Also avaliable at:
20:01 < fedbot`> Meeting started Thu Jun 18 20:01:09 2009 UTC. The chair is
20:01 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Who's
20:01 < fedbot`> Information about MeetBot at http://wiki.debian.org/MeetBot
Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:01 * ricky
20:01 * lmacken
20:02 < nirik> there it goes.
20:02 * sthistle says hi.. 1st time
20:02 < ricky> Welcome :-)
20:02 -!- fedbot [n=supybot(a)scrye.com] has quit Read error: 113 (No route to host)
20:02 -!- fedbot` is now known as fedbot
20:02 < mmcgrath> Ok, so lets get started
20:03 < mmcgrath> #topic Infrastructure -- Tickets
20:03 < Jorn23> I'm here too
20:03 -!- fedbot changed the topic of #fedora-meeting to: Infrastructure -- Tickets
20:03 < mmcgrath> .tiny
20:03 < zodbot> mmcgrath: http://tinyurl.com/47e37y
20:03 < mmcgrath> .ticket 1464
20:03 < zodbot> mmcgrath: #1464 (Puppet Web Apps Refactoring) - Fedora
Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/1464
20:03 < mmcgrath> ricky: you want to take this one?
20:03 < mmcgrath> Jorn23: hey
20:03 < ricky> Sure
20:03 < Jorn23> heya :)
20:03 < ricky> I've been working on converting all of our apps to puppet
20:04 * dgilmore is here
20:04 < ricky> Right now, most of our applications are running off of the modules
under the stg.fedoraproject.org
20:04 < ricky> For this first piece, my goal is to only touch the app and proxy
servers to keep things manageable.
20:04 * abadger1999 here
20:05 < mmcgrath> ricky: if you were to guess, what % done are you?
20:05 < ricky> The way that I'd *like* to move things over is to get the staging
environment into the exact configuration that we want, and then just drop it right into
the master branch.
20:06 < ricky> I have all of our large apps moved, but I haven't gotten all of
the behaviors of our production configs, like some redirects and stuff.
20:06 < ricky> So maybe 70-80% or so. I'm going to try and get the details
matched up today, then send an email out asking people ot make sure that
*.stg.fedoraproject.org behaves exactly like they want it.
20:07 < mmcgrath> excellent.
20:07 < mmcgrath> ricky: thanks for that
20:07 < mmcgrath> ricky: should we avoid messing staging up much until that's
20:07 -!- knurd is now known as knurd_afk
20:08 -!- sgodsell [n=sgodsell@unaffiliated/sgodsell] has joined #fedora-meeting
20:08 < ricky> Most of the puppet modules aren't changing much at this point
20:08 < ricky> Just keep in mind if that if you want configuration changes to show
up in staging, you need to edit the files in the module
20:08 < mmcgrath> <nod>
20:08 < mmcgrath> ricky: anything else on that?
20:09 < ricky> Nope, thanks
20:09 < mmcgrath> coolz.
20:09 -!- JukeBoxHero [n=gorillaJ@fedora/hitboxx] has quit "No Ping reply in 90
20:10 < mmcgrath> Ok, so thats the last ticket there.
20:10 -!- JukeBoxHero [n=gorillaJ@fedora/hitboxx] has joined #fedora-meeting
20:10 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Cloud stuff
20:10 -!- maploin [n=mapleoin@fedora/maploin] has joined #fedora-meeting
20:10 < mmcgrath> So there's not been too much going on with this in the past
month or so but if you've been watching the commit logs you'll notice I've
been rebuilding it all.
20:10 < mmcgrath> I'm hoping to get something up and going fairly soon.
20:11 -!- kital [n=Joerg_Si@fedora/kital] has quit Remote closed the connection
20:11 < mmcgrath> I've been using it off and on with mostly success.
20:11 -!- BaRRa [n=tmakinen(a)tk-cn0002.oulu.fi] has joined #fedora-meeting
20:11 -!- denise [n=ddumas(a)126.96.36.199] has quit Read error: 110 (Connection timed
20:11 < mmcgrath> There's a whole other web interface that some various people
are working on to make it less 'ovirt ui' ish and more 'end user fedora
20:11 < mmcgrath> No ETA on that.
20:11 < dgilmore> mmcgrath: :)
20:12 < Jorn23> What exactly is "cloud stuff" ?
20:12 < mmcgrath> but I am hoping to queue up some of the queues.
20:12 < mmcgrath> Jorn23:
20:12 < mmcgrath> I think that's fairly up to date.
20:12 < mmcgrath> with what we're doing
20:12 < Jorn23> Cheers, I'll read that up
20:12 < mmcgrath> but it's basically an http://ovirt.org/
20:13 < mmcgrath> Anyone have any questions on that?
20:13 -!- denise [n=ddumas@nat/redhat/x-7d59d02879b98a23] has joined #fedora-meeting
20:13 < mmcgrath> Ok.
20:14 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Lots of
20:14 < mmcgrath> so we've had a lot of changes since F11 shipped.
20:14 < mmcgrath> probably more changes in the last couple of weeks then we've
20:14 < mmcgrath> some have been major, some minor, but there's been a bunch of
20:14 < mmcgrath> I only bring it up because we've been shaking out bugs and
20:14 * dgilmore has one coming up that id like to talk about at some point
20:14 < mmcgrath> so keep your eyes open.
20:15 < mmcgrath> and if you find something, please don't assume we know about
20:15 < mmcgrath> Thats really all I had on that :)
20:16 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- dgilmore
wants to talk about "one coming up"
20:16 < mmcgrath> :)
20:16 < mmcgrath> dgilmore: whats the poop?
20:16 < dgilmore> mmcgrath: im killing plague tomorrow
20:16 * mmcgrath cheers
20:16 < mmcgrath> well sort of.
20:16 * nirik jumps up and down in joy.
20:16 < mmcgrath> dgilmore: whats our process going to be for signing and pushing?
20:16 < mmcgrath> do I need to build another signing server?
20:16 < dgilmore> epel will switch to koji
20:16 < dgilmore> i think i have bodhi ready
20:16 < dgilmore> but i need to run some things by lmacken
20:17 < dgilmore> who i can never get ahold of
20:17 < ricky> Feel free to update it in staging if you have a newer package.
20:17 < jwb> another signing server. ew
20:17 * mmcgrath wonders if koji.stg.fedoraproject.org
is back up and running.
20:17 < dgilmore> i need to setup epel_releng
20:17 < dgilmore> epel-releng1
20:17 < dgilmore> something like that
20:17 < mmcgrath> dgilmore: thats a group or a host?
20:17 < dgilmore> mmcgrath: a host
20:18 < mmcgrath> dgilmore: who's going to be able to sign and push?
20:18 < dgilmore> to do mash and package signing on for epel
20:18 < mmcgrath> how long is that process going to take
20:18 < mmcgrath> etc
20:18 -!- kital [n=Joerg_Si@fedora/kital] has joined #fedora-meeting
20:18 < dgilmore> mmcgrath: epel_signers
20:18 < dgilmore> mmcgrath: until we can test not sure
20:18 < dgilmore> but much shorted than fedora updates
20:18 < dgilmore> shorter
20:18 < mmcgrath> so maybe we don't want to kill plague until we know the rest
of that stuff? :)
20:19 < mmcgrath> dgilmore: so aside from the additional signing server what do we
20:19 < jwb> dgilmore, why is it going to be shorter?
20:19 < f13> less packages?
20:19 < mmcgrath> or do we need another releng box and not a signing server for
20:19 < dgilmore> jwb: alot less packages
20:19 < nirik> many fewer packages.
20:19 < jwb> ah
20:19 < f13> less arches too
20:19 < dgilmore> mmcgrath: just another releng box
20:19 < f13> does EPEL do ppc(64) ?
20:19 < jwb> i was thikning frequency. not per-push duration
20:19 < f13> will it?
20:19 < dgilmore> epel does ppc
20:19 < nirik> f13: it does.
20:20 < dgilmore> but not ppc64
20:20 < f13> ah ok
20:20 < dgilmore> i was approached this weeke about doing s390x for epel also
20:20 < mmcgrath> dgilmore: what about disk? right now releng2 requires direct
write access to /mnt/koji
20:20 < nirik> the only weird thing is there is no desktop RHEL for pcc.
20:20 < nirik> ppc
20:20 < mmcgrath> I'd assume we don't want that for EPEL?
20:20 < dgilmore> mmcgrath: the epel releng box will need to write also
20:20 < dgilmore> mmcgrath: the process will be the same
20:20 < mmcgrath> :: sigh ::
20:21 < mmcgrath> are we really sure we want fedora's process infecting epels?
20:21 < dgilmore> and i guess we ould use the existing releng resources
20:21 < mmcgrath> epel used to have such low overhead.
20:21 < f13> mmcgrath: depends on how y ou look at it
20:21 < dgilmore> mmcgrath: lots of people want it
20:21 < f13> mmcgrath: shuffling the repodata around different locations isn't
exactly "low" overhead
20:21 < dgilmore> mmcgrath: updates notifications will be a big win
20:22 < mmcgrath> f13: I look at it as someone who watched jwb freak out a couple of
days ago because he started a push that he wasn't sure would finish in time for the
pkgdb update which was something like 18 hours away at the time.
20:22 < nirik> also epel has no deltas to worry about.
20:22 < f13> mmcgrath: we're talking about different orders of magnitute
20:22 < f13> magnitude
20:22 < jwb> mmcgrath, yeah, what f13 said
20:22 < mmcgrath> right now we are.
20:22 < lmacken> the pkgdb update doesn't effect pushing updates
20:22 < lmacken> at all
20:22 < dgilmore> nirik: right at least until we get to rhel 6
20:22 < jwb> also what lmacken said
20:22 < abadger1999> lmacken: <nod> But jwb didn't *know* that.
20:22 < mmcgrath> lmacken: thats my other point. No one was quite sure about that
because of how many moving parts were involved.
20:22 < jwb> but i see where mmcgrath is coming from
20:23 < lmacken> right
20:23 < mmcgrath> anywho, if we're going forward with it we're going forward
20:23 < mmcgrath> dgilmore: you have a pretty good idea of how a relepel1 would
20:23 < dgilmore> mmcgrath: i think we should go ahead
20:23 < dgilmore> mmcgrath: i do
20:24 < mmcgrath> k, then we'll go forward.
20:24 -!- Dp67 [n=rberry@fedora/Dp67] has joined #fedora-meeting
20:24 < mmcgrath> dgilmore: ping me after the meeting about getting that server up,
I'm positive we've got space for it somewhere.
20:24 < mmcgrath> f13: and that reminds me, I owe you a shiny new signing server :)
20:25 < dgilmore> mmcgrath: :) thanks
20:25 < f13> mmcgrath: cool
20:25 < mmcgrath> it's racked, I need to throw an OS on there and talk to you
about migrating to it whenever you guys are ready.
20:25 < mmcgrath> hell I might just be able to use dd and copy the old signing
20:25 < dgilmore> mmcgrath: if we get the signing server up and running it should be
useable for EPEL as well as fedora
20:25 -!- cassmodiah [n=cass@fedora/cassmodiah] has quit Remote closed the connection
20:25 < mmcgrath> well, this signing server is just different hardware.
20:25 -!- warren [n=warren@redhat/wombat/warren] has quit "Leaving"
20:26 < mmcgrath> as for the signing server software, I think that's still a
20:26 < mmcgrath> dgilmore: anything else on the EPEL move away from plague?
20:26 < dgilmore> mmcgrath: nope
20:26 < mmcgrath> coolz
20:26 < mmcgrath> #topic Infrastructure -- Signing Server
20:26 -!- fedbot changed the topic of #fedora-meeting to: Infrastructure -- Signing
20:27 < f13> we have signing server software to start using
20:27 < mmcgrath> f13: since you're here, do you want me to build the new
signing server virtual so you guys can still log in and virsh and stuff?
20:27 < f13> and with the new hardware, now would be a good time to try and deploy
20:27 < mmcgrath> f13: or do you want it to be physcial and require console access?
20:27 < mmcgrath> oh, excellent.
20:27 < f13> that's... a good question
20:27 < f13> can we get back to you on that? (:
20:27 < mmcgrath> f13: that'd probably work best then, it'll allow you guys
to keep using the current signing box while the new one is getting ready.
20:27 < lmacken> f13: what software?
20:27 < f13> lmacken: mitr wrote software
20:27 < lmacken> f13: where is it?
20:27 < mmcgrath> f13: yeah, I'm on whatever schedule you guys want me to be :)
20:28 * ricky had no idea people were working on signing software
20:28 < f13> https://fedorahosted.org/sigul/browser
20:28 -!- kital [n=Joerg_Si@fedora/kital] has quit Remote closed the connection
20:28 < lmacken> oh, nice :)
20:28 -!- tatica is now known as tatica-out
20:29 < f13> http://git.fedorahosted.org/git/?p=sigul.git;a=summary
if you prefer
20:29 < f13> it hasn't been touched in a while because we wanted to get F11 out
before we disrupted the signing process.
20:29 < mmcgrath> heh
20:30 -!- dthomasdigital [n=dthomasd@ubuntu/member/dthomasdigital] has quit
20:30 < mmcgrath> K
20:30 < mmcgrath> we'll take a look at that
20:30 < mmcgrath> any other questions on the signing server?
20:31 < mmcgrath> Alllrighty.
20:31 -!- N3LRX [n=rberry@fedora/Dp67] has joined #fedora-meeting
20:31 * mmcgrath thinks of anything else...
20:31 < mmcgrath> I think that's really all I've got right now
20:31 < mmcgrath> #topic Infrastructure -- Who's here?
20:31 -!- fedbot changed the topic of #fedora-meeting to: Infrastructure -- Who's
20:31 < mmcgrath> #topic Infrastructure -- Open Floor
20:31 < mmcgrath> oops :)
20:31 -!- fedbot changed the topic of #fedora-meeting to: Infrastructure -- Open Floor
20:31 < nirik> I have one quick item.
20:31 < ricky> Heh
20:32 < mmcgrath> nirik: have at it
20:32 -!- kolesovdv [n=kolesovd(a)188.8.131.52] has quit Remote closed the connection
20:32 < nirik> I've been working with the meetbot plugin author... he hopes to
have a release in the next few days with more features/cleanups/fixes. At that point I
would like to set it up with zodbot and see about having it log to a fedoraproject.org
area... also a search engine of some kind on it would be nice.
20:33 -!- sping [n=sping@gentoo/user/sping] has joined #fedora-meeting
20:33 < dgilmore> nirik: its a neat bot
20:33 < nirik> If anyone has feedback for it, let me know. If I need to request
anything or get anything approved as far as setting it up, let me know. ;)
20:33 < nirik> I already have it packaged. Just waiting to import it for the changes
20:34 < mmcgrath> nirik: I've been fine with it.
20:34 < mmcgrath> can anyone use the start and stop meeting bits?
20:34 < nirik> yep.
20:34 < mmcgrath> how's that access figured out?
20:34 < nirik> it's open to anyone on a channel where that plugin is enabled.
20:34 < mmcgrath> excellent
20:35 < nirik> however, there are now some native supybot commands to let bot admins
close meetings/add chairs
20:35 < mmcgrath> nirik: we'll throw up some sort of web deal for logging
20:35 < mmcgrath> if we haven't already
20:35 < mmcgrath> ricky: have we already?
20:35 < ricky> Not now
20:35 < nirik> so, an admin could override and close a meeting that was mistakenly
20:35 -!- mchua [n=mchua(a)184.108.40.206] has quit Read error: 60 (Operation timed out)
20:36 -!- sdziallas [n=sebastia@fedora/sdziallas] has joined #fedora-meeting
20:36 < nirik> the upstream author is working on converting it to use Restructured
text for it's summary... and then spit out html and a text suitable for posting to
20:36 -!- mchua [n=mchua@nat/redhat/x-a4d6433a89727a6d] has joined #fedora-meeting
20:36 < nirik> anyhow, just thought I would bring it up. Thats it.
20:36 -!- kolesovdv [n=kolesovd(a)220.127.116.11] has joined #fedora-meeting
20:37 < mmcgrath> nirik: excellent, thanks for keeping up with that
20:37 < mmcgrath> Anyone have anything else they'd like to discuss?
20:38 < Jorn23> Just had a question for ricky
20:38 < Jorn23> Or well, not really only for ricky
20:38 < Jorn23> Anyway
20:39 < ricky> Ask away :-)
20:39 < Jorn23> A couple of days ago we talked about me helping with the migrations
to ha-proxy, when do you think you have time for that ricky?
20:39 < mmcgrath> Jorn23: have at it
20:39 < mmcgrath> Jorn23: yeah we found lots of small odds and ends that were still
in balancer :)
20:40 < ricky> There are a few that we could get after the meeting if you're
20:40 < ricky> It'd be helpful for the work I'm doing in staging, actually.
20:40 < mmcgrath> ricky: yeah, I figure if we could go through one or two with
Jorn23 he'd be able to get the rest
20:40 < ricky> There might also be a few that I did in staging that would be good to
20:40 -!- Dp67 [n=rberry@fedora/Dp67] has left #fedora-meeting ["Leaving"]
20:40 < Jorn23> That sounds good yeah
20:41 < Jorn23> But we can discuss the details outside the meeting I imagine.
20:42 < mmcgrath> Jorn23: sounds good, thanks for helping us work on that :)
20:42 < mmcgrath> anyone have anything else they'd like to discuss?
20:44 < mmcgrath> ok, well with that we'll close the meeting
20:44 < f13> I should bring up autotest
20:44 < mmcgrath> f13: oh, have at it
20:44 < mmcgrath> #topic Infrastructure -- autotest
20:44 -!- fedbot changed the topic of #fedora-meeting to: Infrastructure -- autotest
20:44 -!- tibbs [n=tibbs@fedora/tibbs] has quit "Konversation terminated!"
20:44 < f13> the QA team and I have been working on getting something ready to go in
Fedora Infra for automated testing
20:44 < f13> There was a work in progress "invented here" effort called
AutoQA, however we recently discovered Autotest from the kernel project
20:45 < f13> it seems to fit our needs nicely and already has an upstream and
community around it
20:45 < f13> so I've spent the week trying to get it packaged
20:45 < f13> and we plan to deploy it within FI shortly to do some initial automated
20:45 < f13> however there is a bit of a problem
20:45 < f13> the server side of autotest requires an older Django than what's in
20:45 < f13> and building autotest properly requires use of Google Web Toolkit
20:46 < f13> gwt isn't in Fedora, and will be an extreme pain to get in, due to
the snowball effect of piles of java
20:46 < mmcgrath> hahahah
20:46 < ricky> Blech.
20:46 < mmcgrath> this is going to be a pile of fun :)
20:46 < f13> however gwt is in a RHEL5 product, and I think the Django version
available for EL5 matches what we need
20:47 < f13> so I'm working on a package I can build against RHT's brew that
has gwt, to produce something suitable for an EL5 host. gwt isn't needed at rumtime,
20:47 < f13> we'll continue to try and get gwt into Fedora, but that's going
to have to be a project all on its own
20:47 -!- Sonar_Guy [n=Who@fedora/sonarguy] has joined #fedora-meeting
20:47 < mmcgrath> f13: k, I'll be happy to help with that when the time comes.
We do have a free software policy to make sure we follow -
20:47 < f13> If I get something built via RHT's brew, would we be able to use it
within FI to run an autotest host?
20:47 < mmcgrath>
20:48 < mmcgrath> But yeah, I think we'll have hosts for it and stuff
20:48 < f13> all the software involved is free, it's just a huge hassle to get
20:48 < f13> I think james may have some machines to throw at it too, mostly I was
looking for a policy decision
20:49 < f13> the other thing we've done is had upstream pre-translate the java
20:49 < mmcgrath> yeah, we'll figure something out.
20:49 < mmcgrath> f13: anything else on that for now?
20:49 < f13> but it did remove our need of gwt during package build. I'd prefer
that we stuck with using gwt at build time, from RHT
20:50 < f13> nope. We'll likely be looking to host it after FUDCon Berlin
20:50 < mmcgrath> f13: sounds good
20:50 < f13> one last thing, has folks here had enough time to review the proposals
that came from the recent FAD?
20:50 < mmcgrath> f13: There were proposals from FAD?
20:50 < f13> oh my yes
20:51 < f13>
20:51 < f13> please do read and comment if there are any concerns, as we move these
proposals further into planning and polishing up for submitting to FESCo
20:52 < mmcgrath> f13: k, I'll send a note to the infrastructure list and let
people to know to read over them, I'll do the same.
20:52 < abadger1999> f13: Which ones affect FI?
20:52 -!- openpercept [n=openperc@fedora/openpercept] has quit Read error: 110 (Connection
20:52 < abadger1999> Everything but hte first?
20:52 < f13> milestone does, as it reduces the freeze points (no more alpha), and
20:52 < f13> pretty much all of them in one way or another.
20:53 < mmcgrath> f13: was anyone from Infrastructure at FAD?
20:53 * mmcgrath isn't sure who attended that
20:53 < ricky> I think skvidal might have been there
20:53 < mmcgrath> I think seth, jesse and luke?
20:54 < f13> yep
20:54 < mmcgrath> Ah, I see the Attendees section in the link.
20:54 < mmcgrath> K, anyone have anything else to discuss?
20:54 < mmcgrath> If not we'll close in 30
20:55 < mmcgrath> #endmeeting