19:00:00 <nirik> #startmeeting Infrastructure (2013-02-07)
19:00:00 <zodbot> Meeting started Thu Feb  7 19:00:00 2013 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:01 <nirik> #meetingname infrastructure
19:00:01 <zodbot> The meeting name has been set to 'infrastructure'
19:00:01 <nirik> #topic welcome to all
19:00:01 <nirik> #chair smooge skvidal CodeBlock ricky nirik abadger1999 lmacken dgilmore mdomsch threebean
19:00:01 <zodbot> Current chairs: CodeBlock abadger1999 dgilmore lmacken mdomsch nirik ricky skvidal smooge threebean
19:00:07 * skvidal is here
19:00:08 <nirik> welcome everyone.
19:00:09 * puiterwijk 
19:00:13 * mattdm is here
19:00:15 * pingou is here
19:00:22 * maayke is here
19:00:31 * relrod is here
19:00:42 * threebean is here
19:00:55 * abadger1999 here
19:01:01 * nirik will wait a min more for folks to wander in. ;)
19:01:05 <ausmarton> hi
19:01:11 * herlo is here
19:01:15 * mdomsch 
19:01:44 * lmacken 
19:01:47 <nirik> ok, lets go ahead and get started. ;)
19:01:50 <nirik> #topic New folks introductions and Apprentice tasks.
19:02:01 <nirik> any new folks like to say hi? or apprentices have questions or commends?
19:02:03 <nirik> comments even
19:02:08 <cmullead> !
19:02:31 <nirik> cmullead: go ahead...
19:03:08 <cmullead> I am new to the group.  I have been a user/admin of Red Hat/Fedora for many years
19:03:25 <cmullead> and I thought it was time to give something back
19:03:29 <nirik> welcome.
19:03:29 <cmullead> eof
19:03:42 <nirik> are you more interested in sysadmin type tasks? or application development? or both? :)
19:03:59 <cmullead> system admin
19:04:33 <nirik> cool. ;) See me after the meeting over in #fedora-admin and I can point you in the right direction to get started. ;)
19:04:34 <housewifehacker1> This is my first meeting. Im an intern with Gnome OPW program working on datanommer with threebean
19:04:48 <cmullead> ok
19:04:49 <nirik> welcome housewifehacker1.
19:05:10 * nirik is very much looking forward to datanommer. ;)
19:05:16 <pingou> \ó/
19:05:28 <skvidal> ah ha! that answers a question I had about your nick
19:05:32 <skvidal> housewifehacker1: welcome
19:05:52 <skvidal> hmmm actually it doesn't really answer my question
19:06:01 <skvidal> but it is a not an interesting or useful question
19:06:11 <threebean> :)
19:06:21 <abadger1999> :-)
19:06:22 <nirik> ok, any other new folks? Please do let us know if we can assist or direct you in helping out any. Thanks for contributing.
19:06:46 <nirik> #topic Applications status / discussion
19:06:48 <dkanbier> this is my first meeting as well. I'm a sysadmin willing to give something back to the community. I applied to the bug zappers group but would like to help out on anything infrastructure related as well.
19:06:53 <nirik> #undo
19:06:54 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0xd333e10>
19:07:24 <nirik> dkanbier: welcome. Please do also see me after the meeting in #fedora-admin and I can point you in the right direction... ;)
19:07:35 <dkanbier> will do :)
19:07:35 <nirik> #topic Applications status / discussion
19:07:44 <nirik> ok, any application news from this week/upcoming?
19:08:04 <puiterwijk> I want to call out for testers for FAS-OpenID again ,if I may?
19:08:08 <pingou> copr-cli progressed over the week-end (even with a very interesting fosdem)
19:08:09 <nirik> puiterwijk: please do.
19:08:17 * mdomsch fixed several bugs in MM 1.4 in staging; getting to the point where I need to start running cronjobs against live data to find more bugs
19:08:25 <pingou> jenkins-master is also asking for testers
19:08:34 <puiterwijk> as I have not yet received any comments on FAS-OpenID last week, I would like to again ask for testers
19:08:52 <pingou> puiterwijk: smooth for me
19:08:53 <nirik> should we perhaps send out a devel-announce post asking for more testers? or is that too widespread?
19:08:54 <puiterwijk> #info OpenID ID for FAS-OpenID: http://<username>.id.stg.fedoraproject.org/
19:09:00 <lmacken> pingou: I'd like to migrate bodhi & kitch over at some point
19:09:04 <lmacken> **kitchen
19:09:13 <pingou> lmacken: sounds great
19:09:20 <puiterwijk> nirik: I think that's a bit too wide
19:09:23 <skvidal> pingou: have you or slavek tried out copr-fe against it?
19:09:31 <lmacken> pingou: is there an easy way to migrate jobs?
19:09:32 <skvidal> or a test of copr-fe
19:09:37 <pingou> skvidal: I didn't :]
19:09:45 <nirik> puiterwijk: ok, how about infra list?
19:09:53 <puiterwijk> nirik: that would be a good one, I think
19:09:54 <pingou> skvidal: just against local (and not for build)
19:09:59 <skvidal> nod
19:10:14 <pingou> lmacken: I copy/pasted from jenkins to cloud when I did it for fedocal/f-r
19:10:33 <lmacken> pingou: cool
19:10:55 <pingou> lmacken: the only thing to take care of is 1 build = 1 python
19:10:58 <pingou> builder*
19:11:09 <puiterwijk> nirik: do you want me to send it after the meeting?
19:11:14 <nirik> puiterwijk: please do.
19:11:16 <puiterwijk> (or do you want to do yourself?)
19:11:24 <puiterwijk> sure, will do
19:11:53 <nirik> pingou: perhaps a jenkins post too... ?
19:12:26 <pingou> nirik: I wanted to write down a wiki page on what's our idea of use-case for jenkins
19:12:37 <pingou> but yeah after a general announcement is a good idea
19:12:41 <nirik> thats also a good idea. agreed.
19:13:18 <threebean> so, housewifehacker1 has put a ton of improvements into datanommer and we're looking to upgrade the instance in staging soon
19:13:24 <nirik> mdomsch: I think early next week I might try and split out mm mirrorlists to ther own servers. Or should I wait on that until 1.4 is in prod?
19:13:27 <threebean> here's our todo list for it https://github.com/fedora-infra/datanommer/issues/27
19:13:34 <pingou> cool threebean housewifehacker1
19:13:47 <mdomsch> nirik: we can split them out first
19:14:04 <nirik> mdomsch: ok. I would think it would be pretty easy...
19:14:18 <mdomsch> should be
19:14:22 <threebean> if I'm not around and she happens to need help with various fi-apprentice things, any extra help or advice would be appreciated.
19:14:39 <nirik> My door/irc client is always open. ;)
19:14:47 <abadger1999> nirik: devel would be fine for fa-openid mail I think, devel-announce is supposed to be even lower volume, though.
19:15:20 <puiterwijk> abadger1999: so you think the email should go to devel or devel-announce?
19:15:24 <nirik> sure, but this seemed like a one off announcement. ;)
19:15:27 <abadger1999> puiterwijk: devel
19:15:33 <abadger1999> not devel-announce
19:15:45 <nirik> but devel sounds fine to me too
19:15:51 <puiterwijk> ah ok. but not just infra?
19:15:52 <pingou> shouldn't wait for the design to announce it?
19:16:14 <threebean> (lastly, on datanommer.. the upshots to the latest changes are three-fold:  1) we get a smooth upgrade path for the DB with python-alembic, 2) we get an API we need to write datagrepper, and 3) we get some CLI scripts we need for nice nagios checks )
19:16:19 <abadger1999> dvel-announce is for people who are too cranky to subscribe to devel :-/
19:16:44 * SmootherFrOgZ is here now.
19:17:00 <abadger1999> pingou: fine with that too -- if we want testing now, though, I think most of hte infra people are already aware
19:17:05 <nirik> threebean / housewifehacker1: question: how is the db looking for datanommer so far? is it growing very big? should we split it in prod? or we don't know yet?
19:17:26 <nirik> pingou: on jenkins? or fas-openid?
19:17:32 <pingou> nirik: openid
19:18:06 <nirik> ok, I'm fine waiting on a design writeup on those, I agree it would be good to have. more so on jenkins...
19:18:12 <puiterwijk> I tend to agree with pingou on that, but maybe someone could help me with getting the design
19:18:15 <threebean> nirik: don't know yet -- we haven't checked.
19:18:34 <pingou> nirik: you want a fedora design for jenkins??
19:18:50 <puiterwijk> I could use help either with getting the design team a kick or with a temporary design until they are finally done
19:18:59 <nirik> oh I see. I was misunderstanding 'design' here.
19:19:10 <nirik> you are talking about theme/appearance?
19:19:12 <threebean> nirik: oh, that's not true.  A dump was 236M back at the end of january.
19:19:15 <pingou> nirik: yes
19:19:21 * ianweller is here
19:19:38 <nirik> I was talking about a wiki page with 'here's what this is, and how you can use it, and what problems it solves, etc'
19:19:52 <relrod> puiterwijk: I can probably whip up at least a temporary design for it, poke me after the meeting.
19:20:17 <puiterwijk> relrod: ok, thanks :) (design would have been ready about two weeks back, but they're "still busy"...)
19:20:29 <abadger1999> threebean: If it's postgres, point me at it and I can get numbers for you.
19:20:30 <nirik> so, if we ask for testing we can point people at that and they can see how it's used, etc
19:20:49 <abadger1999> threebean: or a df -k /var/lib/pgsql should do it if it's the only database on that server.
19:20:54 <pingou> nirik: +1, that's kind of my idea for jenkins
19:21:10 <threebean> abadger1999: it is.. I'll check here in a moment.
19:21:56 <nirik> pingou: yeah. with how to add things, what things we can't add, etc.
19:22:21 <pingou> nirik: I'll send it around for review/comments before I announce it :)
19:22:27 <nirik> sounds good.
19:22:43 <nirik> with fas-openid, it's much less a issue because thats "anywhere you want to use openid" :)
19:23:00 <nirik> #info will work out a theme/skin for fas-openid and then ask for wider testing soon.
19:23:02 <puiterwijk> <nod>
19:23:20 <nirik> #info datanommer to be updated in stg and tested by threebean / housewifehacker1
19:23:31 <threebean> abadger1999: its db01.phx2.fp.o.. so, there are other dbs there.
19:23:49 <nirik> #info jenkins will get a faq doc and then more widespread testing.
19:24:06 <nirik> Any other application news or plans?
19:24:18 <threebean> oh, I've been fighting with askbot in stg with fedmsg plans.
19:24:26 * lmacken planning a bodhi update this week. masher changes only.
19:24:31 <abadger1999> threebean: k.  I'll get it with a query
19:24:33 <threebean> sorting out fedora/epel django packaging issues for most of the week.
19:24:45 <nirik> #info askbot upgrade being prepped in stg.
19:24:49 <lmacken> threebean and I have also been fighting with a mod_wsgi+fedmsg threading issue recently
19:24:51 <nirik> threebean: thanks a lot for that. ;(
19:25:14 <nirik> #info bodhi update for prod soon with bugfixes
19:25:20 <pingou> threebean: if there is anything special it might be good to talk with Aurélien since HK is django
19:25:29 <pingou> might prevent us from having the problem later
19:26:21 <nirik> ok, moving on then...
19:26:22 * threebean nods
19:26:29 <nirik> #topic Sysadmin status / discussion
19:26:50 <nirik> So, we hopefully have some arm machines being racked today. Will need to get them up and running.
19:26:59 <pingou> arg on apps, HK has made some really nice progress last week
19:27:00 <lmacken> nice!
19:27:20 <nirik> Hopefully we can use a few of them for infra stuff too, which will be nice.
19:27:32 <SmootherFrOgZ> \o/
19:27:33 <puiterwijk> nirik: need help on the arm boxes?
19:27:36 <nirik> pingou: cool. I really want it to be ready. :)
19:27:37 <lmacken> pingou: wrt HK, we need to change the login form to https
19:27:55 <pingou> lmacken: mentionned, forgot to open a ticket for that though
19:28:04 <nirik> puiterwijk: well, we first need network and such. ;) Hopefully it will be pretty easy to get them setup
19:28:23 <puiterwijk> nirik: hehe, would be useful I guess yeah
19:28:25 <nirik> I'd like to discuss a bit a few bigger picture plans...
19:28:34 <nirik> a) splitting out apps from app servers.
19:28:50 <nirik> I'm going to start on this likely next week... with mirrorlist server.
19:29:12 <nirik> but at the end of it, it would be nice if we had no need for app boxes...
19:29:19 <skvidal> +billion
19:29:28 <nirik> does anyone have concerns or questions about that being the way to go? :)
19:29:37 <lmacken> which cloud are we going to migrate the apps to?
19:29:47 <skvidal> are we migrating to the cloud?
19:29:51 <nirik> no cloud yet, just seperate instances
19:29:54 <skvidal> I thought we were just going to split out instances
19:30:02 <lmacken> okay
19:30:12 <nirik> so, make mirrorlist01 02 03 in phx2 and move that to those, etc.
19:30:28 <abadger1999> threebean, nirik: datanommer on db01 is currently 305MB
19:30:41 <abadger1999> select pg_database_size('datanommer');
19:30:50 <nirik> Down the road / rhel7 or something we might be able to use containers for smaller apps
19:31:03 <nirik> but thats someday/sidetrack
19:31:17 <abadger1999> nirik: works for me.
19:31:44 <threebean> abadger1999: we're looking at a rate of ~1GB/year
19:31:51 <nirik> short term this means more instances, but hopefully they can be smaller and help us with debugging/prevent one thing from messing up another
19:31:59 <abadger1999> nirik: you going to just work on separating the web front ends right now?  (db's at a later date)?
19:32:07 <nirik> abadger1999: yeah.
19:32:14 <abadger1999> cool.
19:32:23 <nirik> I still hope someday we can figure out a better way to replicate/deal with dbs
19:32:28 <lmacken> do we still plan on sticking with i686 vms?
19:32:38 <puiterwijk> nirik: I have a replication setup in cloud I believe
19:32:39 <abadger1999> nirik: oh something else to think about --
19:32:53 <nirik> lmacken: good question. we could try 64bit again... but the memory thing is anoying.
19:32:57 <abadger1999> nirik: If we go with openid everywhere, I believe we'll be able to have SSO but different domains.
19:33:13 <lmacken> nirik: yeah :\
19:33:29 <puiterwijk> abadger1999: completely correct
19:33:31 <nirik> is there any new/cool way around it? perhaps we should ask dmalcolm again?
19:33:37 <abadger1999> ie: packages.apps.fedoraproject.org, ask.fedoraproject.org, pkgdb.fedoraproject.org, etc will all be transparently logged into because of openid.
19:33:44 <nirik> puiterwijk: cool. write it up and we can discuss it on list?
19:33:55 <pingou> abadger1999: I do like that idea :)
19:33:57 <puiterwijk> nirik: sure, will try to find the servers back again :)
19:34:02 <puiterwijk> abadger1999: +1
19:34:05 <lmacken> nirik: I don't think there is any way around it (PyObject pointers double in size on 64bit). I'll ask dmalcolm though to make sure.
19:34:25 <nirik> rhel doesn't ship 32bit python on x86_64 do they?
19:34:42 <abadger1999> lmacken: there are ways but -- not as in, "oh, I'll just patch this and[...]"
19:35:10 <lmacken> nirik: not sure
19:35:17 <nirik> yeah, I don't think so.
19:35:18 <puiterwijk> nirik: I believe they actually do, I'll check
19:35:32 <puiterwijk> nirik: they do
19:35:33 <nirik> anyhow, worth looking into before we build a bunch of servers. ;)
19:35:42 <abadger1999> it would require changing CPython data structures.  One thing dmalcolm might know, though, is whether we could run on an alternate interpreter and gain memory.
19:35:49 <nirik> but then if we use that, we need to build our stack of packages against that?
19:36:14 <abadger1999> and if we're split out, we might be able to try that on just a few apps which have a compatible stack.
19:36:20 * lmacken asked dave about it in #fedora-python
19:36:34 <puiterwijk> nirik: RHEL does ship 32-bit python on x86_64
19:36:35 <nirik> cool. lets continue looking at this out of meeting?
19:36:39 <skvidal> nirik: so.. .about building a bunch of servers..
19:36:54 <abadger1999> nirik: <nod>  -- Anyhow -- I don't htink any of the potential solutions would be ready in time for this split to multiple app servers.
19:36:54 <nirik> puiterwijk: yeah, but how many of the things we need are either noarch or have a 32bit version?
19:36:55 <skvidal> nirik: the koji builders are spun up using ansible + kickstart
19:37:19 <skvidal> nirik: we could think about implementing the new app-specific instances like that
19:37:25 <skvidal> nirik: ie - puppetless
19:37:26 <nirik> skvidal: an _excellent_ segway into my next topic...
19:37:31 <nirik> b) migration to ansible
19:37:33 <abadger1999> So if memory is our constraint, I think we should stick with 32bit and think about 64bit+other solution in a year or more.
19:37:43 <nirik> abadger1999: fair enough
19:37:54 <pingou> dmalcolm | dmalcolm: or rather:  yes: it will use nearly double the ram
19:38:11 <nirik> so, I think we need to whip up a plan for our ansible migration... and I can work on that. Then start working away at the little parts...
19:38:14 <threebean> skvidal: exciting prospect
19:38:28 <nirik> short term we need:
19:38:32 <skvidal> nagios
19:38:44 <skvidal> it'd be nice to have fas-push
19:38:46 <nirik> yeah, nagios was my c) topic. ;)
19:39:01 <nirik> it would be nice to move builder ansible repo into main ansible repo
19:39:02 <skvidal> the cron runner stuff - which is on my plate and something I'm actively working on both parts of
19:39:06 <nirik> yep.
19:39:17 <skvidal> agreed on moving builder ansible into main ansible
19:39:56 <skvidal> if anyone wants to do that and work on it
19:40:01 <nirik> we will end up with some machines puppet and newer stuff ansible, but I think thats ok for migrating... and we can keep converting away until we get to where we can have a flag day for the rest.
19:40:04 <skvidal> and I'm the blocking point
19:40:08 <skvidal> don't block on me
19:40:10 <skvidal> yell at me
19:40:18 <skvidal> and take on the process
19:40:24 <skvidal> we have lots of builders
19:40:34 <skvidal> and taking one out to test a new rebuild playbook from the main repo
19:40:39 <skvidal> is super-duper easy
19:40:42 <nirik> skvidal: I can look at doing the new mirrorlist boxes via ansible... that should possibly get us a global/whatever conversion to base others on
19:41:08 * mdomsch may need a 64-bit box for s3-mirror
19:41:08 <skvidal> nirik: I think I need to look at global again - istr it is full of crap
19:41:28 <skvidal> nirik: and by that I mean - it has a bunch of completely unnecessary system changes in it
19:41:29 <nirik> skvidal: yes, we should only convert over the parts that make sense. ;)
19:41:33 <skvidal> nod
19:41:41 <nirik> this is a good chance to clean that junk up
19:41:44 <skvidal> okay - sorry for telling you a bunch of things you already know :)
19:41:48 <nirik> mdomsch: sad that it takes so much mem
19:41:55 <nirik> not at all. :)
19:42:08 <nirik> ok, so c) on my list is 'redo nagios'
19:42:08 <mdomsch> nirik: it's tracking nearly 1M files
19:42:26 <nirik> mdomsch: yeah, understandable. ;(
19:42:51 <mdomsch> 2M really, as it does so on each end
19:42:56 <nirik> I'm going to whip up a mail on nagios and we can come up with a design... anyone interested in helping welcome.
19:43:46 <nirik> ok, any other sysadminy stuff folks had?
19:44:05 <nirik> #topic Private Cloud status update / discussion
19:44:22 <nirik> so, we need/want to update and reboot clouds... possibly next week?
19:44:22 <puiterwijk> nirik: I'm interested in helping with nagios
19:44:31 <nirik> puiterwijk: excellent. ;)
19:44:49 <dkanbier> me too, quite some experience with zabbix
19:44:51 <nirik> also, I'm interested in knocking off anything we have left before calling them production.
19:45:18 <nirik> dkanbier: cool.
19:45:26 <puiterwijk> nirik: so all current instances will get killed?
19:45:43 <nirik> puiterwijk: yep. we will need to restart them all.
19:45:54 <puiterwijk> nirik: restart or terminate?
19:46:17 <nirik> terminated. ;)
19:46:23 <nirik> since hosts will get rebooted.
19:46:36 <puiterwijk> ok
19:46:50 <nirik> but we will provide notice on that, etc.
19:47:01 <puiterwijk> ah ok, great.
19:47:14 <nirik> #topic Upcoming Tasks/Items
19:47:25 <nirik> we now have an actual schedule for f19... so I added it in:
19:47:35 <nirik> #info 2013-02-08 mass rebuild for f19 starts.
19:47:35 <nirik> #info 2013-02-11 pkgdb update.
19:47:35 <nirik> #info 2013-02-18 to 2013-02-19 smooge on site at phx2.
19:47:35 <nirik> #info 2013-02-28 end of 4th quarter
19:47:36 <nirik> #info 2013-03-29 - spring holiday.
19:47:37 <nirik> #info 2013-04-02 to 2013-04-16 ALPHA infrastructure freeze
19:47:38 <nirik> #info 2013-04-16 F19 alpha release
19:47:40 <nirik> #info 2013-05-07 to 2013-05-21 BETA infrastructure freeze
19:47:42 <nirik> #info 2013-05-21 F19 beta release
19:47:44 <nirik> #info 2013-06-11 to 2013-06-25 FINAL infrastructure freeze.
19:47:46 <nirik> #info 2013-06-25 F19 FINAL release
19:47:51 <nirik> anything folks would like to schedule or note?
19:47:53 <mattdm> yes
19:48:01 <nirik> mattdm: fire away. ;)
19:48:10 <mattdm> The cloud feature asks for a koji update in there at the end of march
19:48:29 <mattdm> (cloud image build feature)
19:48:46 <nirik> ah ha.
19:48:47 <nirik> ok.
19:48:57 <mattdm> Specifically given the current schedule, sometime between March 19th and March 26th.
19:49:03 <mattdm> Which isn't a huge window, I know.
19:49:12 <nirik> mattdm: do you want me to pencil it in at a specific date? or ... ok.
19:49:37 <mattdm> It a little bit depends on the actual koji code development.
19:49:53 <nirik> yeah.
19:49:58 <nirik> ok, will add it in. ;)
19:50:01 <mattdm> March 26th is chosen as 1 week before alpha freeze, and it kinda works back from there.
19:50:10 <mattdm> thanks. that's all. :)
19:50:20 <nirik> ok, keep us posted. ;)
19:50:26 <mattdm> yep, i will.
19:50:35 <nirik> #topic Open Floor
19:50:42 <nirik> anyone have anything for open floor...
19:50:57 <abompard> almost made it this time... :/
19:51:13 <nirik> pingou: say, whats the status of fedocal?
19:51:30 <nirik> abompard: welcome. ;) whats the word on hyperkitty/mm3? ;)
19:51:33 <abompard> hey everyone, just joining late to say hello :-)
19:51:37 <pingou> nirik: I have some RFE do cover, need to start packaging it
19:52:02 <abompard> nirik: progress ! mm3test should be up-to-date, much debugging today thanks to pingou
19:52:08 <pingou> nirik: basically, still in 0.1.0-alpha but little bugs were reported, so either it works fine or few people tested it :)
19:52:23 <pingou> http://mm3test.fedoraproject.org/hyperkitty/
19:52:32 <nirik> pingou: thats another one I think we could move to wider testing soon/stg?
19:52:37 <pingou> it has in-thread reply!
19:52:46 <pingou> nirik: it would be nice for sue
19:53:38 <pingou> lmacken: https://fedorahosted.org/hyperkitty/ticket/38
19:54:14 <nirik> abompard: can you add openid to the login auth?
19:54:18 <nirik> or is that planned.
19:55:03 <abompard> nirik: hmm, I'm not sure we should open any type of auth, I'd rather keep it to the email providers and browserid
19:55:12 <abompard> nirik: since it's a mailing-list app
19:55:26 <lmacken> pingou: thanks!
19:55:30 <abompard> nirik: I need to make sure an email will be provided
19:55:30 <pingou> abompard: fas provides openid, so does google/yahoo
19:56:00 <puiterwijk> abompard: when you would configure it for FAs-OpenID, you're sure the user has a working email
19:56:00 <nirik> we can get email from fas openid I think...
19:56:05 * nirik nods.
19:56:07 <abompard> pingou: I know, but any openid provider does not provide an email address (it's the other way around)
19:56:26 <puiterwijk> abompard: FAS-OpenID provides email addresses
19:56:31 <puiterwijk> abompard: and even if it wouldn
19:56:35 <pingou> puiterwijk: HK isn't fedora specific
19:56:41 <puiterwijk> ah ok
19:57:01 <puiterwijk> sorry, then I was misled by the demo I got linked to last week ;)
19:57:05 <nirik> can't you ask for email in the openid request and fail if you don't get one back?
19:57:09 <puiterwijk> yes
19:57:30 <puiterwijk> you can say that you require an email address, and the openid protocol will abort if the provider doesn't want to give it
19:57:32 <nirik> but of course someone could send foo@bar.com...
19:57:53 <abompard> nirik: why not, I also must be sure that the openid provider validates the email adress
19:57:55 <puiterwijk> nirik: then that's the provider providing that email. most providers check it
19:58:02 <abompard> or you could impersonate anyone
19:58:36 <puiterwijk> abompard: in that case, we have the provider url and blacklist him?
19:58:39 <nirik> wlel, it's already pretty easy to do with smtp, but yeah, there's issues.
19:58:54 <nirik> anyhow, openid would be nice if we can figure a good way to do it.
19:59:14 <threebean> abompard: I love HK, btw.  nice work.
19:59:15 <abompard> nirik: true. But the user has settings here, like his password that you could reset, etc.
19:59:20 <abompard> threebean: thanks !
19:59:40 <nirik> yeah.
19:59:53 <nirik> ok, we are almost out of time now. any last minute open floor items?
20:00:17 <nirik> ok, thanks for coming everyone. ;)
20:00:20 <nirik> #endmeeting