Infrastructure/Meetings/2007-01-25

= Meeting of 2007-01-25 =


 * Time shown in EST

15:02 Everyone ready ? 15:02 fer sure 15:03 abadger1999, f13, c4chris|w, iWolf, mdomsch, rordway, skvidal, tibbs: ping 15:03 yes? 15:03 What's up? 15:04 dgilmore: hi. May be in/out. Another meeting starting soon 15:04 meeting 15:04 f13: ok 15:04 in 15:04 Sorry, which meeting is this? 15:04 infrastructure 15:04 tibbs: maybe we dont need you :D 15:05 OK, never been to an infrastructure meeting before. 15:05 abadger1999: whats the status of the package DB 15:05 tibbs: out of habit i pinged you 15:06 lmacken: Firewalls 15:06 <@mmcgrath> macken won't be here, neither will iwolf. 15:06 bahh 15:07 <@mmcgrath> should be a quick meeting. 15:07 * mmcgrath is back 15:07 mmcgrath: were done i think 15:07 all the people on things are not here 15:07 Free chat time 15:07 <@mmcgrath> hah! 15:08 <@mmcgrath> I wouldn't mind doing a bit of planning for the hackfest. 15:08 :) 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 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 jcmoore, from where? 15:09 <@mmcgrath> dgilmore: what do you think about that? 15:09 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 warren, you talkin' to me? 15:09 mmcgrath: if we vpn in and give them access 15:09 -!- dbewley [n=dlbewley@tofu.lib.ucdavis.edu]  has joined #fedora-admin 15:09 ppc we are stuck with out 3 boxes 15:10 ppc1  could use more ram 15:10 <@mmcgrath> yeah, ppc's wer'e limited on. 15:10 mmcgrath: ppc1 has 2gb ram ppc2 and 3 have 4gb 15:10 all are quad cpu boxes 15:10 * mmcgrath notes the xen boxes have way more ram than I thought. 15:10 it would also help if there were a local mirror on the hackfest lan that people could point mock to 15:10 mmcgrath: they have alot 15:11 jcollie: i believe that will be there 15:11 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 so it may not be in the same room, but it'll be pretty close 15:11 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 mmcgrath: its not critical would help a little 15:12 mmcgrath: yeah 15:12 <@mmcgrath> Ok, so here's the high level things I'd like to make sure get accomplished. 15:12 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 maybe i should try the kvm thing 15:12 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 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 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 dgilmore, I can. 15:13 warren: thanks 15:14 mmcgrath, blade + NAS would be the ideal I guess. 15:14 warren: yep 15:14 mmcgrath, what does #1 above entail? 15:14 <@mmcgrath> blade + SAN would be ideal. 15:14 <@mmcgrath> blade + NAS would be affordable. 15:14 then we could hot swap guests if need be 15:14 warren, that's me :-) 15:14 <@mmcgrath> mdomsch: what do you think? 15:14 I've got a really good start - learning turbogears and all as I go 15:14 glad to talk about there 15:15 What is the goal of the mirror project? 15:15 what works: creating sites and site admins, hosts 15:15 warren, there's like 5 things 15:15 1) let mirror admins tell us they exist 15:15 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 3) auto-generate yum and download.fp.o redirects based on the data in the database 15:15 mdomsch, ah... I suppose that would be much better than the ftp@redhat.com queue which sometimes takes forever to act. 15:16 4) get the rsync ACLs to take effect directly from the DB rather than waiting on ftp@redhat.com 15:16 :-) 15:16 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 5) no more manual text lists to maintain 15:16 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 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 pls resend to me 15:17 no problem doing so 15:17 <@mmcgrath> k. 15:18 mmcgrath: to me, too. 15:18 <@mmcgrath> k 15:18 b/c this will impact a lot of mirrors, I think 15:18 * mdomsch hates those "will not export" clauses 15:18 as there's no good software to prevent it 15:18 indeed 15:18 < daMaestro> yeah? what do you do? blacklist IP blocks? 15:18 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 so, I need a lot of python yet 15:20 but will make good progress next week and at the hackfest 15:20 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 is there a list of these smaller projects out there? 15:21  * symerian ears small projects words ;) 15:21 hears instead 15:21 damn 15:21 anyway 15:22 account system / turbogears integration ? 15:22 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 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 if the new account system isn't online by test3 15:23 I'll need to plug into the existing account system 15:23 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 mmcgrath, that'd be mgalgoci, 15:23 who will be there 15:24 jcollie, would be grateful for the help 15:24 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 he responded to my question about preferred licenses 15:25 I'm pushing my stuff through my mgmt chain / legal right now 15:25 but am leaning towards MIT/X11 15:25 <@mmcgrath> k 15:25 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 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 and it's fun for me to learn TG 15:26 <@mmcgrath> mdomsch: isn't TG awesome? 15:27 mmcgrath, yeah, but it's a bit of a learning curve for someone not too familiar with python 15:27 but learning lots every night 9pm-1am 15:27 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 thl is herding package reviewers 15:27 <@mmcgrath> k 15:27 moo 15:27 <@mmcgrath> I noticed that. 15:27 mdomsch, any resource you recommend for learning TG? 15:27 there's quite a few to do 15:27 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 they don't work 15:28 mmcgrath: yum hackfest going on - I was working on some plans for that right now 15:28 mmcgrath: what do you need? 15:28 -!- japj [n=japj@a82-95-0-148.adsl.xs4all.nl] 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 it goes chapter 1, 2, 3, 5, 16, redo all work done following 3,4,5 15:28 skvidal: oooh, yum hackfest. I wanted to poke your brain a bit about a situation we're seeing with obsoletes adn multilib 15:29 f13: obsoletes don't know about arch 15:29 <@mmcgrath> crap gotta run again. 15:29 <@mmcgrath> meeting over? 15:29 sounds good 15:29 mmcgrath, sure. 15:30 yup waiting for the list ;) 15:31 did anyone make an irc log? 15:32 japj: i do 15:32 symerian: can you make it available somewhere? 15:32 <@mmcgrath> Sorry guys --- MEETING END --