September 2007 - infrastructure - Fedora mailing-lists

Planet for Fedora Users
by Rahul Sundaram 28 Sep '07

28 Sep '07

Hi I like this idea. http://blog.thedebianuser.org/?p=239 Maybe we could have planet.fedoraproject.org/users or users.planet.fedoraproject.org similar to the art feed. Rahul

2 2

Meeting Log - 2007-09-27
by Mike McGrath 28 Sep '07

28 Sep '07

2007-09-27T13:54:22 *** mmcgrath changes topic to "Infrastructure -- Role Call" 2007-09-27T13:54:37 * mmcgrath notes jeremy and warren are about to drop 2007-09-27T13:54:50 <warren> drop? 2007-09-27T13:55:07 <lmacken> heheh 2007-09-27T13:55:09 * couf listens in 2007-09-27T13:55:25 * jeremy is here 2007-09-27T13:55:27 <warren> oh you're doing it on asterisk? 2007-09-27T13:55:30 <mmcgrath> abadger1999: dgilmore: f13 jeremy lmacken mbonnet paulobanon jima: ping 2007-09-27T13:55:32 <warren> I odn't have my headset today =( 2007-09-27T13:55:43 <abadger1999> pong 2007-09-27T13:55:46 <jima> oops 2007-09-27T13:55:49 <jima> hey folks :) 2007-09-27T13:55:53 <mmcgrath> warren: nope, its just that every meeting you and jeremy seem to drop within seconds of me changing the topic. Its happened for the last 4 or 5 meetings. 2007-09-27T13:55:53 <f13> mmcgrath: I'm not really here, leaving shortly 2007-09-27T13:56:00 <mmcgrath> f13: k. 2007-09-27T13:56:13 <mmcgrath> skvidal: ping 2007-09-27T13:56:15 <warren> mmcgrath, I wonder if something in IRC causes the proxy to crash 2007-09-27T13:56:17 * lmacken is here 2007-09-27T13:56:30 * skvidal is here 2007-09-27T13:56:38 <skvidal> when do jeremy and warren drop off? 2007-09-27T13:56:50 <mmcgrath> skvidal: almost every meeting right after we start. 2007-09-27T13:56:52 <lmacken> i think this was the first week it hasn't happened ;) 2007-09-27T13:57:01 <warren> mmcgrath, how many weeks in a row? 2007-09-27T13:57:02 <mmcgrath> its very amusing. 2007-09-27T13:57:10 <skvidal> at least a couple of months 2007-09-27T13:57:13 <warren> wow 2007-09-27T13:57:14 <mmcgrath> warren: many, I'm not sure the exact number. 2007-09-27T13:57:29 <skvidal> it became a running gag 2007-09-27T13:57:30 <mmcgrath> we always laugh at it but you two probably don't see it because you're off :) 2007-09-27T13:57:34 <warren> mmcgrath, perhaps our network has a hiccup every week at the same time 2007-09-27T13:57:44 <skvidal> warren: it's a bit uncanny, though. 2007-09-27T13:57:49 <mmcgrath> maybe a clear xlate or something. alrighty, lets get going. 2007-09-27T13:57:51 <jeremy> I got a core file out of dircproxy, but it was basically garbage 2007-09-27T13:57:56 <warren> ah! 2007-09-27T13:58:01 <warren> jeremy, I use dircproxy to 2007-09-27T13:58:02 <warren> too 2007-09-27T13:58:11 <lmacken> our meetings have found an exploit! 2007-09-27T13:58:14 * dgilmore is here 2007-09-27T13:58:19 *** mmcgrath changes topic to "Infrastructure -- Tickets" 2007-09-27T13:58:34 <mmcgrath> .tiny https://hosted.fedoraproject.org/projects/fedora-infrastructure/query?statu… 2007-09-27T13:58:40 <mmcgrathbot> mmcgrath: http://tinyurl.com/yth34b 2007-09-27T13:59:03 <dgilmore> mmcgrathbot: they just dont like you ;) 2007-09-27T13:59:06 <mmcgrath> .title https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/154 2007-09-27T13:59:08 <mmcgrathbot> mmcgrath: #154 (DNS) - Fedora Infrastructure - Trac 2007-09-27T13:59:10 <mmcgrath> could be :) 2007-09-27T13:59:42 <mmcgrath> So the first ticket about DNS stuff is on going. I've got the ttl for fedoraproject.org set to 60 seconds right now. It seems fairly responsive. I've been able to direct traffic to the tummy site or PHX site pretty much at will. 2007-09-27T14:00:13 <mmcgrath> We've been balancing between the two for the last week or so without issue. 2007-09-27T14:00:28 <mmcgrath> The last two pieces of that are to get the rest of the hosts on the VPN and setup our secondary VPN site. 2007-09-27T14:00:29 <jima> cool 2007-09-27T14:01:01 <mmcgrath> basically we're going to have two vpn servers one in PHX one, as a backup, elsewhere. If one fails, the others will automatically connect to the other one within 20 seconds. 2007-09-27T14:01:24 <mmcgrath> I'm going to see what problems that will cause, like will TCP be interupted/ etc. 2007-09-27T14:01:31 <jima> any idea where "elsewhere" will be? 2007-09-27T14:01:31 <mmcgrath> any questions comments on that before we move on? 2007-09-27T14:02:01 <mmcgrath> jima: initially it will be at the denver colo with tummy.com though I'll be moving it somewhere else if we get some space for it. 2007-09-27T14:02:10 * mmcgrath notes - http://mmcgrath.fedorapeople.org/GlobalNetwork.png 2007-09-27T14:02:22 <mmcgrath> Ok, next ticket. 2007-09-27T14:02:30 *** ricky has joined #fedora-meeting 2007-09-27T14:02:35 * ricky is here now. 2007-09-27T14:02:38 <mmcgrath> .title https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/170 2007-09-27T14:02:40 <mmcgrathbot> mmcgrath: #170 (Hosting respins) - Fedora Infrastructure - Trac 2007-09-27T14:02:48 <mmcgrath> ricky: yo 2007-09-27T14:03:20 <mmcgrath> So I think there's been some miscommunication or me just not paying attention to whats going on with the respins. We've been talking about it a bit on f-a-b. 2007-09-27T14:03:20 <dgilmore> mmcgrath: do we have space for hosting random spins? 2007-09-27T14:03:24 <jima> mmcgrath: careful, that map is a tad bit depressing. ;) 2007-09-27T14:03:28 <lmacken> has anyone looked at wevisor, and what it offers ? 2007-09-27T14:03:28 *** JSchmitt has quit IRC 2007-09-27T14:04:11 <mmcgrath> dgilmore: we have spce for hosting *some* space now - 242G 123G 107G 54% /srv 2007-09-27T14:04:13 <abadger1999> lmacken: wevisor needs some work. The RH interns left and it's been stalled. 2007-09-27T14:04:23 <mmcgrath> I'm working to aquire much more space in the 1T range. 2007-09-27T14:04:24 <abadger1999> lmacken: Otherwise, its mostly done. 2007-09-27T14:04:42 <VileGent> fyi fedora unity has respins to go now 2007-09-27T14:04:44 *** LetoTo has left #fedora-meeting 2007-09-27T14:04:48 <mmcgrath> jeremy: is this a releng thing only or is this something infrastructure could or should be helping out with? 2007-09-27T14:05:07 <dgilmore> mmcgrath: can that grow? 2007-09-27T14:05:11 <mmcgrath> I get emails from time to time about it and the whole process is a bit murky in my mind. 2007-09-27T14:05:17 <abadger1999> lmacken: It's current goal, though is just to get you a kickstart file. You still have to pass the kickstart through pungi and other tools to get a spin out. 2007-09-27T14:05:29 <lmacken> abadger1999: ah, gotcha. 2007-09-27T14:05:30 <dgilmore> mmcgrath: i dont want to see us get to the point where we have to say sorry we cant host things anymore 2007-09-27T14:05:33 <mmcgrath> dgilmore: the 1T server or the 242G we have? 2007-09-27T14:05:50 <dgilmore> mmcgrath: well for future needs? 2007-09-27T14:06:01 <mmcgrath> dgilmore: I say we do a trial (we talked about it a bit on FAB) with first come first surve until we get a more dedicated source. 2007-09-27T14:06:01 <dgilmore> i know its not accurate but really both 2007-09-27T14:06:03 <nirik> could jigdo help here? 2007-09-27T14:06:50 <dgilmore> mmcgrath: ok. as long as we communicate that 2007-09-27T14:07:25 <mmcgrath> nirik: does anyone really use jigdo though? - http://www.google.com/trends?q=jigdo%2C+bittorrent 2007-09-27T14:07:32 <mmcgrath> jeremy: ? 2007-09-27T14:07:42 <nirik> debian does a lot. 2007-09-27T14:07:45 <jeremy> mmcgrath: infrastructure definitely could help. there are more people really workin gon infrastructure than rel-eng 2007-09-27T14:07:56 <jeremy> nirik: jigdo can't help with live images at all 2007-09-27T14:08:02 <EvilBob> People don't use jigdo because it is not offered as an atractive option 2007-09-27T14:08:04 <jeremy> nirik: which are the "interesting" things to provide really 2007-09-27T14:08:15 <VileGent> jeremy, ?? 2007-09-27T14:08:16 <nirik> yeah, I guess thats true. ;( 2007-09-27T14:08:36 <VileGent> nm 2007-09-27T14:08:42 <mmcgrath> jeremy: so my concern here is to make sure that this process doesn't rely soley on just you and jesse (you all have enough to worry about with the official spins) 2007-09-27T14:08:44 <EvilBob> people sit an leach off of BT and act like they are doing the world a service by capping their upstream 2007-09-27T14:08:53 <nirik> since the livecd is already installed image, it can't get a jigdo template to get packages from different mirrors. ;( 2007-09-27T14:08:57 <mmcgrath> Is there things we can do now or in the next couple of weeks? 2007-09-27T14:09:00 <jeremy> nirik: correct 2007-09-27T14:09:01 <mmcgrath> s/Is/Are/ 2007-09-27T14:09:19 <jeremy> mmcgrath: it's something that we could do to get more people involved with rel-eng, though. and in a pretty easy way 2007-09-27T14:09:48 <jeremy> mmcgrath: to be honest, even if it's right now still just me doing the extra spins, I'd still like to get the infrastructure in place (and have been meaning to send a mail for, oh, six weeks now but haven't gotten around to it) 2007-09-27T14:09:50 <mmcgrath> we do have a sysadmin-torrent group already for building torrents. We could delegate some of that to rel-eng. 2007-09-27T14:10:15 <jeremy> mmcgrath: I might say introduce a new spinmaster group or some such. plus, that sounds cool :) 2007-09-27T14:10:19 <mmcgrath> My main question is actually building the images from a ks and who can be trusted to do that? 2007-09-27T14:10:24 <mmcgrath> heh 2007-09-27T14:10:31 *** glezos has joined #fedora-meeting 2007-09-27T14:10:37 * glezos jumps in 2007-09-27T14:10:38 <dwmw2_gone> we should remember to seed the trackers at http://www.sixxs.net/tools/tracker/ when we release 2007-09-27T14:11:00 <jima> dwmw2_gone: yeah, someone with ipv6 needs to remember to do that ;) 2007-09-27T14:11:02 <jeremy> the big thing is that the config needs to just be checked over -- you can have %post --nochroot, and it's running as root, so .... 2007-09-27T14:11:16 * jima finds himself wishing, yet again, that he had native ipv6. 2007-09-27T14:11:30 <dgilmore> jima: as do i 2007-09-27T14:11:35 <dwmw2_gone> jima: doesn't take much to arrange it 2007-09-27T14:11:48 <dwmw2_gone> just set up a tunnel from sixxs.net 2007-09-27T14:11:49 <jeremy> but as looking over and doing sign off of the config is needed anyway, that shouldn't be a big deal 2007-09-27T14:11:50 <jima> dwmw2_gone: i did say "native" :P 2007-09-27T14:11:50 <mmcgrath> dwmw2_gone: add that to - http://fedoraproject.org/wiki/Infrastructure/SOP/Release if you get a moment 2007-09-27T14:12:01 <dwmw2_gone> mmcgrath: will do 2007-09-27T14:12:08 <dgilmore> dwmw2_gone: i have a tunnel. just not native support from my isp 2007-09-27T14:12:10 <mmcgrath> jeremy: Do we have any place on the wiki for this stuff or should I create one? 2007-09-27T14:12:36 * jima has numerous tunnels 2007-09-27T14:12:43 <dwmw2_gone> jima: even I don't have "native" at the moment -- I have a tunnel from my ISP. The ISP does support native IPv6 on PPP on the DSL, but I'm temporarily using a non-Linux box as a router (since I blew up the USB ports on my telephone exchange/router) 2007-09-27T14:12:45 <jeremy> mmcgrath: no place at the moment... if you want to start one, I can help flesh it out 2007-09-27T14:13:06 <mmcgrath> jeremy: k, I'll get some initial stuff setup. Including the spins.fp.o site. 2007-09-27T14:13:23 <jima> dwmw2_gone: even a tunnel from your immediate upstream isp is better than from halfway across the internet :( 2007-09-27T14:13:28 <dwmw2_gone> mmcgrath: or on the TorrentRelease page which that page refers to? 2007-09-27T14:13:30 <jeremy> mmcgrath: sounds good. and hopefully the board will agree with us :-) 2007-09-27T14:13:33 <mmcgrath> jeremy: the other question is who should approve secondary spins. I think FESCo would be good for that but some think board. What do you think? 2007-09-27T14:13:33 <dwmw2_gone> oh. The TorrentRelease page doesn't exist 2007-09-27T14:13:43 <f13> why not rel-eng itself? 2007-09-27T14:14:00 <jeremy> mmcgrath: it's trademark guidelines. the board can delegate if they want, but ultimately, control of the trademark rests with the board at the moment 2007-09-27T14:14:02 <f13> there are already guidelines, releng (who knows how to validate this stuff) could just once over and make sure things are being done right. 2007-09-27T14:14:09 <f13> or that. 2007-09-27T14:14:13 <EvilBob> rel-eng +1 2007-09-27T14:15:03 <mmcgrath> so a two part approval? releng says this is technically ready and the board says "give it our trademark" ? 2007-09-27T14:15:06 <dgilmore> mmcgrath: right now the board due to what jeremy said 2007-09-27T14:15:18 <dgilmore> mmcgrath: yup 2007-09-27T14:15:25 <mmcgrath> f13: that sound reasonable to you? 2007-09-27T14:16:38 *** bklemm has joined #fedora-meeting 2007-09-27T14:17:12 <mmcgrath> Ok, we'll just go with that for now. it will likely take some time before this completely gets flushed out anyway. I'll start an initial page and point rahul at it with the exception that he be patient as we're still working it out. 2007-09-27T14:17:19 <mmcgrath> Anyone else have anything to discuss on that topic? 2007-09-27T14:17:58 <jeremy> thanks for jumping in and doing some quick thinking on it 2007-09-27T14:18:42 <dwmw2_gone> mmcgrath: http://fedoraproject.org/wiki/Infrastructure/SOP/Release?action=diff 2007-09-27T14:18:46 <f13> mmcgrath: seems fine to me. 2007-09-27T14:18:59 <mmcgrath> solid. 2007-09-27T14:19:04 * mmcgrath is glad to get moving on that. 2007-09-27T14:19:37 <mmcgrath> Ok, on to the next topic 2007-09-27T14:19:45 *** mmcgrath changes topic to "Infrastructure -- Schedule - http://fedoraproject.org/wiki/Infrastructure/Schedule" 2007-09-27T14:20:18 <mmcgrath> Corporate Sponsorship. We're getting some more interest in this but no new sponsors this week. 2007-09-27T14:20:40 <mmcgrath> I've been asking for some collaboration servers to host dedicated instances of gobby, pastebins, asterisk and a few other things we use to work together. 2007-09-27T14:21:02 <mmcgrath> I'm looking for a 1T torrent server both as an upgrade to our current server and because the current torrent is getting old. 2007-09-27T14:21:03 <lmacken> sweet 2007-09-27T14:21:18 <mmcgrath> I've also been pushing to find a new hosting server with about 1T of storage and onsite backup. 2007-09-27T14:21:45 <f13> nice 2007-09-27T14:21:47 <mmcgrath> I figure if we can get that setup and keep an off site backup in PHX hosted will be a very viable and attractive solution when combined with the Fedoraverse. 2007-09-27T14:22:00 <f13> mmcgrath: what ever happened to the blade center thing? 2007-09-27T14:22:17 <mmcgrath> f13: the last thing I heard from Jay Madison was "we'll tell you when we hear more" 2007-09-27T14:22:29 <mmcgrath> We're quickly coming up on a year of that thing just sitting on the floor. 2007-09-27T14:22:45 <jima> how big is it? 2007-09-27T14:22:58 <mmcgrath> jima: I think its more the power thats the problem, not the height. 2007-09-27T14:23:01 * mmcgrath digs up last email 2007-09-27T14:24:22 <jeremy> jima: it's a bladecenter. they're 14U iirc 2007-09-27T14:24:44 <jima> jeremy: i don't know bladecenters; that second part helps. ;) 2007-09-27T14:24:55 <jeremy> (but my RC may well not be... waaaayy too much rattling around in my head today :) 2007-09-27T14:25:28 <mmcgrath> f13: yeah, last email I got was on 9/04/07 that said "none yet, standby please. We'll let you know when we have everything with this sorted out. " 2007-09-27T14:25:42 <mmcgrath> So thats the latest. 2007-09-27T14:26:17 <mmcgrath> So anyway, thats the latest on corporate sponsorship, any questions before moving on? 2007-09-27T14:27:04 <mmcgrath> The next item on the schedule is architecture documentation. I've added a few bits to - http://git.fedoraproject.org/hosted/fedora-infrastructure.git/architecture/ 2007-09-27T14:27:06 <jeremy> jima: whoops. 7U, 14 blades. 2007-09-27T14:27:09 <mmcgrath> I haven't updated the wiki yet. 2007-09-27T14:27:18 <mmcgrath> jeremy: there you go, that sounds better :) 2007-09-27T14:27:29 <dgilmore> 20amps :) 2007-09-27T14:27:38 <jima> 20amps. hmm. 2007-09-27T14:27:42 <dgilmore> actually 25 i think 2007-09-27T14:27:50 <jeremy> also serves as a space heater for the winter 2007-09-27T14:27:51 <jima> 25 makes it a bit trickier :P 2007-09-27T14:28:02 <dgilmore> would need to double check 2007-09-27T14:28:11 <dgilmore> its 20 or 25 2007-09-27T14:28:24 <mmcgrath> ricky: I've added that image to point to a proper location :) 2007-09-27T14:29:04 <mmcgrath> Next item - SOP's - http://fedoraproject.org/wiki/Infrastructure/SOP/Bodhi 2007-09-27T14:29:17 <mmcgrath> I think lmacken's sop about bodhi is the only new one. 2007-09-27T14:29:18 <mmcgrath> thanks lmacken 2007-09-27T14:29:21 <lmacken> no prob 2007-09-27T14:29:25 <ricky> mmcgrath: Ah, thanks :) 2007-09-27T14:29:36 <lmacken> the data was already on bodhi's wiki.. i just pointed to it :) 2007-09-27T14:29:56 <mmcgrath> And as far as new sponsorship goes, no new sponsors this week though kyriakos_ has been working on learning the build system 2007-09-27T14:30:11 <mmcgrath> a bold move :) 2007-09-27T14:31:00 <jima> heh 2007-09-27T14:31:40 <mmcgrath> Ok, so thats all I've got. I'll open the floor. 2007-09-27T14:31:45 *** mmcgrath changes topic to "Infrastructure -- Open Floor" 2007-09-27T14:31:50 <mmcgrath> anyone have anything to dicsuss? 2007-09-27T14:32:08 <lmacken> I'm going to be working with dan walsh at some point soon to resolve our SELinux "situation" 2007-09-27T14:32:40 <mmcgrath> lmacken: thanks. I'm fine with using SELinux but we need to have a valid way to manage it and deploy with puppet. 2007-09-27T14:32:51 <lmacken> so if you've had any gripes /concerns with selinux in our infrastructure, please let me know so we can address it 2007-09-27T14:33:00 <lmacken> mmcgrath: ok 2007-09-27T14:33:01 <mmcgrath> skvidal: ? 2007-09-27T14:33:03 <mmcgrath> :-P 2007-09-27T14:33:25 *** rdieter has quit IRC 2007-09-27T14:33:25 <skvidal> sorry 2007-09-27T14:33:27 <skvidal> one sec 2007-09-27T14:33:38 <skvidal> what is our 'selinux situation' 2007-09-27T14:33:45 <skvidal> define it 2007-09-27T14:33:48 <mmcgrath> skvidal: just that we don't use it. 2007-09-27T14:33:51 <lmacken> yes. 2007-09-27T14:33:59 <skvidal> uh huh.... 2007-09-27T14:34:07 <skvidal> lmacken: here's what I have to say and people can be pissed at me 2007-09-27T14:34:09 <mmcgrath> and it'd be good for Fedora if we did use it. 2007-09-27T14:34:32 <lmacken> mmcgrath: ourselves too.. in that it would have helped mitigate some issues in the past 2007-09-27T14:34:39 <skvidal> If you break a server "fixing" selinux I'm going to be angry :) 2007-09-27T14:34:54 <mmcgrath> welllllll, mitigate==completely hide in the last case. 2007-09-27T14:35:16 <mmcgrath> I go back and forth about selinux, I'm rarely in the middle about it but I'm willing to give it a shot. 2007-09-27T14:35:18 <f13> lets start with permissive and work from there. 2007-09-27T14:35:25 <lmacken> mmcgrath: yeah.. but those are the two biggest concerns we're faced with that we need to resolve: 1) shit doesn't break 2) we know about selinux violations 2007-09-27T14:35:26 <f13> that's sufficiently middle. 2007-09-27T14:35:33 <skvidal> let's start with a justification for why we need it on 2007-09-27T14:35:35 <skvidal> then work from there 2007-09-27T14:35:39 <f13> we can see the denials and yell at people about them, but it wont' actually stop us from doing anything. 2007-09-27T14:35:47 <mmcgrath> skvidal: it would add security and thats a good thing. 2007-09-27T14:36:00 <skvidal> mmcgrath: I have a lot of questions about the net win we get from it 2007-09-27T14:36:02 <jima> mmcgrath: [tm] 2007-09-27T14:36:26 <skvidal> in much the same way I'm illconvinced of how things like mugshot are obviously beneficial - but then most people call me a stick-in-the-mud 2007-09-27T14:36:42 <dgilmore> skvidal: your an old fuddy duddy 2007-09-27T14:36:45 <dgilmore> but so am i 2007-09-27T14:36:56 <mmcgrath> skvidal: I agree with you on it. But having said that selinux has changed a lot since I last started looking at it. 2007-09-27T14:37:02 <jima> geez, all you fogeys 2007-09-27T14:37:06 <mmcgrath> I'm willing to give it a second shot, especially if its not me doing the work :) 2007-09-27T14:37:20 *** tibbs has quit IRC 2007-09-27T14:37:23 <jeremy> skvidal: stick-in-the-mud 2007-09-27T14:37:36 <mmcgrath> lmacken: if you and dan come back with some 20 step process we have to follow every time we change a config... we'll probably not use it. 2007-09-27T14:37:45 <mmcgrath> but if you can make it dead simple, I'm for it. 2007-09-27T14:37:49 <lmacken> mmcgrath: understandable.. the point is to make it Just Work with our setup. 2007-09-27T14:38:03 <dgilmore> lmacken: selinux needs alot of work to work with mock/koji/plague 2007-09-27T14:38:07 <f13> which may be hard, as Ii'm sure we color outside the lines a bit. 2007-09-27T14:38:14 <mmcgrath> lmacken: work closely with dan on it and we'll start with permissive mode. 2007-09-27T14:38:16 <f13> ya know, /srv/ and all 2007-09-27T14:38:21 <lmacken> mmcgrath: sounds good 2007-09-27T14:38:27 <f13> I'm not terribly interested in spending a ton of time rewriting policy 2007-09-27T14:38:36 <dgilmore> mmcgrath: i usually just put the boxes in permissive mode 2007-09-27T14:38:42 *** GeroldKa has joined #fedora-meeting 2007-09-27T14:38:44 *** fabian_a has joined #fedora-meeting 2007-09-27T14:39:00 <skvidal> lmacken: and make no modifications to production machines 2007-09-27T14:39:15 <lmacken> skvidal: definitely won't until we have a game plan 2007-09-27T14:39:42 <mmcgrath> lmacken: yeah, come up with a full plan and send it to the list before we get started. starting on the xen dom0's would probably be a good place to start. 2007-09-27T14:39:52 <lmacken> mmcgrath: will do 2007-09-27T14:40:11 <mmcgrath> Ok, if no one has anything else we'll move on. 2007-09-27T14:40:22 * dgilmore has nothing 2007-09-27T14:40:44 <mmcgrath> ok, we'll close in 30 2007-09-27T14:40:49 *** kital has quit IRC 2007-09-27T14:41:00 <mmcgrath> 15 2007-09-27T14:41:11 <mmcgrath> 5 2007-09-27T14:41:18 *** mmcgrath changes topic to "Infrastructure -- Meeting End"

3 4

Fusion io solid-state storage
by Ignacio Vazquez-Abrams 28 Sep '07

28 Sep '07

Just saw this, figured someone here might be interested. http://www.fusionio.com/ -- Ignacio Vazquez-Abrams <ivazqueznet(a)gmail.com> PLEASE don't CC me; I'm already subscribed

4 3

Intro / application for Sysadmin group. Request for sponsor
by Mark Hutchinson 26 Sep '07

26 Sep '07

I need a sponsor. My name is Mark Hutchinson. I have about 10 year experience with Linux/Unix. I am an RHCE and working toward my RHCA. My experience is primarily in Oil/Gas and ISP. ( I will attach my resume as well ) I am a contractor currently for a Large Oil/Gas company and look after local and global Linux environments. I have also worked with NIS and LDAP quite a bit over the years. I am In Calgary, Alberta Canada ( MST ) I would like to be involved in the build and admin of environments. Though testing would also interest me as well. Giving a little back would be great as I have been using Fedora and RedHat for many years as an end user. My experience related to what is needed: * Web clustering - done with HA - Failover * Building and build environments - I have built many many environments for multiple purposes * Monitoring - Nagios and BigBrother extensively * Virtualization - Mostly with VMware but learning XEN now as well. * Source control management - I would like to learn more here * Script writers and automators - Lots of standard admin scripts and notification / automation script work in bash. * Years experience. * Skills (especially those listed below in "What we are looking for"). * Specific tasks you are interested in working on. * Current Location / Time Zone. * Why you are interested in working on the team. * Resume CV (not required, but a good idea)

1 0

Introduction, what is required.
by Mark Hutchinson 26 Sep '07

26 Sep '07

Good day, I have just signed up for the FedoraProject. What is required of me at this point to introduce myself etc... Just looking for protocol on this. Also, how do I get and use my fedoraproject.org email address? Thank you, Mark

2 1

Missing kernel update (with security fix!)
by Axel Thimm 26 Sep '07

26 Sep '07

Hi, a kernel rpm with a security fix has been pushed 24h ago by bodhi, but no mirror has it yet, all donwloadX show the old kernel but a couple that give connection refused. -- Axel.Thimm at ATrpms.net

3 2

Outage Notification - 2007-09-26 04:00 UTC (DB2)
by Mike McGrath 25 Sep '07

25 Sep '07

There will be an outage starting at 2007-09-26 04:00 UTC UTC, which will last approximately 5 hours. To convert UTC to your local time, take a look at http://fedoraproject.org/wiki/Infrastructure/UTCHowto or run: date -d '2007-09-26 04:00 UTC' Affected Services: Websites Buildsystem Database Unaffected Services: CVS / Source Control DNS Mail Torrent Ticket Link: https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/163 Reason for Outage: DB2 is running out of disk space. This is a preventative maintenance to ensure that we don't completely run out. There are a couple of avenues I'm still examining but if they fail I'll be forced to rebuild db2 with more storage and that's the reason for the outage. Its possible the actual outage will only take an hour or two but just in case I've created a 5 hour block. Contact Information: Please join #fedora-admin in irc.freenode.net or respond to this email to track the status of this outage.

1 0

Builders...
by Oliver Falk 25 Sep '07

25 Sep '07

Hi! Please have a look. somethings wrong. tagNotifications are hanging around and builds are not working... I've seen some mail subject stating something about db2 outage. Might this be the reason? Thx, -of

2 3

db2 outage
by Mike McGrath 24 Sep '07

24 Sep '07

I need to work on db2 asap (space issues). I'd like this to be tomorrow night around midnight Eastern. Is there any reason we should not do this? -Mike

1 0

url rewriting for mitwi project
by ria das 24 Sep '07

24 Sep '07

Hi, I am working on the project mitwi (http://publictest1.fedora.redhat.com/mitwi/index.cgi) for SoC. I was suggested for url rewriting and so I was looking for it. For this I wrote : RewriteEngine on RewriteBase / Rewritecond %{REQUEST_URI} !^/(index|language|direct).cgi RewriteRule ^(.*)$ /direct.cgi?data=$1 [L,QSA] I tried out this, but did not get any result. So can anyone please help me out to solve the problem. Ria Das

2 1

Trying to recreate Fedora 7 Everything structure
by Imtiaz Rahi 24 Sep '07

24 Sep '07

Hi, Hope I asking question at the right place. I want to re-spin Fedora 7 with updated packages with revisor. So, I am creating a "fedora" and "updates" repository. My updates repo is fine. I am running into issues with the fedora everything repo ( http://download.fedora.redhat.com/pub/fedora/linux/releases/7/Everything/i3… ). For that I am trying to recreate the structure with whatever downloaded packages I have. Every time I run createrepo the primary.xml generates lines like this; e.g. <location href="ConsoleKit-0.2.1-2.fc7.i386.rpm"/> I am in /disk/repository/fedora/7/i386/. My current command for createrepo is $ createrepo -pd -g /disk/repository/fedora/7/i386/repodata/comps-f7.xml -o ./ Fedora/ And my directory structure look like this fedora/7/i386]$ ls -CF Fedora/ isolinux/ RPM-GPG-KEY RPM-GPG-KEY-fedora-test fedora.css README-BURNING-ISOS-en_US.txt RPM-GPG-KEY-beta RPM-GPG-KEY-rawhide GPL RELEASE-NOTES-en_US.html RPM-GPG-KEY-fedora stylesheet-images/ images/ repodata/ RPM-GPG-KEY-fedora-rawhide TRANS.TBL All the rpms are in Fedora directory. And I know from previous experience installation will fail as yum will look for the RPM in "fedora/7/i386/" whereas they are in "fedora/7/i386/Fedora" Also from the DVD's primary.xml and http://download.fedora.redhat.com/pub/fedora/linux/updates/7/i386/repodata/… I can see the file location looks like this; e.g. <location href="Fedora/ ConsoleKit-0.2.1-2.fc7.i386.rpm"/> How can I make my primary.xml file look like the one in get in CD/DVD? Please help me here. Give me some pointer. What I am doing wrong here? Thanks, Imtiaz

2 2

TurboMail patch
by Luke Macken 21 Sep '07

21 Sep '07

So bodhi has recently been having some issues with TurboMail, which seems to stop dispatching mid-push. Bodhi still 'turbomail.enqueue's the messages, but the worker threads seem to be dead. I created a new package with a patch from upstream Ticket #17 (http://trac.orianagroup.com/turbomail/ticket/17/) http://lmacken.fedorapeople.org/rpms/python-TurboMail-2.0.4-3.fc7.noarch.rpm I installed this package on releng1, so we'll see how it works during the next updates push. More testing is welcome :) luke

1 1

Introduction (Christopher Vandeventer)
by Life Frenzy 21 Sep '07

21 Sep '07

Good Day, I'm a student entering my senior year of Software Engineering in the USA, Nashville TN. I made the jump to Fedora about 6 months ago to help me with my 'weekend'-business. I'll be lurking at the meetings to keep up, and I'll offer any assistance I can, however much of this information is still beyond my grasp at the moment. I'm focusing most of my current resources on learning how to provide & maintain system services, however my future goal is in project management on open source platforms so I do tend to have a bit of the benefits and problems of 'jack of all trades.' I'm going to be working with Ambassadors lists as I'll be doing the people-person thing for a few years while I gain more experience. I currently do remote maintenance for the Novell system & network at a hospital in NYC, and have both CNA & Net+ certifications - attempting for the LPIC-1 by mid '08. I have started applying for junior level admin positions, I'll update the profile page on the status if I do receive a position supporting linux. My Fedora wiki will have my contact info & website within the next week. Thanks for giving people the chance to gain experience on Fedora through volunteering. As the years pass, I plan to work with the Ambassadors and Translators (Spanish), and possibly some developing, before focusing entirely here. I thought it would be appropriate to say Hello though, as I will be reading your communications. Van -

2 1

Infratructure Meeting Log for 2007/09/20
by Ricky Zhou 20 Sep '07

20 Sep '07

16:00:33 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Role Call 16:00:35 < mmcgrath> Who's here? 16:00:36 * ricky 16:00:36 -!- warren [i=warren@redhat/wombat/warren] has quit Remote closed the connection 16:00:37 -!- jeremy [i=katzj@nat/redhat/x-0f82d1e06695232a] has quit Remote closed the connection 16:00:39 < mmcgrath> quick before they drop 16:00:40 < ricky> Haha. 16:00:41 < mmcgrath> doah, too late. 16:00:48 < jima> oops 16:01:01 * jima here 16:01:18 * kyriakos_ (not that it really makes any difference :P) 16:01:23 < mmcgrath> skvidal: abadger1999 paulobanon f13 ivazquez ricky jima lmacken dgilmore kyriakos_ ping? 16:01:24 < londo> heh 16:01:28 < mmcgrath> londo: ping :) 16:01:30 < paulobanon> here 16:01:35 < abadger1999> pong 16:01:38 < londo> here 16:01:40 < jima> pong 16:01:40 < ivazquez> Pong. 16:02:01 < jima> (not that sets off my nick detection...maybe i should work on that) 16:02:17 < paulobanon> can we change the meeding for friday, to see if they still disconnect :D 16:02:29 < mmcgrath> paulobanon: we could :) 16:02:39 < mmcgrath> Ok, I think we have enough to get started. 16:02:46 < ricky> Or move the time :) 16:02:57 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- First tickets 16:02:59 < mmcgrath> https://hosted.fedoraproject.org/projects/fedora-infrastructure/query?statu… tus=assigned&status=reopened&group=milestone&keywords=%7EMeeting&order=priority 16:03:49 < mmcgrath> Ok, one thing I wanted to talk to everyone about is some of the architectural changes I've been planning / making. 16:04:00 < mmcgrath> Long story short we're slowly decentralizaing our infrastructure. 16:04:10 < mmcgrath> this is A) cool and B) not simple. 16:04:15 < mmcgrath> B's the part I'm worried about. 16:04:29 < mmcgrath> Basically we're adding a bunch of redundancy to our environment but also adding complexity and points of failure. 16:04:38 < jima> mmhmm 16:04:42 * daMaestro is here 16:04:45 < mmcgrath> I recently created another domain to help ease this transition, right now its public but in the future it probably won't be. 16:04:48 < mmcgrath> daMaestro: yo 16:05:15 < mmcgrath> Once complete, every machine will be able to get to every other machine via "hostname.vpn.fedoraproejct.org" once you're connected to a machine. (firewall permitting) 16:05:34 < jima> oh, neat. 16:05:42 < mmcgrath> Part of this is the vpn configuration and part of this is naming our machines. 16:05:51 -!- rdieter_away is now known as rdieter 16:06:00 -!- jeremy [i=katzj@nat/redhat/x-824cfb21e0d420e3] has joined #fedora-meeting 16:06:00 < mmcgrath> Long story short, once you're on the network, use hostname.v.fp.o (vpn.fedoraproject.org) 16:06:16 < mmcgrath> whereas all other external requests will come through just fedoraproject.org 16:06:31 < mmcgrath> we'll no longer have the fedora.redhat.com domain (including the test boxes) and we'll be done with fedora.phx.redhat.com. 16:06:32 < paulobanon> when will this be fully functional ? 16:06:37 < jima> mmcgrath: GOOD! 16:06:47 < mmcgrath> paulobanon: *fully* functional, probably after F8 but long long before F9 16:06:51 < ricky> Nice. 16:06:57 * jima is a little tired of guessing "fedoraproject.org? or fedora.redhat.com?" 16:07:00 < mmcgrath> but we will have at least one remote proxy. 16:07:06 < mmcgrath> jima: I think others are as well. 16:07:22 * mmcgrath realizes its not second nature for most people. 16:07:29 * jima nods 16:07:46 < mmcgrath> I did test the proxy2 box, it was handling all of the fp.o traffic yesterday on a xen guest, with one processor and 1G ram. 16:07:48 < ricky> But does this mean that simply ssh puppet1, for example will need to be ssh puppet1.vpn.fedoraproject.org instead? 16:07:53 < mmcgrath> the physical box itself will allow for MUCh more than that. 16:08:03 < mmcgrath> ricky: its all in how we decide to search domains. 16:08:09 < ricky> Aha, OK. 16:08:16 < ivazquez> And configure ssh. 16:08:29 < ricky> Good point :) 16:09:00 < mmcgrath> I'm also slowly getting together a network map, this will greatly complicate our current network setup which is currently "Its in PHX or a one off in duke" 16:09:05 < paulobanon> ~when do we need to start renaming everything _ 16:09:06 < paulobanon> ? 16:09:12 < mmcgrath> hopefully the day to day functionality will be different. 16:09:22 < mmcgrath> paulobanon: not sure yet, we may not need to rename anything. 16:09:32 < mmcgrath> just change to the new scheme when we rebuild. 16:09:54 < paulobanon> k k 16:09:59 < mmcgrath> The biggest hangup I have right now is bootstrapping a build on a box that is off of the network. 16:10:21 < mmcgrath> I'd like to build over vpn so that the ks isn't sent in clear text and anaconda doesn't seem to support https (I could be wrong on that) 16:10:24 < mmcgrath> jeremy: ping? 16:11:13 < mmcgrath> I've given some thought to having xen do a bridge on the tap device, that way the xen guests wouldn't need VPN at all, they'd use the xen bridge and it'd go over the vpn from there but there are some security worries I have with that, as well as SPOF worries. 16:11:13 < jeremy> mmcgrath: what's up? 16:11:20 < londo> mmcgrath: you can do a wget, %include from kickstart would that be enough? 16:11:28 < mmcgrath> jeremy: does anaconda support https to get a ks? 16:11:34 < notting> no 16:11:39 < mmcgrath> notting: thanks 16:11:54 < jeremy> mmcgrath: well, it's more complicated than that 16:12:00 < mmcgrath> londo: the problem is getting the ks file in the first place, we'll just have to figure something else out. 16:12:05 < jima> mmcgrath: bridge + ebtables to redirect the traffic to the vpn? 16:12:10 < jima> (or such) 16:12:12 < jeremy> mmcgrath: you can have a minimal kickstart config that is just enough to get to the second stage. then you can have it include %ksappend https://... 16:12:24 < mmcgrath> jima: yeah. 16:12:39 < mmcgrath> jeremy: I'm mostly worried about sending even a fake, encrypted root password over the net. 16:13:09 < mmcgrath> no worries, we'll figure something out. 16:13:13 < jeremy> mmcgrath: you don't include the root pass in the first snippet 16:13:30 < jeremy> mmcgrath: you have lang, keymap, network, and url (or nfs or whatever) + the %ksappend line 16:13:31 < kyriakos_> mmcgrath: how feasible would it be to have local buildboxes with http proxies for the packages? 16:13:40 < mmcgrath> <nod> we could do that. 16:13:52 < mmcgrath> kyriakos_: for personal or global use? 16:13:56 < mmcgrath> s/global/public/ 16:14:07 < kyriakos_> mmcgrath: global 16:14:30 < mmcgrath> kyriakos_: people actually do all the time for local builds + squid and such 16:14:44 < mmcgrath> jeremy: ahh, I can give that a go. 16:15:05 < mmcgrath> Ok, anyone have any other questions on the vpn + new domain topic? 16:15:06 -!- GeroldKa [n=GeroldKa@fedora/geroldka] has joined #fedora-meeting 16:15:09 < mmcgrath> if not we'll move on. 16:15:10 < nirik> just FYI, we have a pretty complete mirror at our site local to proxy3, so if it pulls packages from there it should be quite zippy. 16:15:22 < mmcgrath> nirik: actually thats good to know, thanks. 16:16:05 < mmcgrath> That was ticket https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/154 BTW 16:16:06 < nirik> (mirrormanager should already point fedora stuff using mirrorlists to the right place, but you would need IP for centos/debian/ubuntu/whatever other things) 16:16:25 < mmcgrath> its still in the very early stages so I hope to keep communcations open on ideas and such when we get to actual implementation. 16:16:35 < mmcgrath> nirik: <nod> 16:16:48 < kyriakos_> is there a standard vpn package that you use? 16:16:53 < mmcgrath> Ok, next ticket is the VCS choice. 16:16:57 < mmcgrath> kyriakos_: we're using openvpn. 16:17:14 < mmcgrath> jcollie is absent again so we'll skip that. /me wonders how he's doing its been a while. 16:17:36 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Schedule 16:17:38 < mmcgrath> http://fedoraproject.org/wiki/Infrastructure/Schedule 16:17:54 < mmcgrath> Ok, Corporate Sponsorship has gone ok. 16:18:06 -!- warren [i=warren@nat/redhat/x-8a9f6cb294f7e3f1] has joined #fedora-meeting 16:18:14 < mmcgrath> right now we're still waiting for legal to get back to us with the official ok for tummy.com but its all setup and ready for the go ahead 16:18:19 * mmcgrath makes note to follow up about that. 16:18:46 < mmcgrath> Nothing terribly new this week, we have funding to purchase a server for the colo in Germany. 16:18:54 < jima> oh, cool. 16:18:56 < paulobanon> mmcgrath: nice! 16:18:56 < mmcgrath> Just waiting on the quote to come back and that should be a pretty new/good thing. 16:19:08 * mmcgrath thanks paulobanon, it could be EXTREMELY useful in the coming months. 16:19:22 < mmcgrath> I mean, a half rack in Europe is nothing to shake a stick at. 16:19:30 < paulobanon> nothing to thank for :P 16:19:47 < paulobanon> i had the contacts, so i provided them thats it :) 16:20:05 < mmcgrath> I've sent a couple of more emails out but had nothing concrete come back with a yes or no. 16:20:13 < mmcgrath> ricky: ping 16:20:18 < ricky> mmcgrath: pong 16:20:20 -!- giarc [i=hidden-u(a)gnat.asiscan.com] has joined #fedora-meeting 16:20:32 < mmcgrath> ricky: I've kind of ignored the status of that sponsorship page, are we just waiting on the new templating system? 16:20:35 < mmcgrath> how close is it? 16:20:39 < jima> a half rack? wow. 16:20:53 < mmcgrath> jima: no kidding. 16:21:13 < ricky> mmcgrath: Well, I'd say that it works now (as in can generate the static pages that we have now). 16:21:21 < mmcgrath> ..but ? 16:22:00 < ricky> It could possibly use some cleanup, though- I might not have done things in the smartest way. 16:22:08 < mmcgrath> k 16:22:14 < ivazquez> I can take a look after if you like. 16:22:44 < mmcgrath> ricky: is your stuff in the fedora CVS already? 16:22:46 < ricky> I'd like to possibly try to setup a generated site at /_/ or something and hope that we can use templates for F8. 16:22:47 < paulobanon> ricky/mmcgrath: is this something for pre-F8 or after ? 16:22:49 < mmcgrath> you're using genshi or kid or something else? 16:22:57 < paulobanon> ricky: already replied :) 16:23:07 < mmcgrath> paulobanon: pre-F8, I'm actually hoping for it in the next week or so (the sponsorship page that is) 16:23:15 < mmcgrath> and if its blocking on the templating system thats ok. 16:23:22 < ricky> mmcgrath: Genshi, and it's currently in http://ricky.fedorapeople.org/fedora-web/.git/. 16:23:42 * mmcgrath forgot about that. 16:23:54 < mmcgrath> ricky: remind me after the meeting, I'll get the websites team setup with control over that. 16:24:00 < ricky> Sure thing. 16:24:48 < ivazquez> Hrm. I can't seem to clone it. 16:25:02 -!- notting [i=notting@redhat/notting] has quit "Ex-Chat" 16:25:18 < ricky> ivazquez: Oops, running that now. 16:25:31 < paulobanon> ricky: if this is something that will go forward, why not get it into hosted ? 16:25:36 < ricky> ivazquez: Try now. 16:25:48 < paulobanon> as an actual project :) 16:25:52 < ivazquez> Much better. 16:25:59 < mmcgrath> paulobanon: welll, this one's actually going to be a place just for the websites team. 16:26:11 < mmcgrath> so it'll be going on git.fedoraproject.org, I've just been bad about getting it on there :( 16:26:21 < paulobanon> ahh ok ok 16:26:39 < mmcgrath> ricky: ivazquez: can you two give that a look over and get it up early next week? We can test in /_/ 16:27:03 < ivazquez> I'm a bit busy here, but I'll do what I can. 16:27:11 < ricky> Thanks. 16:27:18 < mmcgrath> ivazquez: thanks, I'd greatly appreciate it. 16:27:22 * jima has to roll out before the meeting endtime 16:27:26 < mmcgrath> Ok we'll move on to architecture. 16:27:48 < mmcgrath> Is there anyone here that'd be willing to document some stuff for me on SOP's or in kivio/dia? 16:28:03 < mmcgrath> I'm working on some of this as well but we can always use help :) 16:28:23 < paulobanon> mmcgrath: if u drop me what u want, i can give u a hand 16:28:37 < mmcgrath> paulobanon: excellent, I'll take you up on that. 16:28:40 -!- clarkbw [i=clarkbw@nat/redhat/x-a033520974148b46] has quit "Ex-Chat" 16:28:54 < mmcgrath> not much has happened during this week on that but more is on the way. 16:29:04 < mmcgrath> Next thing on the Schedule is SOP's, nothing new there. 16:29:08 < mmcgrath> So I'll open the floor 16:29:16 < paulobanon> proxies 16:29:17 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor 16:29:21 < paulobanon> caching that is 16:29:27 < mmcgrath> paulobanon: yes, discuss the caching on the proxies. 16:29:56 < paulobanon> so we had a nice "impersonating experience" of lmacken in bodhi this week 16:30:00 -!- mdomsch [n=mdomsch(a)cpe-70-124-62-55.austin.res.rr.com] has joined #fedora-meeting 16:30:13 < paulobanon> mod_cache was playing some tricks on us 16:30:21 < jima> heh 16:30:25 < lmacken> :) 16:30:31 -!- stahnma [n=stahnma(a)c-76-18-178-254.hsd1.tn.comcast.net] has joined #fedora-meeting 16:30:38 < mmcgrath> that was fun. 16:30:47 < paulobanon> a fully functional caching bodhi is setup in pt1.f.rh.c/updates 16:31:07 < jima> i tried impersonating lmacken at a store, but they didn't believe me. 16:31:17 < mmcgrath> For those that are interested - wget -SO/dev/null http://fedoraproject.org/wiki/ 16:31:28 < paulobanon> one thing we need to make sure we do, is standardize the static content 16:31:31 < jima> lmacken: btw, if you hear something about a shoplifting trial, it wasn't me. 16:31:34 < mmcgrath> thats a good way to get the headers (and thus information about the content you're looking at) 16:32:10 < paulobanon> so if we could take a look into our TG apps, and make sure that everything is using /static/ for images, CSS, etc 16:32:36 < mdomsch> paulobanon, mm does 16:32:37 < mmcgrath> <nod> 16:32:41 < lmacken> jima: haha 16:32:46 < paulobanon> right now, smolt/stats and docs.fp.o/ are being cached 16:32:53 < abadger1999> Cool. Will do 16:33:11 < paulobanon> hopefully early next week, bodhi will be the first app to be cached also 16:33:21 -!- Aaronfc7 [n=Aaron(a)76.216.128.74] has joined #fedora-meeting 16:33:22 < paulobanon> so testing is appreciated in PT1/updates 16:33:25 * mdomsch needs db2 cached 16:33:26 < lmacken> I will probably be updating bodhi tonight or tomorrow with TG 1.0.3.2, so we can utilize the secure cookies, and some other fixes 16:34:03 < paulobanon> if u guys have suggestions, please comment/talk/whatever :) 16:34:16 < Aaronfc7> b43 module 16:34:19 < mmcgrath> <nod> cool. 16:34:21 * jima maintains no TG apps :) 16:34:30 < lmacken> paulobanon: i'll play around with it tonight, thanks for setting it up 16:34:33 < paulobanon> if you guys want to test your app with mod_cache let me know where the testing app is, and ill setup some rewrites in PT1 16:34:37 < ivazquez> Aaronfc7: Wrong group. 16:34:38 < lmacken> jima: want to help ? :) 16:34:51 < mmcgrath> paulobanon: no doubt, thanks for getting that all setup and tested in our environment. 16:34:53 < jima> lmacken: wouldn't that typically require knowing...what, python? 16:34:54 < Aaronfc7> still learning 16:35:09 < paulobanon> mmcgrath: no prob 16:35:18 < ivazquez> jima: So... in 2 hours then? 16:35:26 < mmcgrath> paulobanon: I'd love to get some of our WikiGraphics cached 16:35:26 < paulobanon> another thing, stickum :) 16:35:30 < mmcgrath> see: http://fedoraproject.org/wiki/WikiGraphics?action=AttachFile&do=get&target=… 16:35:33 < mmcgrath> for example 16:35:38 < jima> ivazquez: ...? 16:35:44 < paulobanon> mmcgrath: pt1/wiki ;) 16:35:52 < ivazquez> In about 2 hours you'll be able to help with TG. 16:35:53 < lmacken> jima: TG turns python into a different sort of beast.. it's usually just best to dive in head first 16:35:55 < jima> ivazquez: well, for starters, i have to roll out in about 5 minutes, so definitely not. :P 16:36:25 < paulobanon> mmcgrath: forget the PT1/wiki, its not defined in modcache.conf 16:36:31 < mmcgrath> paulobanon: I actually added some caching to the production wiki (they're in puppet) 16:36:46 < paulobanon> mmcgrath: ill take a look tomorrow 16:37:07 < paulobanon> lmacken / ricky: daMaestro was interested in joining your stickum interest group 16:37:13 < mmcgrath> paulobanon: cool, anything else? If not we'll move on 16:37:14 * jima doesn't know any python, and has things like a job and family that make free time a bit erratic. :| 16:37:41 < daMaestro> +1 with helping with stickum devel 16:37:47 < ricky> daMaestro: Ask abadger1999 about getting SVN access when you see him. 16:37:51 < kyriakos_> what's stickum ? 16:37:52 < daMaestro> sure 16:37:54 < paulobanon> ricky: you wanna try pushing a testing version under pastebin.fp.o ? 16:37:58 < paulobanon> :P 16:38:00 < ricky> (Google accounts required, of course) 16:38:03 < daMaestro> kyriakos_, pastebin: example: http://f3dora.org./ 16:38:10 < daMaestro> damnit, http://f3dora.org/ 16:38:26 < daMaestro> there is also a fedora project test one, i don't have the url handy 16:38:33 < abadger1999> paulobanon, mdomsch:BTW, there's some ExpiresActive lines in pt1's mirrors.conf file that don't work. 16:38:36 < ricky> paulobanon: Hm, would we need it to be packaged first? I think mmcgrath mentioned that on the ticket. 16:38:55 < paulobanon> ricky: true true 16:38:56 < abadger1999> Not sure who's working on that but I commented them out for now 16:39:04 < ricky> publictest5.fedora.redhat.com/stickum/, may not be latest SVN- I will update it when I have the chance. 16:39:11 < paulobanon> abadger1999: mirrors.conf its not me 16:39:20 < mdomsch> abadger1999, oh? 16:39:31 < mdomsch> probably me, but I don't recall doing it on pt1 16:39:33 < jima> okay, i'm off -- have a nice night everyone. 16:39:39 < paulobanon> abadger1999: im usually under modRewrite.conf and modcache.conf 16:39:44 < ricky> jima: See you. 16:39:46 < mmcgrath> jima: later 16:39:48 < abadger1999> mdomsch: /etc/httpd/conf.d/publictest1.fedora.redhat.com/mirrors.conf 16:39:50 < paulobanon> jima: later 16:39:53 * mmcgrath attempts to get the meeting back up 16:40:11 < mmcgrath> do we have anything else we need to discuss in the meeting or should we head on over to #fedora-admin and continue discussing some of this there? 16:40:24 < abadger1999> mdomsch: I thought it was something puppet dragged in but I didn't see it in the configs on puppet1 16:40:48 < paulobanon> mmcgrath, ricky: is the pastebin something we still want for pre f8 ? 16:41:19 < mdomsch> odd 16:41:23 < mmcgrath> paulobanon: It'd be nice but we have some other priorities. 16:41:43 < mdomsch> how are we on donated resources? 16:41:46 < paulobanon> mmcgrath: that was what i was thinking 16:41:48 < mdomsch> sorry if it was covered earlier 16:41:52 < mmcgrath> but if it will just take a couple of hours to get up and running, I say have at it. 16:42:22 < mmcgrath> mdomsch: ahh, we talked about it a bit. 16:42:31 < mdomsch> ok, I'll read the logs later 16:42:44 < mmcgrath> so the tummy.com stuff is up and ready, we actually ran fedoraproject.org and wiki off of it yesterday for a couple of hours without incident. 16:42:49 < mmcgrath> mdomsch: cool 16:42:56 < lmacken> daMaestro: nice! f3dora.org++ 16:42:56 < mmcgrath> ok, if no one has anything else we'll close the meeting in 30 16:43:34 < mmcgrath> 10 16:43:40 < paulobanon> 5 16:43:42 < paulobanon> :) 16:43:51 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting End 16:43:52 < daMaestro> lmacken, yeah.. it's up just so i could learn the stickum codebase and learn TG 16:43:56 < mmcgrath> Thanks for coming everyone. 16:44:01 < ricky> Thanks a lot. 16:44:04 < paulobanon> daMaestro: cool!! 16:44:09 < paulobanon> mmcgrath: thanks! 16:44:11 < abadger1999> Thanks!

1 0

Mirrors page
by Florian La Roche 18 Sep '07

18 Sep '07

Hello all, The fullsearch link at http://fedoraproject.org/wiki/Mirrors does not make a lot of sense (just look at the result page if you let this run through all pages). Who has access to this page? regards, Florian La Roche

3 4

Introduction
by Kyriakos Oikonomakos 17 Sep '07

17 Sep '07

Hi All, My name is Kyriakos and I am interested in joining the infrastructure group. I've been working with Linux for a few years now as a sys-admin and I guess it's time to give something back to the community. My background until recently was in the telecoms industry and I'm familiar with monitoring tools such as nagios, mrtg and cacti. Also i've worked with most of the "usual" services such as apache, postfix, sendmail, bind and other. >From time to time I've written scripts in bash and perl and part of my current position is to work with a fair amount of python and java code. Currently I am based in London, UK and I work as a software support engineer for an automation software vendor. Looking forward to an interesting journey. Regards, Kyriakos

2 1

[Fwd: Fedora 7 update data not valid, do not sync.]
by Mike McGrath 16 Sep '07

16 Sep '07

FYI everyone (see FWD). I've also disabled the MM crawler until this re-sync is finished. I predict that the result will be MM sending people to some mirrors which will, at some point, contain the bad data. Once the sync is finished and we are comfortable with it we'll re-enable the job. -Mike

1 0

Infratructure Meeting Log for 2007/09/13
by Ricky Zhou 13 Sep '07

13 Sep '07

16:00:44 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting - Role Call 16:00:45 < mmcgrath> Who's here? 16:00:46 * ricky waves to warren and jeremy :P 16:00:55 * skvidal is here 16:00:57 * lmacken 16:00:57 -!- warren [i=warren@redhat/wombat/warren] has quit Remote closed the connection 16:00:57 -!- jeremy [i=katzj@nat/redhat/x-282f78b789f9fc9d] has quit Remote closed the connection 16:00:58 < mmcgrath> heheheh 16:01:00 < mmcgrath> there they go 16:01:00 < skvidal> hahahaha 16:01:00 < ricky> Hehe. 16:01:02 < skvidal> how on earth 16:01:05 < mmcgrath> like clock work 16:01:48 -!- jeremy [i=katzj@nat/redhat/x-0f82d1e06695232a] has joined #fedora-meeting 16:02:09 < mmcgrath> mbonnet_: mdomsch lmacken abadger1999 dgilmore jima ping 16:02:13 < mmcgrath> anyone I forgot ping 16:02:37 < abadger1999> heheh, I wonder what would happen if I started watching for "anyone" 16:02:46 * nirik sits in the spectator seats. 16:02:47 -!- warren [i=warren@nat/redhat/x-f451cb84616c1460] has joined #fedora-meeting 16:03:07 < warren> connection died 16:03:20 < mmcgrath> Ok, lets get started 16:03:28 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Tickets 16:03:39 < mmcgrath> https://hosted.fedoraproject.org/projects/fedora-infrastructure/query?statu… 16:03:40 -!- giarc [i=hidden-u(a)gnat.asiscan.com] has joined #fedora-meeting 16:03:59 < mmcgrath> So first a new ticket # 152 16:04:06 -!- glezos [n=glezos@fedora/glezos] has joined #fedora-meeting 16:04:07 < mmcgrath> VPN setup. 16:04:18 < mmcgrath> Did everyone get my email to the list earlier this week? 16:04:32 * jima stumbles in 16:04:38 * glezos jumps in too 16:04:48 < mmcgrath> Did anyone get it? 16:04:49 < ricky> mmcgrath: Sorry, what was the subject again? 16:05:06 < jima> oh, about vpn stuff 16:05:07 < skvidal> architecture changes 16:05:23 < ricky> Oops, got it- missed that the first time somehow. 16:05:28 < warren> "external systems" meaning what? 16:05:32 < abadger1999> Yep. 16:05:37 < mmcgrath> I'm still deciding on some of the technical bits on it. For example do we want bridged or routed, etc. 16:05:41 < mmcgrath> warren: anything not in PHX. 16:05:54 < mmcgrath> which at this point would include Duke and the tummy.com servers. 16:06:12 < mmcgrath> s/servers/hosts/ 16:06:16 * dgilmore is here 16:06:18 < warren> what is running on tummy.com? 16:06:38 < skvidal> nothing, yet 16:06:46 < mmcgrath> Right now xen9 and proxy3. Its not official as its blocking on an ok from legal and the VPN (described in #152) 16:06:49 < ricky> proxy3 in the future? 16:06:52 < mmcgrath> ricky: yep 16:07:17 < mmcgrath> xen9 is actually up and using puppet and everything, proxy3 is up but not configured yet. 16:07:43 < mmcgrath> Anyone have any comments on the email? Did it sound sane enough to try? 16:07:50 < ricky> So I assume that the way the Duke servers was setup required going through Redhat IS to get firewall stuff? 16:07:54 < mmcgrath> paulobanon: ping (forgot to ping you earlier) 16:08:07 < mmcgrath> well, there's a couple of problems we're solving by using vpn. 16:08:31 < mmcgrath> the biggies are access to the internal network, and encrypted communications (for example with bacula) 16:08:38 * nirik is a big fan of openvpn. Works great and is very flexable. 16:08:47 < mmcgrath> nirik: I am too, I've had great success with it in the past. 16:08:59 * jima is a big openvpn fan, as well. 16:09:05 < ricky> I've toyed with OpenVPN a bit- I like it a lot. 16:09:18 < mmcgrath> So there's a few questions still floating around in my head. 16:09:36 < mmcgrath> 1) name space 16:09:46 < mmcgrath> 2) ip space (routed vs bridged) 16:09:48 < jima> iow, i think there might be a general consensus that openvpn was the right software to use ;) 16:09:50 < mmcgrath> 3) bootstrapping 16:10:08 * jima has only done routed 16:10:19 < mmcgrath> With name space as it is we have proxy[1-2].fedora.phx.redhat.com and proxy3.fedoraproject.org 16:10:40 -!- fab__ [n=bellet(a)bellet.info] has quit Connection timed out 16:10:48 * jima winces a little 16:11:02 < mmcgrath> We (or I) have to figure out a proper domain for all of our stuff. I think this line of thought will ultimately end in us running our own DNS and ridding ourselves of fedora.phx.redhat.com 16:11:13 < mmcgrath> 2) Is still up for debate. 16:11:21 < mmcgrath> and 3) I think will just be a technical implementation. 16:11:39 < mmcgrath> does anyone have any comments/ideas/concerns for what I'm going for here? 16:11:52 -!- fab__ [n=bellet(a)bellet.info] has joined #fedora-meeting 16:12:24 < skvidal> any concerns this will play silly buggers with xen or kvm? 16:12:53 < mmcgrath> skvidal: AFAIK its gone fine. 16:12:54 < jima> what will? openvpn? 16:13:06 < mmcgrath> I've actually got proxy3 connected to bastion right now in a test via an SSH tunnel + openvpn. 16:13:10 < mmcgrath> seemed to play just fine. 16:13:16 < skvidal> mmcgrath: cool. thanks 16:13:31 * jima could spin up an openvpn link involving xen, if mmcgrath hadn't 16:13:40 < mmcgrath> :) 16:13:54 < mmcgrath> ok, so I'll keep everyone informed on that more when the blocks (RHIS mostly right now) are all figured out. 16:14:06 < mmcgrath> Next ticket is.... 16:14:13 < mmcgrath> #14 which I'd imagine is still on hold 16:14:22 < mmcgrath> no jcollie, we'll skip. 16:14:33 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Schedule 16:14:35 < mmcgrath> http://fedoraproject.org/wiki/Infrastructure/Schedule/ 16:14:37 -!- frankc [i=824c6013(a)gateway/web/cgi-irc/ircatwork.com/x-31ec561b1be00be3] has joined #fedora-meeting 16:14:51 < mmcgrath> First item is Corporate sponsorship. 16:15:06 < ricky> :( at the websites team response. 16:15:14 < skvidal> mmcgrath: you agreed to have the starbucks logo tatoo'd on your forehead? 16:15:20 < ricky> Hehe. 16:15:25 < mmcgrath> I've been in a couple of meetings this week. Some stuff looking good but nothing in stone yet. 16:15:30 < mmcgrath> skvidal: inner thy. 16:15:31 < glezos> ricky: sorry, I've been really busy lately :( 16:15:37 < mmcgrath> glezos: not your or ricky's fault. 16:15:40 < skvidal> mmcgrath: more people would see that, yah 16:15:41 < ricky> glezos: No problem. 16:15:46 < jima> mmcgrath: that sees much traffic, then? 16:15:47 * f13 peeks in. 16:15:48 < dgilmore> skvidal: we need to leave space there for a second logo also 16:15:50 < mmcgrath> I was actually thinking about sending a cattle call out for more 'web masters' 16:16:00 < mmcgrath> since it turns out both ricky and glezos actually know how to code :) 16:16:06 < skvidal> dgilmore: :) 16:16:18 < mmcgrath> It'd be nice to have one or two people who can focus ENTIRELY on the view and look of our pages. 16:16:26 < ricky> For the "record," I don't have anything started with the sponsorship page, as some people on the list thought. 16:16:27 -!- jwb is now known as jwb_gone 16:16:36 < jima> and who can do design worth >crap? wow! 16:16:40 < mmcgrath> ricky: thats my thought :) 16:16:43 < mmcgrath> err fault. 16:16:56 < ricky> Ah, no problem. 16:17:27 * mmcgrath sees if he can get mizmo's attention real quick for some consultation. 16:17:37 < ricky> I was expecting a slightly more excited response, though. It's basically just a "design and plug into template" type of thing. 16:17:51 -!- mizmo [i=duffy@nat/redhat/x-686300b264b66012] has joined #fedora-meeting 16:17:52 < mizmo> yo 16:17:55 < mmcgrath> ricky: me too, I think we have a shy bunch there. 16:18:19 < mmcgrath> mizmo: Who do you think would be more interested in the design / presentation aspect of our websites. The art-list or the marketing-list? 16:18:29 < glezos> mmcgrath: or puzzled on how to start/publish somthing.. 16:18:32 < mmcgrath> I'm thinking about sending a notification that we're looking for some more people in the websites list. 16:18:36 < mmcgrath> glezos: yeah. 16:18:51 < mizmo> mmcgrath: i think art list 16:19:04 < ricky> I'm really hoping that website buildsystem thingy that I'm playing with will stir up some life on the list. 16:19:06 < mmcgrath> k, thanks. 16:19:15 < mmcgrath> ricky: <nod> 16:19:30 < glezos> ricky: cool 16:19:39 < mmcgrath> so How about this, I'll send an email to the art list to see if we can get a few people (3 or 4) to join the websites team to help work on just those aspects of the website. 16:19:48 < mizmo> sounds good 16:19:52 < mmcgrath> that should help take the load off of ricky and glezos since they're actually working on some apps right now. 16:20:18 < mmcgrath> The thing thats tricky here is that I'd really like the websites team to be able to take and do requests similar to how the art team has their page now. 16:20:40 < mmcgrath> Anyone have any comments on that? If not we'll move on to the next topic. 16:20:54 < dgilmore> mmcgrath: sounds good to me 16:20:58 * jima avoids design like it might be contagous 16:21:20 < glezos> mmcgrath: +1. Some more documentation on how to checkout code and build a local fpo website could help newcomers. 16:21:21 < jima> contagious, even 16:21:23 < mmcgrath> jima: I do too, I'm just terrible at it. 16:21:27 < jima> mmcgrath: ditto 16:21:36 < mmcgrath> glezos: indeed, is there even a site on the wiki for the websites team? 16:21:38 < ricky> +1 for the websites team taking requests, etc. like the art team. 16:21:44 < mmcgrath> I guess now's a good time to bring up governance a bit. 16:21:44 < ricky> mmcgrath: fp.o/wiki/Websites 16:21:47 < dgilmore> thast why we have people like mizmo to make things look pretty 16:21:53 < glezos> mmcgrath: there is, but it seems dead :/ 16:21:57 < ricky> May be slightly outdated :( 16:22:03 < jima> more power to those who can do design :| 16:22:10 < mmcgrath> dgilmore: more than just that, people like mizmo makes them usable in many cases. 16:22:24 < glezos> I think the overlapping between art, infra, docs and websites has caused the latter to become a bit stalled 16:22:44 < mmcgrath> So here's a general question, should the websites team remain a separate entity or should it be considered a subset of the infrastructure group? 16:23:06 < ricky> glezos: Once we expand the static pages a bit, the websites team will have a much more defined/specific goal. 16:23:07 < mmcgrath> On the one hand I'd like it to be different, but it feels like the infrastructure team has a total overlap with it and that we keep 'kick starting' it over the last couple of years. 16:23:08 < skvidal> does it matter? 16:23:14 < mmcgrath> skvidal: you've been aroudn longest, what do you think? 16:23:18 < ricky> (Which is my larger goal with what I'm doing now). 16:23:20 < mmcgrath> skvidal: I have no idea, just thinking out loud. 16:23:27 < skvidal> well, I mean - from a hierarchy standpoint what does it get us? 16:23:42 < skvidal> if the websites team stops doing things we kick start it and maybe it does something 16:23:43 < f13> that's a question of content vs service 16:23:44 < mmcgrath> I have no idea. 16:24:00 < skvidal> if the websites team is a subset and it stops doing things, we kick start it and maybe it does something 16:24:04 < skvidal> sounds like a toss up either way 16:24:05 < f13> infrastructure team owns the services, httpd, daemons whatever. content folks own teh content served by those services. 16:24:23 < skvidal> f13: that's a fair distinction and it speaks to maintaining the status quo 16:24:29 < mmcgrath> f13: I don't want that to change, Infrastructure + content scares the crap out of me :) 16:24:34 * ricky adds the l10n team into the websites-related teams medley. 16:25:09 < f13> mmcgrath: right, keep the content out of infrastructure. 16:25:28 < mmcgrath> So lets leave it as it is, even if it is an artificial separation right now. That may not be the case in the future. 16:25:47 < mmcgrath> One of the problems is that the websites team has only just started seeing work to do over the last couple of months for the most part so its hard to gauge it. 16:25:51 * mmcgrath will contact the arts list. 16:25:56 < glezos> FWIW, we use templating a lot, so it shouldn't be hard for the designers to contribute *only* on design. 16:26:05 < mmcgrath> <nod> 16:26:14 < mmcgrath> anything else on this topic? If not we'll move on. 16:26:35 < mmcgrath> k, next topic 16:26:48 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Architectural Documentation 16:27:00 < mmcgrath> When the VPN roles out we'll actually be in an environment that is reasonably complex. 16:27:20 < mmcgrath> In theory we could have 4 or 5 external proxy/cache servers at the end of the year. Hopefully in different countries. 16:27:43 < mmcgrath> how they all communicate and what they all do will be, by far, the most complex environment this team has seen so we'll have to do a better job of documentation. 16:28:03 < mmcgrath> This has typically fallen on me for most stuff (I was very happy to see glezos add an SOP this week) 16:28:10 -!- rdieter_away is now known as rdieter 16:28:24 < mmcgrath> but we may have to spend a week or two, as a team, getting http://fedoraproject.org/wiki/Infrastructure/Architecture/ setup better. 16:28:36 < mmcgrath> anyone have any ideas on how best to allow others in the team to do this? 16:29:23 < mmcgrath> jima: this would be a good thing from the noc point of view btw. 16:29:38 < mmcgrath> We'll talk about that as we start to implement it more. 16:29:39 < ricky> Split it up into categories/pieces that need documentation? 16:29:42 * jima nods 16:29:52 * jima looks at that wiki page 16:30:04 < mmcgrath> ricky: yeah, I could put a section under there that has everything listed including what has not yet been documented. 16:30:24 < mmcgrath> I'll move on for now though 16:30:30 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- SOP 16:30:36 -!- clarkbw [n=clarkbw(a)static-71-243-117-136.bos.east.verizon.net] has quit Read error: 110 (Connection timed out) 16:30:39 < jima> AIEE! drawings! 16:31:16 * jima flees 16:31:16 < mmcgrath> New SOP from this week was provided by glezos, everyone take a look - http://fedoraproject.org/wiki/Infrastructure/SOP/Translations 16:31:23 < mmcgrath> glezos: good job with that. 16:31:38 < mmcgrath> Next topic 16:31:47 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- User Sponsorship 16:31:59 < mmcgrath> No new infrastructure users 16:32:06 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor 16:32:16 < mmcgrath> Ok, so we're at the half hour mark or so, anyone have anything they'd like to discuss? 16:32:48 < skvidal> facter 16:32:48 < nirik> anyone have ideas on how to fix https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/89 it's been busted for a while. ;( 16:32:51 -!- tibbs [i=tibbs@fedora/tibbs] has quit "Konversation terminated!" 16:33:17 < mmcgrath> nirik: we can disable the plugin I suppose. Thats an upstream thing though so I think many have been timid to take it on. 16:33:20 < mdomsch> nirik, it tracebacks... 16:33:20 < mmcgrath> <nod> 16:33:33 < mmcgrath> skvidal: talk about what you've done with facter 16:33:55 < f13> yeah, drop the plugin. 16:33:56 < skvidal> after a great deal of hating myself I've added a couple of facts to facter for puppet 16:33:57 < nirik> yeah, the ticket iself, not the problem it describes. ;) I mailed the submitter direct and I think it's complete. Perhaps there is some way for someone to mark it done in the db? 16:33:59 < ricky> mdomsch: It seems to be a bug with the git plugin. 16:34:08 < ricky> mdomsch: And when the git plugin is disabled, it complains about the SVN plugin- grrr. 16:34:13 < mdomsch> doh 16:34:20 < mdomsch> redirect git URLs -> gitweb 16:34:21 -!- couf [n=couf@fedora/couf] has quit "leaving" 16:34:31 < f13> ricky: huh? 16:34:44 < f13> ricky: perhaps the disabling was done wrong. 16:34:47 < skvidal> puppet uses the program facter to determine all sorts of things about the box - login to any infrastructure box and type 'facter' to see the results - I've added a file in cvs called fedora-local.rb 16:34:58 < f13> I've successfully disabled thigns in teh revisor trac so that they can deal with a ticket, then re-enable it. 16:34:59 < mmcgrath> f13: one sec, we'll talk about that in a bit. 16:35:01 < skvidal> we can edit that file to update the facts puppet can use as $variables 16:35:35 < skvidal> so right now I added: $distrorelease which is the result of rpm -q --qf "%{version}\n" --what-provides redhat-release 16:35:45 < skvidal> and I've added mmcgrathisawanker 16:35:48 < skvidal> which is always true 16:35:50 < skvidal> :) 16:35:50 < mmcgrath> facter | grep mcgrath 16:35:53 < mmcgrath> is amusing. 16:36:02 < mmcgrath> heh 16:36:03 * dgilmore is scared 16:36:05 < skvidal> I did that to show how to set facts as simple strings 16:36:18 < mmcgrath> And it is pretty simple. It will be nice to be able to use this. 16:36:30 < f13> skvidal: hrm, eventually that should probably change to "system-release" but not for a while. 16:36:33 < mmcgrath> skvidal: thank's for doing the research and getting that working and for the whole rhel5.repo file thing :) 16:36:34 < skvidal> the point is if we need to make a decision about something based on some value on the system we can add facts, push the file then use those facts in the next run of uppet 16:36:49 < skvidal> f13: not any time soon for rhel and centos :) 16:36:58 < f13> yeah, RHEL6 timeframe. 16:37:00 < skvidal> oh yah and I closed the reposync ticket 16:37:13 < skvidal> which just means we're using an 'ahem' local cache 16:37:14 -!- walters [n=walters(a)static-71-243-117-136.bos.east.verizon.net] has quit Read error: 110 (Connection timed out) 16:37:21 < skvidal> of things for updating our rhel5 boxes 16:37:23 < mmcgrath> hurray for a local cache. 16:37:28 * jima doesn't even know if he has shells on any infrastructure servers 16:37:31 < mmcgrath> skvidal: thanks again. 16:37:40 < skvidal> so the annoying rhn errors will be gone 16:37:48 < skvidal> mmcgrath: sorry it took so long 16:37:51 < mmcgrath> jima: you do, I thought I even asked you to look into somethign once. I just thought you were too busy / not interested. 16:37:57 < mmcgrath> skvidal: mostly my fault for not giving you a place to store it. 16:38:04 < mmcgrath> ok, we'll move on to the plugin error. 16:38:13 < mmcgrath> f13: actually I ran into that problem too, whats the right way to disable all plugins? 16:38:17 < jima> mmcgrath: mostly i was like "...huh?" 16:38:54 < f13> just a tic 16:39:14 * jima logs into bastion. wow! 16:39:17 < mmcgrath> err all plugins for a specific repo 16:40:04 < f13> oh haha. 16:40:10 < skvidal> f13: ? 16:40:13 < f13> looks like you tried commenting with #, I think the config file comment is ; 16:40:24 < mmcgrath> yeah that sounds like me :) 16:40:37 < ricky> f13: I'm pretty sure that both are valid. 16:40:43 < mmcgrath> though I thought it just made it use the svn plugin instead. 16:40:48 < f13> is somebody editing it right now? 16:40:55 < mmcgrath> <nod> see - https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/89 16:40:58 * mmcgrath is not. 16:41:01 * dgilmore goes to kick jima :) 16:41:03 < f13> there is a .swp file there. 16:41:19 < jima> dgilmore: ? 16:41:27 * f13 writes anyway 16:41:31 < dgilmore> jima: * jima logs into bastion. wow! 16:41:54 < ricky> f13: Oops, sorry- that's me. 16:42:06 < f13> Found a swap file by the name "/srv/web/trac/projects/fedora-infrastructure/conf/.trac.ini.swp" owned by: apache dated: Thu Sep 13 13:42:36 2007 16:42:15 < ricky> f13: I closed it- write away. 16:42:16 < f13> ricky: can you quit so I can write? 16:42:40 -!- rdieter [n=rdieter(a)sting.unl.edu] has quit Remote closed the connection 16:43:36 < f13> interesting. 16:43:41 < mdomsch> mmcgrath, reason for httpd being stopped on app4? 16:43:55 < f13> it didn't used to do this. Wonder how 'svn' is not supported anymore. 16:44:20 < mmcgrath> mdomsch: nope, and actually puppet should be enabling it if its off, let me look at it real quick. 16:44:36 < ricky> trac.versioncontrol.* = disabled in [components] didn't even do it. Grrr. 16:44:43 < mdomsch> tuesday 11:39am it stopped with a SIGTERM 16:45:11 < mmcgrath> I seem to remember shutting it down for a test but I'm wondering why puppet didn't turn it back on. 16:45:26 < mdomsch> puppet will, or supervisor (which doesn't) 16:45:48 < f13> ricky: yeah, this is bothersom. svn should be a valid scm 16:45:51 < mmcgrath> puppet should turn on httpd. 16:46:04 < mmcgrath> f13: so you're seeing the same thing we are with that? Something strange is going on? 16:46:23 < ricky> My other question is.. why does the tickets component *care* about SCM? 16:46:50 < f13> ricky: some twisted path of Trac. You can have wiki entries that are SCM checkin IDs and it will tooltip the checkin comment 16:47:08 < f13> kind of neat, but means that anything wiki based (which tickets are) go through the scm plugin code path. 16:47:14 < f13> but svn /should/ be there, it's stock in trac. 16:47:18 < ricky> Sounds cool, but adds more points of failure. 16:47:31 < abadger1999> ricky: IIRC, someone added a link into the repository in that ticket (changeset:1111 type thing) 16:47:40 < ricky> Ahh. 16:47:57 < ricky> We could remove it manually, but it'd be nice if this kind of thing worked too :) 16:48:03 < mmcgrath> lets discuss this in #fedora-admin after the meeting (Since the meeting is almost done) 16:48:13 < ricky> Sure thing. 16:48:28 < mmcgrath> does anyone have anything else to discuss? If not we'll close the meeting in 30 16:48:38 * lmacken is upgrading bodhi as we speak :) 16:48:45 < mmcgrath> 15 16:48:58 < mmcgrath> 5 16:49:03 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting End 16:49:05 < jima> lmacken: yay! 16:49:06 < mmcgrath> thanks for coming everyone 16:49:08 < ricky> Thanks. 16:49:12 < mmcgrath> mizmo: thanks for stopping by. 16:49:22 < mizmo> yep :) 16:49:27 -!- mizmo [i=duffy@nat/redhat/x-686300b264b66012] has left #fedora-meeting ["w00tw00t"] 16:49:33 < glezos> mmcgrath: thanks

1 0

mock-0.7.2-1.fc6.1
by Mike McGrath 12 Sep '07

12 Sep '07

Anyone know any reason why we shouldn't upgrade mock? Our last upgrade borked a bunch of stuff. I'm under the impression that has all been fixed but thought I'd ask. -Mike

2 1

rename InitialCC in CVS Request Template to watchbugzilla
by Till Maas 11 Sep '07

11 Sep '07

Aloas, I am not sure, but I guess that InitialCC: in New Package CVS Request (and the CC entries in Package Change Request) correspond to the "watchbugzilla" right in the PackageDB. Therefore please rename it. And how is one supposed to add someone to watchcommit and commit with a New Pacakge CVS Request? Or is this something that needs to be done via the PackageDB interface? Regards, Till

2 1

Architectural Changes
by Mike McGrath 10 Sep '07

10 Sep '07

As we talked about in the meeting yesterday we have a new sponsor (http://www.teliasonera.com/) There are a couple of others in the works (I don't want to officially announce until its finalized) but one thing is clear. Pretty soon we're going to have multiple proxy servers outside of PHX. The end goal here would be to use mod_geoip to re-direct people to their nearest location but we're going to take baby steps to get there. Here are the steps as I see them. 1) Finalize the caching stuff paulobanon has been working on. 2) VPN 3) Setup 1 remote proxy server and test 4) Get DNS setup properly to direct people to the proxy servers in a RR format 5) mod_geoip. 4) is still a little fuzzy in my mind. Right now we're using Bind for DNS and, AFAIK, the version we're using does not have support for geoip. So my thought is using mod_geoip to direct people to (for example) de1.fedoraproject.org or us2.fedoraproject.org. I'm still a little unclear on the best way to do this in our environment. Those keeping an eye on the commit logs will have noticed the odd commit for t.fedoraproject.org. So, for example: ping -c1 t.fedoraproject.org For me seems to do the right thing. I get basically a RR balanced IP between 3 addresses (fp.o, yahoo and google) I just picked two ip's that weren't ours to balance around. The thing, for me at least, is I get fp.o every time if I use FireFox. This is over many days on different computers. I've seen FF bring up the google ip once. So I ask those on the list to go to http://t.fedoraproject.org/ and just tell me what you get. Or, even better, explain to me what the heck is going on there, I have one theory about first requests to DNS vs named caching in FF and name caching elsewhere. But we've had different people get many different results (some get wget to RR, some with wget always get the same thing, same with curl, lynx, w3m, and HEAD) More investigation is needed. 2) is something I'm working on now. VPN will only be for external servers (not users). We've actually already had a few issues we've had to overcome in strange ways from external servers that could have been fixed by a VPN. (puppet and bacula backups immediately come to mind) We'll tightly control (iptables) what these boxes have access to on the vpn server (bastion). We'll keep the ttl on our load balanced products lower so that if something does go wrong with one of them, we can easily take it out of the mix. The reason for 2) is so we don't have to maintain multiple different proxy server types. If we use VPN we can treat each server the same, just like the ones we have now which keeps it maintainable. Questions / Comments / Suggestions? -Mike

10 10

Off on Monday
by Mike McGrath 10 Sep '07

10 Sep '07

Just letting you guys know I'll be off on Monday, going to a cubs game and generally doing everything you've seen in: http://imdb.com/title/tt0091042/ -Mike

8 13

Fedora Infrastructure IRC Meeting Log from 2007-09-06
by Ricky Zhou 06 Sep '07

06 Sep '07

16:04:35 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Role Call 16:04:37 < mmcgrath> who's here? 16:04:39 < skvidal> I am 16:04:45 < skvidal> one sec I'll be on the bridge 16:04:55 * ricky is here, but will probably have to go very shortly. 16:04:58 * glezos is here 16:05:19 * lmacken 16:05:19 -!- jeremy [i=katzj@nat/redhat/x-48f778bf0db0139e] has quit Remote closed the connection 16:05:19 -!- warren [i=warren@redhat/wombat/warren] has quit Remote closed the connection 16:05:21 * abadger1999 here 16:05:22 < lmacken> heh 16:05:24 < lmacken> clockwork 16:05:29 < mmcgrath> there they go. 16:05:41 < abadger1999> Some people just don't want to do an honest days work :-) 16:05:43 < mmcgrath> I know dgilmore is traveling a lot this week. 16:05:49 < frankc> Frank is listening. 16:05:59 < mmcgrath> hello frank 16:06:01 < mmcgrath> paulobanon: ping? 16:06:37 < mmcgrath> Ok, lets get started. 16:06:42 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Tickets 16:06:50 < mmcgrath> https://hosted.fedoraproject.org/projects/fedora-infrastructure/query?statu… 16:06:54 < mmcgrath> We really only have the one on the list 16:07:09 < mmcgrath> and jcollie isn't around right now. 16:07:25 < mmcgrath> The lack of interest in that from the community in general leads me to believe that we'll be on CVS for a long time to come. 16:07:28 < mmcgrath> just a prediction :) 16:08:12 < mmcgrath> This will probably be a short meeting, not much happened in the last week. 16:08:17 < skvidal> I think most people don't want to hear the flame war 16:08:19 < abadger1999> heh. It's not (really) broken so don't fix it. 16:08:47 < mmcgrath> heh, thats true. 16:08:52 < mmcgrath> Ok, we'll move on to the schedule next. 16:08:56 < glezos> skvidal: what if there's no flame work and just action? 16:08:57 < lennert> ... 16:09:03 < skvidal> glezos: hahahahahahaha 16:09:05 < skvidal> glezos: that's funny 16:09:12 < skvidal> glezos: are you a comedian for your day job? :) 16:09:15 < glezos> :) 16:09:19 < ricky> Darn- I must go. 16:09:25 < mmcgrath> ricky: later ricky! 16:09:29 * ricky will read the logs in a bit :) 16:09:34 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- http://fedoraproject.org/wiki/Infrastructure/Schedule/ 16:09:40 < glezos> skvidal: we're infra, we can do anything :P 16:09:49 < skvidal> right-o :) 16:09:57 < mmcgrath> I might remove the VCS from the schedule since there's been little progress in it and it is already in the tickets. 16:10:11 < mmcgrath> As far as sponsorship goes we have a new member to add to the web group - glezos 16:10:14 < mmcgrath> welcome glezos. 16:10:23 -!- tibbs|h [n=tibbs@fedora/tibbs] has quit Read error: 104 (Connection reset by peer) 16:10:24 < skvidal> yay glezos! 16:10:28 < mmcgrath> s/web group/sysadmin-web group/g 16:10:30 < glezos> mmcgrath: glad to be here. 16:10:37 -!- tibbs|h [n=tibbs@fedora/tibbs] has joined #fedora-meeting 16:10:42 < mmcgrath> glezos: have you had a chance to verify your access to bastion? 16:10:54 < glezos> mmcgrath: yes, it seems to work fine, thanks 16:11:02 -!- JSchmitt [n=s4504kr@fedora/JSchmitt] has quit "Konversation terminated!" 16:11:25 < mmcgrath> and nothing new on the SOP stuff. 16:11:39 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Corporate Sponsors 16:11:50 < mmcgrath> So there's some good news and some same old news on the sponsorship front. 16:12:02 -!- jeremy [i=katzj@nat/redhat/x-35dd6359cde2db97] has joined #fedora-meeting 16:12:06 < mmcgrath> we've been offered a half rack + BW from a colo in Germany 16:12:09 < skvidal> jeremy: you're late! 16:12:24 < skvidal> mmcgrath: cool, well done 16:12:35 < mmcgrath> The company is TeliaSonera 16:12:42 < skvidal> pretty name 16:12:45 < mmcgrath> :) 16:12:46 < mmcgrath> yeah 16:13:22 < lennert> telia is a carrier 16:13:35 < mmcgrath> and we get free remote hands from Proio.com 16:14:15 < mmcgrath> So for us its just a matter of finding some servers to put over there. 16:14:27 < mmcgrath> I'm working on that but if the IBM's or Dell's out there hear anything let me know. 16:14:33 < skvidal> mmcgrath: so we hav 21U 16:14:41 < skvidal> do we have money to buy servers w/yet? 16:14:45 < mmcgrath> skvidal: I believe so. 16:15:01 < skvidal> ie: can we just spend 9K and get a 1950 in germany? 16:15:05 < mmcgrath> We've never really had a budget so I'm working on that. We did get money for the warrantys. 16:15:09 < mmcgrath> I'm hoping so. 16:15:29 < mmcgrath> Back in July/June I created an end of the year budget that included 2 server purchases (both 1950's) 16:15:40 < mmcgrath> so, if that gets approved as I think it has, then we can purchase and ship. 16:15:54 < mmcgrath> I'd stick one 1950 in PHX and one 1950 in Germany (probably through dell.de) 16:16:01 < skvidal> sounds like a good plan 16:16:14 < mmcgrath> the req process internally isn't the most fun and people are busy and haven't been as responsive as I'd hoped. 16:16:43 < mmcgrath> So thats the latest on that, any questions? 16:17:08 < glezos> mmcgrath: any ideas on how we're going to promote this support? banners, announcement? 16:17:29 < mmcgrath> glezos: I'm glad you asked, especially since you have an eye for good looking pages and access 16:17:33 * mmcgrath grabs ticket 16:17:47 < mmcgrath> This is it here - https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/137 16:18:00 < mmcgrath> I think ricky may have done some preliminary work on this but I'm not sure how far he's gotten. 16:18:13 < mmcgrath> I've sent a few emails to the websites list (including the cattle call one). 16:18:22 < glezos> mmcgrath: I'll take a look as well 16:18:30 < mmcgrath> I asked if there were still people on the list, they said yes, and some said they weren't sure how to get started. 16:18:44 < mmcgrath> so I sent that ticket explictly out to see if anyone would make a page but I've gotten no response :( 16:18:55 < mmcgrath> glezos: I'd appreciate that. 16:18:57 < glezos> mmcgrath: what about an announcement? 16:19:08 < mmcgrath> glezos: ? what did you have in mind? 16:19:46 < glezos> mmcgrath: dunno, if we do need more support then it could act as a "thanks" and invitation for more support 16:19:55 < glezos> on FWN for example 16:20:32 < mmcgrath> we've talked about some of that stuff, I'd like to save some of the announcements for the larger donations, like if IBM gave us 200 grand in hardware or something. 16:20:46 < glezos> mmcgrath: agreed 16:20:56 < mmcgrath> so far though no ones requested it. Its hard to put a value on donation because, in theory, we might have a bunch. 16:21:13 < mmcgrath> But, since the "powered by dell" logo is the only thing we have right now I'd at least like to get the sponsorship page and such setup. 16:21:45 < glezos> mmcgrath: right 16:21:47 < mmcgrath> so anywho, thats all we've got on that for this week. 16:21:52 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor 16:22:02 < lmacken> https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/141 16:22:04 < mmcgrath> Its been a pretty quiet week, I've been making lots of little changes. 16:22:14 < lmacken> are there any account system guys that could take care of that ? 16:22:23 < skvidal> mmcgrath: is lockbox pretty much use-free now? 16:22:46 < mmcgrath> skvidal: pretty close, we still have infrastructure.fp.o to move off of there (its stored on the netapp but served through lockbox) 16:22:56 < mmcgrath> and we still deploy some apps via lockbox though that should be trivial to fix. 16:23:03 < skvidal> mmcgrath: what does infrastructure need? 16:23:17 < mmcgrath> just some place to mount that share elsewhere. 16:23:35 * f13 peeks in 16:23:36 < abadger1999> Are we mounting the full netapp on puppet1 now? 16:23:36 < skvidal> okay b/c when its free I can bump it to rhel5 and setup reposync 16:23:38 < mmcgrath> we need to find a different master server to attach to the netapp, right now lockbox is the only server that can rw to everything on the netapp as root. 16:23:46 < mmcgrath> abadger1999: we might be but it can't write to it. 16:23:52 < abadger1999> k 16:23:55 < mmcgrath> lmacken: we'll talk about that in a sec. 16:23:59 < mmcgrath> skvidal: what did you have in mind? 16:24:21 < skvidal> I don't need much - just a rhel5 instance with entitlements and disk space 16:24:37 < skvidal> so I can get it to sync down rhel5, rhel5-virt, and rhel5-updates 16:24:48 < mmcgrath> <nod> 16:24:58 < mmcgrath> I'll put in a ticket request for that today. 16:25:03 < skvidal> you suggested lockbox before 16:25:16 < skvidal> and I had other things going on so I was waiting for it to free up to do the reposync bit 16:25:37 < abadger1999> The hosted trac repositories sync via the netapp mount on lockbox as well. 16:25:40 < mmcgrath> lockbox would be good but A) its almost dead (EOL'd) and B) I want to stick the sync mechanism on a xen instance. 16:25:58 < mmcgrath> abadger1999: yeah, there's a few cron jobs we'll need to migrate. 16:26:20 < mmcgrath> lmacken: ok, lets talk about #141 16:26:28 < lmacken> ok 16:26:53 < mmcgrath> how are they obtained now? Which pages on the wiki? 16:27:00 < skvidal> mmcgrath: okay if there's a better choice of a box to stick a xen instance on where I can write to the netapp, I can do it there 16:27:02 < lmacken> No idea.. I don't know anything about those certs 16:27:45 -!- caillon [n=caillon(a)mithril.returnzero.com] has left #fedora-meeting [] 16:28:35 * jeremy looks 16:28:56 < mmcgrath> http://fedoraproject.org/wiki/PackageMaintainers/UsingPlagueClientFaq 16:29:10 < abadger1999> lmacken: Does till just want us to host the files available from the wiki on admin.fp.o? 16:29:29 * dgilmore is here 16:29:30 < mmcgrath> abadger1999: I think the concern is that someone could go in and change those files. 16:29:44 < lmacken> yeah, in "a secure way".. so I'm assuming he wants us to host it from https://admin.fp.o/accounts 16:30:02 < jeremy> yeah, that's what it looks like 16:30:07 < jeremy> and we definitely should! 16:30:12 < lmacken> definitely 16:30:25 < mmcgrath> yep, easy fix. lmacken if you don't want to do it assign the ticket to me, I'll do it after the meeting. 16:30:52 < lmacken> sounds good. I don't know my way around the account system.. So i'll reassign :) 16:31:09 < mmcgrath> lmacken: I'm just going to stick it under /accounts/ somewhere and change the links on the wiki. 16:31:21 < mmcgrath> Does anyone have anything else they'd like to discuss? 16:31:26 < glezos> mmcgrath: As a heads-up, I'll create a ticket now for what we need for transifex 16:31:32 < mmcgrath> glezos: k 16:31:35 < abadger1999> wishlist item: I'd like to have a ~/.fedora/ directory for bodhi/koji/etc config, sessioncookies, certs. 16:31:51 < jeremy> abadger1999: makes sense to me 16:31:56 < mmcgrath> abadger1999: we can do that. 16:32:48 < mmcgrath> Ok, anything else? If not I'll close the meeting in 30 16:33:16 < mmcgrath> 10 16:33:21 < skvidal> thanks mmcgrath 16:33:29 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting End 16:33:46 < mmcgrath> thanks for coming guys. 16:33:48 -!- frankc [i=824c405d(a)gateway/web/cgi-irc/ircatwork.com/x-c083a3ac2d3221c1] has left #fedora-meeting [] 16:33:50 < mmcgrath> skvidal: any time :) 16:34:16 -!- mmcgrath changed the topic of #fedora-meeting to: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Communicate/FedoraMeetingChannel for meeting schedule 16:35:05 < dgilmore> mmcgrath: you will need to co-ordinate the change so that fedora-packager-setup.sh gets the certs from the new location 16:35:39 < mmcgrath> dgilmore: where is the canonical source for fedora-packager-setup.sh 16:35:43 < dgilmore> koji 16:35:50 < dgilmore> its in the koji package 16:36:38 * dgilmore has been meaning to do up a fedora-maintainer package that has it and pulls in koji/plague/mock etc so someone can easily get a environment up 16:36:40 < mmcgrath> dgilmore: can do, I'll file a bug. 16:39:47 -!- halfline_ is now known as halfline 16:39:56 < f13> dgilmore: do it as a comps group instead. 16:40:09 < f13> dgilmore: can still do 'yum groupinstall fedora-dev-kit' 16:40:16 < f13> and can select it at install time, or graphically from pirut 16:40:23 < f13> (and can be easily included in spins) 16:41:40 < dgilmore> f13: no because we want a place to put scripts etc 16:42:24 < f13> sure, so you can have a package that does that, but also a comps group that includes the koji package, your package, whatever other tools we pick and choose 16:42:26 < dgilmore> f13: fedora-packager-setup.sh does not really belong in koji. 16:42:36 < f13> and we make those defaults, but not mandatory so that at install time you can adjust. 16:42:41 < f13> dgilmore: I don't disagree 16:42:42 < dgilmore> sure that could be done 16:42:57 < f13> I just don't want you to do a 'metapackage' with unnecessary Requires. 16:43:57 < dgilmore> we can make that work 16:44:10 < dgilmore> I should set aside time and just do it

1 0

imtiaz.rahi@fedoraproject.org address not working
by Imtiaz Rahi 06 Sep '07

06 Sep '07

Hi, I am a fedora ambassador from Bangladesh and also fedora FreeMedia contributor. Just today I found out that my fp.org mail address ( imtiaz.rahi(a)fedoraproject.org) is not working. Mail sent to this address returns with an error user unknown. I understand that imtiaz(a)fp.org addresses are not working anymore. But does my imtiaz.rahi(a)fp.org address got deleted or disabled during that process. Can anyone please shed some light regarding this matter. Cheers, Imtiaz http://fedoraproject.org/wiki/ImtiazRahi

2 1

mirrormanager future features
by Matt Domsch 04 Sep '07

04 Sep '07

MirrorManager, for what I really wanted to see by the Fedora 7 release, has been a success. But there are still several gotchas I'd like to iron out before Fedora 8. * The mirrorlist mod_python applet consumes too much memory on the app servers. It basically reads in a 2MB mirrorlist_cache pickle file which is lists, by directory, of what mirrors hold what content. Handy to have, but in mod_python, that blows the RSS size out to ~27MB per process, times all the httpd processes that have run that code, each with their own private copy. Not pretty. The mirrormanager TurboGears backend isn't fast enough to handle all the client requests for mirrorlists, hence I exported the data for mod_python to use. But the mod_python trick takes too much memory. The way out? Split the mod_python applet into two pieces: 1) Yet another daemon, listening on a local UNIX socket, that has a copy of the mirrorlist cache. It calculates the answers to return. 2) The mod_python applet connects to the daemon, passes it's list of args, and gets back the answer list. It handles redirects too. In this way, the daemon can fork() itself if necessary to handle the traffic, but those forks() use copy-on-write memory, and the children will never touch the pickle, so they'll all share mostly the same memory. One copy of the mirrorlist_cache, used by all children. Since I'm saving so much RSS memory here, I can add back into the mirrorlist_cache all the directories which are being omitted now. So, we will be able to return the list for any dir or file that the public mirrors know about, not just a few as we do now. I've got a stab at this, but am still working on the details. I'll want to do some time tests against the new code, to make sure it isn't too much slower for clients, but a quick swag shows it'll be OK; 0.3sec or so per request, even in parallel, which IMHO is "good enough". * Mike's redirection stuff is included in the above already, so that'll be online as soon as the rest is. Now, to find the time before F8... Still to come, provided I find a lot of time (unlikely), or someone else steps up to help: * Designate a way for mirrors to claim themselves to be always up-to-date. Probably will require a sysadmin to set this bit, as it's somewhat dangerous. But there are cases, e.g. a local out-of-line squid proxy, where it makes sense to do it. This change will change the schema, and has repercussions throughout the code, so I haven't wanted to make it lightly. * Some people want metalink support. Conceptually it's possible, and even pretty easy once we've got the daemon above working right. But as noted on f-d-l, it's been 10 weeks since someone asked for it and even sent some code that doesn't quite integrate but was a starting point, and I haven't had time to get to it. It's not looking good for me to add that right now, but I'd be happy to review patches. * I've wanted to add the libgeoip country->continent mappings, so we can fall back netblock -> country -> continent -> global but I don't know C->Python bindings code at all, and need that exported in python-GeoIP for mm to use. * I've got pending a request to change the fedora.repo files to make yum treat the list as in priority order. I really want the continent mappings in place before doing that though... Should we let countries with <3 mirrors return their own lists? Right now if a country has <3 mirrors, the users get the global list back. Anything else people really need to see? Thanks, Matt -- Matt Domsch Linux Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux

6 8

Re: Improving availability and guaranteeing integrity in ISO downloads
by Anthony Bryan 04 Sep '07

04 Sep '07

> http://www.getright.com/seedtorrent.html > > Supported by Azureus, among others. We already have an extensive > HTTP/FTP mirror system to leverage. > > I've noticed, after the initial release rush, torrents end up being > quite a bit slower than just downloading from a mirror. Especially on a > less popular arch. (cough ppc cough...) In the past I've just stopped > the torrent, downloaded the iso from a mirror, then restarted the > torrent to help seed. > > It would be nice to just have this happen automagically. That's how Metalink works with clients that also support P2P networks. GetRight supports this w/ metalink, hopefully KGet will eventually along with aria2. Phex also supports it over gnutella. (Sorry for the late reply) > On Tue, Jun 19, 2007 at 03:40:59PM -0400, Anthony Bryan wrote: > > >On Sat, Jun 09, 2007 at 07:51:20PM +0200, Ruben Kerkhof wrote: > > > > Have you had a chance to look over Ruben's additions? Any feedback? He > > said he re-licensed it to line up w/ mirrormanager. Any ideas/comments > > for features in Metalink that could be of use to Fedora? > > Yes, I took a quick look; I'll be able to do something with this, but > not for the next 4 weeks, as I'm out of the office moving houses and > on vacation, then catching up on real work. :-) checking in after 10 wks :) Will what Ruben submitted be usable? a few more metalink apps have been released. Free Download Manager (Win) has been released under GPL3. DownThemAll 1.0b2 firefox extension is out, and displays more of the info contained in a metalink to the person downloading, like a how many mirrors are listed, logo, description, version, os/arch, and other stuff that could be useful. here are some screenshots the DTA team put up: http://code.downthemall.net/maierman/metalink-test/metalink-feisty.png http://code.downthemall.net/maierman/metalink-test/metalink-xp.png http://code.downthemall.net/maierman/metalink-test/metalink-comment.png there's also Celerius, a GTK python downloader in progress at http://celerius.tuxfamily.org/ if anyone wants to help out. -- (( Anthony Bryan ... Metalink [ http://www.metalinker.org ] )) Easier, More Reliable, Self Healing Downloads

2 1

Problem with sending e-mails to Fedora mailing list
by Bukovansky Richard 03 Sep '07

03 Sep '07

Hi, I'm not sure I asking the right group, but I need help. When I send an e-mail to any Fedora mailing list I'm subscribed to, my e-mails are not delivered back to me even I have set "Receive your own posts to the list?" to yes. My e-mail is delived to everybody in the list, it's showing in mailing list archive. I.e.: https://www.redhat.com/archives/fedora-cs-list/2007-August/msg00073.html What is suprising me more, I'm getting this kind of error e-mails back after while: This is the Postfix program at host smtp-out4.iol.cz. #################################################################### # THIS IS A WARNING ONLY. YOU DO NOT NEED TO RESEND YOUR MESSAGE. # #################################################################### Your message could not be delivered for 8.0 hours. It will be retried until it is 5.0 days old. For further assistance, please send mail to <postmaster> If you do so, please include this problem report. You can delete your own text from the attached returned message. The Postfix program <fedora-cs-list(a)rehdat.com>: connect to rehdat.com[208.73.212.12]: Connection refused Or once I got this one: This is the Postfix program at host smtp-out3.iol.cz. #################################################################### # THIS IS A WARNING ONLY. YOU DO NOT NEED TO RESEND YOUR MESSAGE. # #################################################################### Your message could not be delivered for 8.0 hours. It will be retried until it is 5.0 days old. For further assistance, please send mail to <postmaster> If you do so, please include this problem report. You can delete your own text from the attached returned message. The Postfix program <fedora-cs-list(a)rehdat.com>: connect to rehdat.com[69.25.212.153]: Connection timed out It was working normally before and I'm not sure what can be wrong. Thank you very much. -- Bukovansky Richard richard.bukovansky(a)atlas.cz

2 2

Private builds in koji
by Ignacio Vazquez-Abrams 03 Sep '07

03 Sep '07

A couple of questions that haven't been brought up yet but have recently passed through my mind: 1) Is it considered an acceptable use of resources to use koji to scratch-build private FOSS packages that will be hosted on fedorapeople.org? 2) What about pointing people to the builds hosted within koji? 3) What about building in koji but hosting elsewhere? I think it's safe to assume that building anything non-FOSS in koji is *completely* out of the question. -- Ignacio Vazquez-Abrams <ivazqueznet(a)gmail.com> PLEASE don't CC me; I'm already subscribed

4 4

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

infrastructure September 2007