2 more items to do'.
19:36:22 <jeblair> zaro: i thought that was something that needed adding to gearman itself.
19:36:53 <mordred> and we're sure we need gearman queue cancelling? I thought we'd come up with a workaround for that...
19:37:01 <zaro> jeblair: no. we can unregister. what we need is the dequeue
19:37:05 <jeblair> there's a workaround but it's crazy silly.
19:37:15 <jeblair> and likely has race conditions..
19:38:16 <zaro> brian told me he would check on why there's no dequeue.  but he did say that it's probably not there for a good reason
19:38:24 <zaro> but he couldn't think of the good reason.
19:38:25 <mordred> ok. do we have a writeup anywhere of the why? last time I tried to tell brian we needed it he asked me about like, 12 different gearman patterns
19:39:08 <zaro> why we need it?
19:39:11 <jeblair> mordred: i know i've written about it in some emails to zaro
19:39:18 <mordred> k
19:39:44 <mordred> I mean, I grok in general - but I find it's good to distil needs into very specific nuggets - brian is a busy guy
19:39:45 <zaro> i gave him my use case and he said it made sense.
19:39:50 <mordred> k. cool
19:40:01 <mordred> I guess it helps that zaro is in the office with him
19:40:18 <jeblair> mordred, zaro: let me know if i can help either by writing more emails or a phone call or something.
19:40:18 <mordred> in any case, I think he said that someone made a gearman extension to do it at some point...
19:40:23 <mordred> cool
19:40:54 <zaro> is there a place to search for gearman extensions?  i must have missed it in the docs.
19:41:37 <jeblair> (i didn't know there were extensions)
19:41:57 <mordred> of course there are extensions. what isn't pluggable these days?
19:42:00 <mordred> and I don't know...
19:42:25 <zaro> ok.  i'll take an action to ressearch.
19:42:33 <jeblair> https://www.google.com/search?q=site%3Agearman.org+extensions
19:42:34 <jeblair> :(
19:43:20 <jeblair> #topic Registering IANA PEN
19:44:13 <jeblair> i've started an email thread with heckj and some others about their need for OIDs
19:44:52 <jeblair> if they do need them, i can register the PEN, and we can adopt a process like debian uses for internal allocation underneath it
19:44:54 <jeblair> #link http://dsa.debian.org/iana/
19:45:17 <jeblair> but i'm not sure it's needed at this point.  i'll let you know how it turns out
19:46:39 <jeblair> #topic open discussion
19:47:03 <heckj> jeblair: howdy
19:47:11 <heckj> I don't think we need it now
19:47:43 <jeblair> heckj: ok.  btw, i believe SNMP and LDAP share PEN allocations
19:47:48 <heckj> just sent you, lauren, and topol an email related - would prefer to discuss at the summit what we should be doing - I just don't know the specifics, but it seemed to heavily overlap with SNMP OIDs
19:47:53 <heckj> Ahhh!!!
19:48:15 <heckj> Well, regardless, I'd like to maybe have that discussion a little more broadly, so if it's not a burning issue I'd rather discuss and then do :-)
19:48:34 <heckj> jeblair: that work for you?
19:48:38 <anteaya> o/
19:48:46 <topol> heckj +1
19:48:49 <jeblair> heckj: absolutely.  :)
19:49:28 <fungi> before i forget, there was a suggestion last week that maybe we should subsume chmouel's openstackwatch.rss into jeepyb, in case anyone's looking for a fun integration project
19:49:31 <fungi> #link https://github.com/chmouel/openstackwatch.rss
19:49:48 <mordred> oh yeah. I forgot about that
19:50:00 <jeblair> heckj, topol: i think the main issues are (a) do we need one (is it appropriate for any activity we're doing), and (b) how to manage sub-allocations internally
19:50:09 <anteaya> fungi, if it doesn't require me to set up an environment, I can work on that
19:50:31 <heckj> jeblair: yeah - that's what I came into as well. I'm guessing that as soon as we open that door, it'll be wanted for lots of pieces
19:50:45 <pleia2> clarkb: you mentioned ssl certificates the other day, are we monitoring thoes expiration dates with any tools? should we? (I use nagios personally, but there are tons of tools/scripts out there)
19:50:56 <topol> jeblair, Makes sense!
19:51:00 <fungi> anteaya: i believe he's been running it by connecting with his personal gerrit account, so probably easy to test with no special access
19:51:36 <anteaya> fungi, I will take a look, is there any bug report or blueprint for me to track my progress?
19:51:38 <jeblair> heckj: you want to propose a summit session?
19:51:45 <clarkb> pleia2: I don't think we are monitoring them with any special tools. I just opened the 4 or so we had in a browser
19:51:46 <jeblair> topol: or you?
19:51:50 <fungi> anteaya: nah, it was an informal discussion
19:51:55 <anteaya> fungi, very good
19:52:06 <clarkb> I am working on making the jclouds jenkins plugin useable as a replacement for the scp plugin
19:52:23 <jeblair> clarkb: pleia2 we could use a cron certwatch script, which is a little lighter weight than nagios
19:52:27 * anteaya has a question when it is her turn
19:52:41 <heckj> jeblair: woll do
19:52:42 <pleia2> jeblair: shall I toss this in a bug report?
19:52:44 <heckj> er, will do
19:53:25 <jeblair> clarkb, pleia2: i'm also okay with nagios if someone wants to set that up.  though i don't think it should page anyone.  :)
19:53:31 <jeblair> heckj: awesome, thanks
19:53:34 <jeblair> anteaya: your turn!
19:53:38 <anteaya> thanks
19:53:46 <jeblair> pleia2: yes please
19:53:47 <pleia2> nagios is a bit much for just this one check, so unless we have other things to monitor...
19:54:01 <anteaya> I come from a ruby and rails background, yesterday I saw the puppet-dashboard module for the first time
19:54:28 <anteaya> it looked like it could tolerate a little tidying, just wondering if anyone had any thoughts
19:54:32 <jeblair> mordred: ^
19:54:37 <zaro> jeblair, mordred : could i get your feedback on my comment to issue https://bugs.launchpad.net/openstack-ci/+bug/1082802  ?
19:54:39 <uvirtbot> Launchpad bug 1082802 in openstack-ci "Write gearman job server jenkins plugin" [Medium,Triaged]
19:54:42 <mordred> I agree. it could tolerate tidying
19:55:05 <anteaya> mordred, very good, I will give it some focus and see what I come up with
19:55:19 <mordred> zaro: I will comment - the job-server plugin I was suggesting is a much more invasive and crazy thing
19:55:59 <jeblair> mordred: are we still running a fork of the dashboard?
19:56:03 <mordred> zaro: tl;dr - replace jenkins' internal job queue implementation with gearman
19:56:17 <mordred> jeblair: no, we're running a fork of the dashboard puppet module
19:56:41 <jeblair> mordred: why is that again?  (sorry)
19:56:56 <mordred> jeblair: because the upstream one was broken in some manner. I honestly don't remember
19:57:18 <jeblair> mordred: so perhaps anteaya could think about upstreaming whatever it was you fixed too?
19:57:35 <mordred> also, I believe upstream trunk was also in an unusable state - so although I think I submitted a pull request, I don't think that helped
19:57:48 <jeblair> oh, ok.
19:57:50 <mordred> also, I believe dashboard is quickly becoming abandonware, if it isn't already
19:58:05 <jeblair> i thought i heard some people stepped up to keep it going?
19:58:23 <mordred> did they? neat!
19:58:23 <anteaya> let's have that discussion then, no sense me wasting time on abandonware
19:58:37 <jeblair> clarkb: am i making things up?
19:58:37 <mordred> don't listen to me - I'm apparently stypid
19:59:28 * olaph types this so meetbot will count him present...
19:59:34 <fungi> heh
19:59:35 <anteaya> so, if I understand correctly, I need to scope the current puppet-dashboard module repo and possibly the .git history?
19:59:55 <jeblair> anteaya: at any rate, you now have some more vague and incomplete information that might help with a code archeology project.
20:00:10 <pleia2> hah
20:00:16 <anteaya> vague and incomplete, the best kind of information
20:00:27 <jeblair> anteaya: but we really do rely on dashboard, so whatever you can do will be appreciated.  :)
20:00:36 * anteaya grabs her bullwhip and fedora
20:00:39 <jeblair> time's up, thanks everyone!
20:00:44 <jeblair> #endmeeting