FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.

» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Redhat > Fedora Infrastructure

LinkBack Thread Tools
Old 09-11-2008, 09:21 PM
Ricky Zhou
Default Meeting Log - 2008-09-11

20:00 -!- dgilmore changed the topic of #fedora-meeting to: infra meeting prep
20:00 * abadger1999 stretches
20:00 < dgilmore> hey all meetingtime
20:00 * ValHolla walks in
20:00 < SmootherFrOgZ> hello guys
20:00 < abadger1999> Hey ValHolla!
20:01 < abadger1999> Hello SmootherFrOgZ
20:01 < brothers> EHLO
20:01 < ValHolla> Hello
20:01 * wakko666 lurks.
20:01 < dgilmore> we will start with https://fedorahosted.org/fedora-infrastructure/query?status=new&status=assigned&status=reopened&g roup=milestone&keywords=~Meeting&order=priority
20:01 < dgilmore> the tickets
20:01 * fchiulli is observing
20:01 * mmcgrath notes G said he won't be around and sends his best.
20:01 < dgilmore> .ticket 753
20:01 < zodbot> dgilmore: #753 (Mini-freeze for Fedora 10 beta) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/753
20:01 * ricky
20:01 * nokia3510 says hello
20:02 < dgilmore> we need to plan infra freeze for F-10 release
20:02 < dgilmore> and a mini one for beta
20:02 < dgilmore> f13: when is beta. what would you like frozen?
20:04 < dgilmore> mmcgrath: thanks
20:04 < f13> mmcgrath write up a chart that showed what would be frozen and what wouldn't. THe beta freeze is supposed to be today
20:04 -!- skvidal [n=nnnnnnsk@fedora/skvidal] has joined #fedora-meeting
20:04 < mmcgrath> f13: ah, k. I wasn't sure about the dates. given the fas issues that are going on, mind if we start the freeze tomorrow?
20:05 < dgilmore> f13: i have a koji outage scheduled for tomorrow
20:05 < dgilmore> i guess we should have talked about this last week
20:05 -!- kital [n=Joerg_Si@fedora/kital] has quit Remote closed the connection
20:06 < mmcgrath> yeah, its the first time we've really done a beta freeze, I wasn't sure when it was going to start either.
20:06 < mmcgrath> f13: so next time we'll be more careful, mind if we start the freeze on the 13th?
20:06 < f13> this beta is looking like a rolling wreck anyway, so sure, why not
20:06 < mmcgrath> hahahah
20:06 < dgilmore> f13:
20:06 < abadger1999> As long as it's rolling :-)
20:06 < mmcgrath> on the f13th
20:06 < f13> so it goes.
20:07 < dgilmore> so that means all changes need to be sent to infra list first and acked by one other?
20:07 < mmcgrath> yeah. I'll send a note out after the meeting similar to the last one. I should update the release SOP as well.
20:08 < dgilmore> do we roll from mini freeze to full freeze? or do we unfrezze once beta is out and frezze again for test and final?
20:09 < dgilmore> lets talk this over on the list
20:09 < dgilmore> .ticket 395
20:09 < zodbot> dgilmore: #395 (Audio Streaming of Fedora Board Conference Calls) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/395
20:09 < dgilmore> jcollie: ping
20:09 < jcollie> dgilmore: ping
20:09 < dgilmore> any update
20:09 < f13> I'm fine with unfreeze/refreeze later
20:09 < SmootherFrOgZ> dgilmore: mail says on Saturnday 13 for the outage
20:09 < dgilmore> SmootherFrOgZ: 1:00am UTC which is 8pm friday my time
20:10 < jcollie> dgilmore: nope, been busy with other stuff
20:10 < dgilmore> jcollie: it happens
20:10 -!- cassmodiah [n=cass@p54AB3DC0.dip.t-dialin.net] has quit Remote closed the connection
20:10 < SmootherFrOgZ> dgilmore: k
20:10 < dgilmore> .ticket 446
20:10 < zodbot> dgilmore: #446 (Possibility to add external links on spins page) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/446
20:10 < dgilmore> this is me
20:10 < dgilmore> i fail
20:10 < dgilmore> .ticket 740
20:11 < zodbot> dgilmore: #740 (Loaning out system time to OLPC participants) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/740
20:11 < dgilmore> so I got from OLPC what it is they wanted
20:11 < dgilmore> as i suspected its a machine where people can log in. have cvs, koji, mock, etc access
20:12 < dgilmore> so they can do test builds. package management etc
20:12 -!- kital [n=Joerg_Si@fedora/kital] has joined #fedora-meeting
20:13 < dgilmore> there is quite a few people at OLPC who don't use fedora and are not familiar with fedora and how it works
20:13 < dgilmore> so part of what they want is also help to train people on how to work with fedora. how to maintain packages.
20:14 < dgilmore> mock is packaged and works on debian
20:14 < dgilmore> koji is not packaged
20:15 < dgilmore> so does anyone have ideas on what we can do to help
20:15 < nokia3510> I don't get it why they need a machine in this case.
20:15 < f13> I really don't understand why they can't supply this themselves
20:15 < f13> it's not very hard to put up a RHEL or Fedora host and install those packages
20:16 < f13> or is the thing they really want help with account management, in the form of FAS logins?
20:16 < dgilmore> nokia3510: so that people who dont use fedora can get access to it
20:16 < dgilmore> f13: some of it is fas
20:16 < dgilmore> and account management
20:17 < dgilmore> When i was there i was pushing to get fas in place. and i think it will happen over time
20:17 < f13> so tell them if they host a quadg5 system, we'll let them use fas to log in on it (:
20:18 < ValHolla> The reply from OLPC sounds to me like they are looking for a "tutor" ... someone to teach the non Fedora folks
20:18 < dgilmore> some of it is trying to get it closer to fedora. so that fedora developers could help train olpc's developers in fedora
20:18 < dgilmore> f13:
20:18 < f13> er, why does it matter where the box lives for that?
20:18 * f13 is seriously confused by the folks over there.
20:19 < dgilmore> it really doesnt
20:19 < dgilmore> they currently have a box they give out accounts on
20:19 < f13> and really, all they have to do is *ask*
20:19 < f13> but we've seen evidence of where that just doesn't happen
20:19 < dgilmore> but there are concerns its not as secure as it could be
20:19 * ValHolla wonders if it wouldn't be a good Idea to create #fedora-OLPC and try to have some of us in there for the OLPC to ask questions to
20:20 < dgilmore> ValHolla: it exists already
20:20 < dgilmore> well #fedora-olpc
20:21 < dgilmore> they are doing a better job of asking for help, than has been the case in the past
20:21 < ValHolla> dgilmore: ok, well a specific channel the OLPC folks know they can go to for a quick answer/tutorial on what they are doing?
20:21 < ValHolla> if they get stuck using mock, koji etc...
20:21 < wakko666> ValHolla: like #fedora, or #koji, or #fedora-admin?
20:21 < dgilmore> ValHolla: they usually come to me and I point them in the right direction.
20:22 < dgilmore> #fedora-devel
20:22 < dgilmore> there are quite a few olpc folks in #fedora-devel
20:22 < ValHolla> yes like those.... but there are so many different places to go.
20:22 < dgilmore> perhaps what we can do is help them with kickstart files, FAS help, etc
20:23 < dgilmore> help get an environment setup where they could rebuild the box that is used by people to test builds and play with things on a regular basis
20:24 -!- stickster_afk is now known as stickster
20:24 < dgilmore> help them to understand fas
20:24 < ValHolla> roughly how large an image do they use? will it all fit on a single dvd?
20:24 < dgilmore> and help get it implemented
20:24 < ValHolla> make them a "Live CD on steroids"
20:24 < dgilmore> ValHolla: image for what?
20:25 < ValHolla> kickstart image
20:25 * ValHolla is throwing solaris terms around ... sorry
20:25 < dgilmore> ValHolla: well it would need to be a developer spin of sorts
20:26 < nokia3510> I'd ask where Fedora-related stuff ends and strict OLPC particularities appear
20:26 < dgilmore> having fedora-packager gcc rpm-build etc installed
20:26 < nokia3510> because it seems to be a big overlap
20:26 -!- warren [n=warren@redhat/wombat/warren] has quit "Leaving"
20:26 < dgilmore> nokia3510: there are a few things in OLPC that can not be in fedora
20:26 < dgilmore> but most of it really should be built in fedora
20:27 -!- TopoMorto [n=TopoMort@] has quit "Sto andando via"
20:27 < dgilmore> anyways. lets move on
20:27 < nokia3510> ok
20:28 < dgilmore> feel free to make comments in the ticket with ideas
20:28 -!- dgilmore changed the topic of #fedora-meeting to: Fedora Infrastructure - Open Floor
20:28 < dgilmore> anyone have anything they would like to bring up?
20:29 < SmootherFrOgZ> what about hosting request ?
20:29 < dgilmore> SmootherFrOgZ: which hosting request?
20:29 -!- LetoTo1 [n=paul@76-10-173-74.dsl.teksavvy.com] has joined #fedora-meeting
20:29 < dgilmore> SmootherFrOgZ: OLPC's?
20:29 < SmootherFrOgZ> nope
20:30 < SmootherFrOgZ> seth ask few weeks ago, that fp.o was looking for more hosting place
20:30 < dgilmore> skvidal: ^^^
20:30 < skvidal> we're always looking for more hosting locations
20:31 < skvidal> we found some new space w/ibiblio
20:31 < skvidal> very kindly of them
20:31 < ricky> So torrent server there?
20:31 -!- dwmw2_gone is now known as dwmw2_HIO
20:31 < skvidal> ricky: xen host there
20:31 < skvidal> torrent will be on it
20:31 < SmootherFrOgZ> , skvidal can you draft me details about what fp.o is looking for
20:31 < ricky> Niice
20:31 < skvidal> SmootherFrOgZ: mmcgrath has a better background in what's needed
20:31 < skvidal> ricky: iirc the box has been ordered
20:32 < SmootherFrOgZ> actually, our hosting sevice need more details to reply on that
20:32 < SmootherFrOgZ> service*
20:32 < ricky> Coolnsess
20:32 < skvidal> SmootherFrOgZ: but the gist is 1-2U +power +network
20:32 < ricky> **Coolness
20:32 < skvidal> SmootherFrOgZ: ideally, a serial console and the ability to power off/on the machine
20:32 < skvidal> SmootherFrOgZ: though we're willing to work with various amount of other things
20:32 < dgilmore> SmootherFrOgZ: or an easy way to get it handled
20:32 < ricky> Non-blocked ports are usually good
20:33 < SmootherFrOgZ> k
20:33 < skvidal> ricky: +1
20:33 < dgilmore> disk is good to
20:33 < skvidal> dgilmore: well, normally we can provide the box
20:33 < skvidal> if we can get the space
20:33 < dgilmore> skvidal: true.
20:34 < dgilmore> we have some hosting that is rack space+power+network. we provide servers. we also have some where servers are provided
20:34 < dgilmore> depends on whats on offere
20:34 < dgilmore> offer
20:35 < skvidal> nod
20:35 < skvidal> SmootherFrOgZ: does that help?
20:36 -!- JSchmitt [n=s4504kr@fedora/JSchmitt] has quit "Konversation terminated!"
20:36 < SmootherFrOgZ> yep, thanks
20:37 < dgilmore> moving on
20:37 < abadger1999> I'm going to be at a memorial service tomorrow. Out all day. I'll probably be *very* spotty on the weekend as well with all the relatives in town.
20:37 -!- stickster is now known as stickster_afk
20:37 -!- stickster_afk is now known as stickster
20:37 < dgilmore> abadger1999: ok
20:37 * dgilmore will be updating koji to 1.2.6
20:38 -!- bzbot is now known as buggbot
20:38 < ricky> Any thoughts about upgrading postgres on db2?
20:38 < dgilmore> so we will need to look and see if we start hitting bugs.
20:38 < dgilmore> abadger1999: ^
20:38 < dgilmore> ricky: im for it.
20:38 < abadger1999> I'd like to do it ASAP.
20:38 < SmootherFrOgZ>
20:38 < skvidal> abadger1999:
20:38 < abadger1999> mmcgrath: Thoughts on when we can do that?
20:38 < dgilmore> from my use of 8.3 its much better than 8.1
20:38 < ricky> No rush or anything, just wondering when it can go down (did we ever end up scheduling an outage for it?)
20:39 < ricky> dgilmore: That's really good to hear :-)
20:39 < mmcgrath> abadger1999: just need someone to do it. I can try to do it next week sometime
20:39 < SmootherFrOgZ> mmcgrath: i can give you a hand on that
20:39 < abadger1999> mmcgrath: k. I can help with the data manipulation. It's just dumping the data on one box and loading on another.
20:40 -!- mbacovsk_ [n=mbacovsk@nat/redhat/x-574ae97d3509588a] has joined #fedora-meeting
20:40 < mmcgrath> SmootherFrOgZ: thanks. I'm hoping it will be pretty straight forward but I'll make sure you can be around the night of.
20:40 < mmcgrath> we have a staging box now too so hopefully that will make things nice and clean
20:40 < ricky> abadger1999: Does it necessarily need to be on different boxes?
20:40 < abadger1999> ricky: Could be the same box even.
20:40 < nokia3510> can I help too ?
20:40 < ricky> Actually, I guess it could be good to test the postgres 8.3 manifest as well
20:40 < abadger1999> It's easier for reverting if they're separate boxes.
20:40 < ricky> And there's waay less data to transfer each time, I forgot
20:41 < SmootherFrOgZ> mmcgrath: no pb
20:41 < ricky> mmcgrath: So how will staging play into this?
20:42 < mmcgrath> ricky: we'll run it there first, if it works fine we'll run it in production
20:42 < ricky> Cool.
20:43 < ricky> Also, any update on the mod_wsgi mirrorlist?
20:43 < ricky> Is it close to ready?
20:43 < abadger1999> I think mdomsch just wanted people to test his new code.
20:43 -!- greenlion [n=greenlio@93-80-108-19.broadband.corbina.ru] has quit Remote closed the connection
20:44 < abadger1999> lmacken: If you're still around, did you get a chance to do that?
20:44 < ricky> Ah, so it's fully running on staging right now?
20:44 < ricky> Anybody have a URL handy?
20:44 < ricky> Ah, I guess http://mirrors.stg.fedoraproject.org/mirrorlist
20:45 < lmacken> abadger1999: nope, I haven't had time to do anything other than a basic sanity check on it
20:45 -!- mbacovsk__ [n=mbacovsk@okr2fw.topnet.cz] has quit Read error: 104 (Connection reset by peer)
20:46 < abadger1999> ricky: In the same vein, I heard you switched trac over to mod_wsgi?
20:46 < ricky> I haven't yet, actually
20:46 < abadger1999> Okay.
20:46 < ricky> I only did moin, but I want to look into trac soon
20:46 < abadger1999> Cool.
20:46 < dgilmore> mmcgrath: want to give an update on staging?
20:46 < ricky> I'll hopefully have a bit to look at it tonight or tomorrow night
20:46 < abadger1999> I want to convert the bzr web viewer to loggerhead... the new version is a wsgi app so when trac switches I try to deploy that.
20:47 < mmcgrath> dgilmore: sure, I can give a quick one.
20:47 < ricky> Nice
20:47 < abadger1999> (And then the code will actually be maintained upstream :-)
20:47 < mmcgrath> right now we have an app1.stg, app2.stg, proxy1.stg and db1.stg. The staging environment can't contact any production hardware except for /accounts/ and the infrastructure repo.
20:47 < mmcgrath> Matt is giving it a try for some actual work, I'm going to continue documenting it and will hold training at some point in the future (separate from the puppet training)
20:48 < dgilmore>
20:48 < lmacken> excellent
20:48 -!- mbacovsk__ [n=mbacovsk@okr2fw.topnet.cz] has joined #fedora-meeting
20:48 < dgilmore> when is puppet training again?
20:48 < mmcgrath> haven't schedule it yet.
20:48 < mmcgrath> its coming though
20:48 < dgilmore> ok
20:49 < dgilmore> If somoene wants to help with some things. I wantto design a way to load balance koji
20:49 < SmootherFrOgZ> dgilmore: go on
20:50 < ricky> As in kojihub?
20:50 * ricky wonders what kind of locking would need to happen
20:50 < dgilmore> ricky: locking is all in the db
20:50 < ricky> Oh, cool.
20:51 < dgilmore> SmootherFrOgZ: either setting up haproxy, pound, or something like that to do load balancing
20:51 -!- warren [n=warren@redhat/wombat/warren] has joined #fedora-meeting
20:51 < ricky> Oh, so it's 100% doing the balancing (as in, the code doesn't really need to change?)
20:51 < dgilmore> so we can do pretty much all maintainence without koji going down. coping with spikes. taking a box off to test something etc
20:52 < ricky> So would we build a separate proxy box just for koji and put it in front of koji1 and koji2 then?
20:52 < dgilmore> ricky: yeah.
20:52 < SmootherFrOgZ> dgilmore: interesting, haproxy is pretty good
20:52 < dgilmore> ricky: or pound would need two boxes
20:53 < ricky> So both pound and haproxy would handle SSL stuff easily, right?
20:53 < dgilmore> ricky: it should. we would need to test it first.
20:53 < ricky> Yeah.
20:54 < SmootherFrOgZ> yeah haproxy handle that well
20:54 < ricky> I wonder if we'd want koji + a builder in staging
20:54 < dgilmore> ricky: we should.
20:54 < ricky> But that sounds really complicated to setup
20:54 < dgilmore> i do my testing of koji on sparc.koji.fedoraproject.org
20:55 < dgilmore> mostly i want to remove the spof in the buildsys
20:55 < dgilmore> so i dont know how well haproxy would help
20:55 < ricky> Hm.
20:55 < dgilmore> since it would replace the spof
20:56 < dgilmore> pound having two boxes. one active, one passive would do it
20:56 < dgilmore> but it needs thought and testing
20:56 -!- wfp [n=wfp5p@viridian.itc.Virginia.EDU] has quit "Leaving"
20:56 < dgilmore> also something that I need help with is dogtag
20:57 < dgilmore> getting it to build in mock. so we can look at migrating the CA to it
20:57 < dgilmore> thats going to need alot of testing.
20:57 < ricky> Yeah, I'll try to ping you about that when I have some more itme
20:57 < ricky> **time
20:57 < mmcgrath> dgilmore: if we're going to load balance it we should probably just put them behind the hardware firewall we have at RH, if not maybe just do heartbeat.
20:57 < dgilmore> when we switch we have to do some funkyness to migrate from the existing CA
20:58 < dgilmore> mmcgrath: id be ok with either
20:58 < mmcgrath> I'm not sure we want to put koji behind our proxy farm thats shared with the other webservers, and doing a dedicated frontend would take our koji server from 1 to 4 (2 koji, 2 proxy) which we could do but I think its probably easier just to use the balancer thats there.
20:58 < dgilmore> mmcgrath: though id kinda prefer load balancing
20:58 * ValHolla wonders if you use pound the setup could be used for more then just koji
20:59 -!- mccann [n=jmccann@] has quit "See ya"
20:59 < dgilmore> ValHolla: we try and keep buildsys seperate from everything else
20:59 < ricky> I guess we don't have very much control over the load balancer between fedoraproject.org and proxy1/2?
20:59 < mmcgrath> dgilmore: yeah, use the front end load balancer then. we'll have two backend servers and the balancer on front.
20:59 < mmcgrath> ricky: yeah its all by request
20:59 < dgilmore> mmcgrath: works for me
21:00 < mmcgrath> really though its been pretty solid so I haven't complained about it.
21:00 < ValHolla> digilmore: yeah I guess that is a good idea:/
21:00 -!- ldimaggi___ [n=ldimaggi@nat/redhat/x-6a2313ae23860d09] has quit "Leaving"
21:00 < dgilmore> so thats some things for thought
21:00 < dgilmore> Anyone have anything else to talk about?
21:00 -!- mbacovsk_ [n=mbacovsk@nat/redhat/x-574ae97d3509588a] has quit Connection timed out
21:00 < dgilmore> if not ill wrap up in 30
21:01 < dgilmore> 20
21:01 < dgilmore> 10
21:01 -!- brothers [n=brothers@66-234-43-147.nyc.cable.nyct.net] has left #fedora-meeting []
21:01 < dgilmore> --meeting wrap-- thanks all
Fedora-infrastructure-list mailing list

Thread Tools

All times are GMT. The time now is 07:05 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org