Meeting of 2007-01-25
*** Time shown in EST
15:02 < dgilmore> Everyone ready ?
15:02 < jcollie> fer sure
15:03 < dgilmore> abadger1999, f13, c4chris|w, iWolf, mdomsch, rordway, skvidal, tibbs: ping
15:03 < skvidal> yes?
15:03 < tibbs> What's up?
15:04 < f13> dgilmore: hi. May be in/out. Another meeting starting soon
15:04 < dgilmore> meeting
15:04 < dgilmore> f13: ok
15:04 < mdomsch> in
15:04 < tibbs> Sorry, which meeting is this?
15:04 < dgilmore> infrastructure
15:04 < dgilmore> tibbs: maybe we dont need you :D
15:05 < tibbs> OK, never been to an infrastructure meeting before.
15:05 < dgilmore> abadger1999: whats the status of the package DB
15:05 < dgilmore> tibbs: out of habit i pinged you
15:06 < dgilmore> lmacken: Firewalls
15:06 <@mmcgrath> macken won't be here, neither will iwolf.
15:06 < dgilmore> bahh
15:07 <@mmcgrath> should be a quick meeting.
15:07 * mmcgrath is back
15:07 < dgilmore> mmcgrath: were done i think
15:07 < dgilmore> all the people on things are not here
15:07 < dgilmore> Free chat time
15:07 <@mmcgrath> hah!
15:08 <@mmcgrath> I wouldn't mind doing a bit of planning for the hackfest.
15:08 < dgilmore> :) we should work out what we want to achieve
15:08 <@mmcgrath> I'd also like to encourage anyone who's interested but can't make it to Boston to be online and available.
15:08 -!- daMaestro [n=jon@fedora/damaestro] has joined #fedora-admin
15:08 < jcollie> mmcgrath, there was something in the fesco meeting about adding some build hosts during the hackfest
15:08 <@mmcgrath> Here are my high level goals.
15:09 <@mmcgrath> Thats true.
15:09 < warren> jcmoore, from where?
15:09 <@mmcgrath> dgilmore: what do you think about that?
15:09 < dgilmore> mmcgrath: we can add x86_64 xen guests
15:09 <@mmcgrath> we could stick a builder on each of the xen boxes and allow people to run mock from there if we wanted.
15:09 < jcollie> warren, you talkin' to me?
15:09 < dgilmore> mmcgrath: if we vpn in and give them access
15:09 -!- dbewley [email@example.com] has joined #fedora-admin
15:09 < dgilmore> ppc we are stuck with out 3 boxes
15:10 < dgilmore> ppc1 could use more ram
15:10 <@mmcgrath> yeah, ppc's wer'e limited on.
15:10 < dgilmore> mmcgrath: ppc1 has 2gb ram ppc2 and 3 have 4gb
15:10 < dgilmore> all are quad cpu boxes
15:10 * mmcgrath notes the xen boxes have way more ram than I thought.
15:10 < jcollie> it would also help if there were a local mirror on the hackfest lan that people could point mock to
15:10 < dgilmore> mmcgrath: they have alot
15:11 < dgilmore> jcollie: i believe that will be there
15:11 < mdomsch> remember, BU has their own local mirror too
15:11 <@mmcgrath> I think we will have an actual budget, any time something like that comes up in the future send it to me. I have no idea what the budget is or what we have to do to spend
it but if more ram can be put to use, let me know.
15:11 < mdomsch> so it may not be in the same room, but it'll be pretty close
15:11 < jcollie> oh yeah... i guess i was thinking that this was at some random hotel...
15:11 <@mmcgrath> I say we keep in mind that we can build xen boxes for hackfest as needed, but not allocate anything until we get there and a clear need is presented.
15:11 <@mmcgrath> It only takes a fe minutes to build a box.
15:11 < dgilmore> mmcgrath: its not critical would help a little
15:12 < dgilmore> mmcgrath: yeah
15:12 <@mmcgrath> Ok, so here's the high level things I'd like to make sure get accomplished.
15:12 < dgilmore> mmcgrath: we need to ping stacy again i still cant acces xen4 through cyclades
15:12 <@mmcgrath> 1) a solid mirror milestone (mdomsch)
15:12 < dgilmore> maybe i should try the kvm thing
15:12 < warren> mmcgrath, I think for the future we should try to get a blade center.
15:12 <@mmcgrath> 2) A clear path for the future of the packaging database (abadger1999)
15:13 < warren> mmcgrath, with all useful services on guests, the boxes themselves matter less. Having more of them with the ability to expand quickly would save us in the long run.
15:13 <@mmcgrath> 3) Working building blocks for our new accounts system (only needs to be proof of concept (mmcgrath)
15:13 < dgilmore> is someone going to be around tonightto walk me through using the kvm software?
15:13 <@mmcgrath> warren: I agree. I'm going to be meeting with people about that.
15:13 < warren> dgilmore, I can.
15:13 < dgilmore> warren: thanks
15:14 < warren> mmcgrath, blade + NAS would be the ideal I guess.
15:14 < dgilmore> warren: yep
15:14 < warren> mmcgrath, what does #1 above entail?
15:14 <@mmcgrath> blade + SAN would be ideal.
15:14 <@mmcgrath> blade + NAS would be affordable.
15:14 < dgilmore> then we could hot swap guests if need be
15:14 < mdomsch> warren, that's me :-)
15:14 <@mmcgrath> mdomsch: what do you think?
15:14 < mdomsch> I've got a really good start - learning turbogears and all as I go
15:14 < mdomsch> glad to talk about there
15:15 < warren> What is the goal of the mirror project?
15:15 < mdomsch> what works: creating sites and site admins, hosts
15:15 < mdomsch> warren, there's like 5 things
15:15 < mdomsch> 1) let mirror admins tell us they exist
15:15 < mdomsch> 2) let mirror admins update their own data in the database
15:15 -!- mmcgrath changed the topic of #fedora-admin to: mirrors, its whats for dinner.
15:15 < mdomsch> 3) auto-generate yum and download.fp.o redirects based on the data in the database
15:15 < warren> mdomsch, ah... I suppose that would be much better than the firstname.lastname@example.org queue which sometimes takes forever to act.
15:16 < mdomsch> 4) get the rsync ACLs to take effect directly from the DB rather than waiting on email@example.com
15:16 < mdomsch> :-)
15:16 < warren> yay!
15:16 < daMaestro> is 3) based on geoip and server "freshness"
15:16 <@mmcgrath> mdomsch: did you get my email about the 'cla' type page?
15:16 < mdomsch> 5) no more manual text lists to maintain
15:16 < mdomsch> mmcgrath, don't think so
15:17 < daMaestro> i hate getting download.fp.o sending me off to NZ or something crazy like that
15:17 <@mmcgrath> When having new users sign up to be a mirror admin, make room for a page requesting something like a CLA.
15:17 < mdomsch> mmcgrath, uh, why?
15:17 <@mmcgrath> Basically I think we're going to have our mirror admins agree to a "will not export" clause or something.
15:17 * mmcgrath notes this is coming from legal.
15:17 <@mmcgrath> I don't know if anything will come of it or not.
15:17 < mdomsch> pls resend to me
15:17 < mdomsch> no problem doing so
15:17 <@mmcgrath> k.
15:18 < skvidal> mmcgrath: to me, too.
15:18 <@mmcgrath> k
15:18 < skvidal> b/c this will impact a lot of mirrors, I think
15:18 * mdomsch hates those "will not export" clauses
15:18 < mdomsch> as there's no good software to prevent it
15:18 < skvidal> indeed
15:18 < daMaestro> yeah? what do you do? blacklist IP blocks?
15:18 < mdomsch> geoip isn't good enough, and doesn't matter if it's an Iranian-controlled person sitting in .eu
15:19 < daMaestro> i've worried about that with the fedora unity torrent
15:19 <@mmcgrath> Its not something we can enforce I don't think.
15:19 <@mmcgrath> anyway....
15:19 < daMaestro> with DHT .. there is no real way for us to shutoff users.
15:19 < mdomsch> so, I need a lot of python yet
15:20 < mdomsch> but will make good progress next week and at the hackfest
15:20 < mdomsch> and hope to have a trial run by test2, live by test3
15:20 <@mmcgrath> oh a,d 4) finish the postfix server.
15:20 <@mmcgrath> So those are the 4 main things I'd really like to see actually make progress.
15:20 <@mmcgrath> Anyone have anything else to add? I have a bunch of smaller projects that I'd like to get out.
15:21 < jcollie> is there a list of these smaller projects out there?
15:21 * symerian ears small projects words ;)
15:21 < symerian> hears instead
15:21 < symerian> damn
15:21 < symerian> anyway
15:22 < mdomsch> account system / turbogears integration ?
15:22 < jcollie> mdomsch, do you have a list of things that need doing for the mirror management?
15:22 <@mmcgrath> mdomsch: thats my number 3
15:22 * mdomsch hasn't loked at abadger1999's sqlalchemy work yet
15:22 <@mmcgrath> There's a list out there that was on the list, I'm going to send a follow up lost soon.
15:22 < mdomsch> jcollie, I'll make such a list, it's all in my head right now
15:23 <@mmcgrath> mdomsch: has the initial mirrors guy ever gotten back in touch with you? I haven't seen him in a month or two
15:23 < mdomsch> if the new account system isn't online by test3
15:23 < mdomsch> I'll need to plug into the existing account system
15:23 < jcollie> mdomsch, mmcgrath great... every now and then i have a few hours to hack so it's nice to have a list like that
15:23 < mdomsch> mmcgrath, that'd be mgalgoci,
15:23 < mdomsch> who will be there
15:24 < mdomsch> jcollie, would be grateful for the help
15:24 < mdomsch> mmcgrath, oh, you're thinking about farshad
15:24 <@mmcgrath> Yeah, thats the guy.
15:24 <@mmcgrath> farshad.
15:24 <@mmcgrath> any word from him?
15:24 < mdomsch> he responded to my question about preferred licenses
15:25 < mdomsch> I'm pushing my stuff through my mgmt chain / legal right now
15:25 < mdomsch> but am leaning towards MIT/X11
15:25 <@mmcgrath> k
15:25 < mdomsch> he was fine with whichever, I've nuked nearly all of his code by now :-)
15:25 <@mmcgrath> heh
15:26 <@mmcgrath> k, well thank you for taking over the project. It'll be a great thing to get out there... seriously.
15:26 < mdomsch> indeed
15:26 <@mmcgrath> So as far as fudcon goes I know a lot of you will be doing non-infrastructure things, I'm going to try to coordinate as much as possible.
15:26 < mdomsch> and it's fun for me to learn TG
15:26 <@mmcgrath> mdomsch: isn't TG awesome?
15:27 < mdomsch> mmcgrath, yeah, but it's a bit of a learning curve for someone not too familiar with python
15:27 < mdomsch> but learning lots every night 9pm-1am
15:27 < warren> mmcgrath, I'm going to be pretty much focused on package reviews, but I'm interested in parts of account system and TG stuff.
15:27 < mdomsch> thl is herding package reviewers
15:27 <@mmcgrath> k
15:27 < skvidal> moo
15:27 <@mmcgrath> I noticed that.
15:27 < warren> mdomsch, any resource you recommend for learning TG?
15:27 < mdomsch> there's quite a few to do
15:27 < mdomsch> warren, I've got the book, it's good, but don't use their examples verbatim
15:27 <@mmcgrath> skvidal: what are your plans for hackfest? Is your time pretty well spoken for?
15:27 < mdomsch> they don't work
15:28 < skvidal> mmcgrath: yum hackfest going on - I was working on some plans for that right now
15:28 < skvidal> mmcgrath: what do you need?
15:28 -!- japj [firstname.lastname@example.org] has joined #fedora-admin
15:28 <@mmcgrath> warren: there's a bit of a learning curve, but once you've got it... you've got it.
15:28 <@mmcgrath> I'm just trying to think of what resources I'll have during hackfest.
15:28 < mdomsch> it goes chapter 1, 2, 3, 5, 16, redo all work done following 3,4,5
15:28 < f13> skvidal: oooh, yum hackfest. I wanted to poke your brain a bit about a situation we're seeing with obsoletes adn multilib
15:29 < skvidal> f13: obsoletes don't know about arch
15:29 <@mmcgrath> crap gotta run again.
15:29 <@mmcgrath> meeting over?
15:29 < skvidal> sounds good
15:29 < warren> mmcgrath, sure.
15:30 < symerian> yup waiting for the list ;)
15:31 < japj> did anyone make an irc log?
15:32 < symerian> japj: i do
15:32 < japj> symerian: can you make it available somewhere?
15:32 <@mmcgrath> Sorry guys ------------------------------- MEETING END --------------------------------------------------