Infrastructure/Meetings/2006-11-16

= Meeting of 2006-11-16 =


 * Time shown in EST

15:00 Everyone here and ready for the meeting? 15:01 meh 15:01 I'm still exhausted from yesterday 15:01 warren: you can sleep latter 15:01 * mmcgrath is here 15:01 yeah 15:02 but also in another meeting so won't be responding extremely quickly 15:02 I just can't think of topics and lead discussion in my current mental state. 15:02 packagedb abadger1999, c4chris anything on this? 15:02 warren: sit back relace and chime in when you need to 15:03 heh -- I'm writing an email to the list write now. 15:03 i know i need to spend some more time on tha package database 15:03 warren: relax even 15:03 Basicaly, the summit had a lot of ideas about what to do ith the new buildsystem and the package db. 15:03 given the requirements of packagedb to handle the release management stuff that Jesse talked about brew currently doing... 15:03 * iWolf_ here, moreso in another minute or two 15:04 has Jesse effectively described what those requirements are, or reviewed the proposed schema? 15:04 -!- You're now known as iWolf 15:04 This is going to be essential to release management in the merged FE+FC 15:04 f13: ping 15:04 I don't think the current schema encompasses everything that as talked about. 15:04 But it can probably be extended to handle it. 15:04 nod 15:05 abadger1999: yeah the buildsys, packagedb and update system are the most important pieces we need to work on 15:05 I have an importer for extra's owners.list and cvs modules almost written so I'm going to finish that up and get a prototype running. 15:05 Then Jesse and others can take a look and we can fix the schema to handle all the tasks we need to do. 15:06 I'm hoping to have that running by Monday night so Jesse can take a look at it early next week. 15:06 cool 15:06 abadger1999: cool 15:07 let us know if you need anything 15:07 At that time we should be able to break off little subprojects as well -- Get a web designer to help work on the front end, people to write other improters, etc. 15:07 awsome 15:08 abadger1999: your the VCS guy. this has kinda been pushed back a little but still is a high prioritory thing 15:08 it seems that we have made some progress on it 15:08 where are we at  other then dwmw2 and alan cox  bitching at each other? 15:09 I get the sense that f13's pretty happy with mercurial. 15:09 :-) 15:09 f13's mercurial work is awesome. 15:09 mmcgrath: i get that also 15:09 dist-hg using the existing system could be deployed very quickly 15:09 the questions remaining have to do with... do we want to remain with tarball + patch 15:09 I'd be happy to use dist-hg. 15:09 although there's still more work to be done to finish it. 15:09 blizzard wants exploded source 15:10 Exploded source has plenty of advantages. 15:10 < halfline_> as long as the srpms are still tarball plus patches 15:10 But it could have disadvantages as well. 15:10 warren: http://cvs.fedoraproject.org/extras-nov-13-2006.tar.gz ? 15:10 < halfline_> (and not just a new tarball, or a pristine tarball and megapatch) 15:10 oh, nm 15:10 halfline_, yeah, I have no clue how that is supposed to work while maintaining the tarballs as being identical to upstream. 15:10 ignore me :D 15:10 https://wiki.ubuntu.com/NoMoreSourcePackages 15:10 That is a proposal that went to the ubuntu lists. 15:11 -!- cyril_FC6 [i=user78@62.Red-217-127-20.staticIP.rima-tde.net] has joined #fedora-admin 15:11 warren: im against exploded tarballs 15:11 ubuntu addressed the split vs. megapatch issue? 15:11 Ubuntu hasn't been so friendly toward upstream as us 15:11 i like the current approach of pristine tarballs, + small patches 15:11 I am against exploded source if that makes it difficult to logically separate patches and feed them upstream one at a time. 15:11 small patches == easier route to upstream 15:12 Ubuntu, being debian based, has to live with megapatches at the root level (within the dpkg) but the proposal I posted makes use of DRCS branching to simulate small patches to the developer. 15:12 where did you post that proposal? 15:13 Those will either be made into a megapatch when it gets turned into a .deb or they won't have a source package. 15:13 ubuntu AFAIKT only cares about  themselves  and  doesnt care to get things upstream  weather that is debain  or the upstream project 15:13 https://wiki.ubuntu.com/NoMoreSourcePackages 15:13 DRCS = Directory of Radioactively Contaminated Sites 15:13 I'm Feeling Lucky! 15:13 The Ubuntu proposal. 15:13 oh 15:14 Note that the loomify plugin to bzr doesn't exist yet. 15:14 It's a specification and the developer is needign to get others to buy into the idea before going forward. 15:14 the current system i feel works well. 15:14 loomify is comparable to quilt? 15:15 Maybe we need to be debating this in the yet to be created fedora-scm group 15:15 dgilmore: I am unhappy with it -- I run quilt on everything I check out of our cvs. 15:15 I shouldn't have to do that. 15:15 abadger1999: hmm ok i would ask what you gain by that but this is not the place or time for it 15:16 lets move onto DB1 upgrade 15:16 warren: Most likely. Like I said in fesco, doing exploded sources is exciting but also much different than what e do now. 15:16 iWolf: status 15:16 < iWolf> dgilmore: waiting for the hardware swap 15:16 dgilmore, from a packager POV the current tarball + patch system works well, but many software developers feel it is an ugly layer of overhead. 15:16 < iWolf> the proxies getting taken out of rotation so we get the new Dells in place again. 15:16 < iWolf> Then I'll be rolling forward on that (beyond what I did for prep to a xen guest we worked with). 15:16 iWolf: ok cool i guess we will have to bug pick see where he is at 15:17 < iWolf> dgilmore: I think he was in meetings this week. 15:17 iWolf: are you using the current db2 xen test 15:17 < iWolf> dgilmore: he said if he didn't get it this week he would get it next. 15:17 iWolf: i believe so 15:17 < iWolf> dgilmore: not at the moment no. I had already restored the dbs to play with them and that went well. 15:17 < iWolf> dgilmore: if we need to re-use it feel free. 15:18 iWolf: i dont know if we do or not but its there if you need to do anything 15:18 lmacken: about ? 15:18 < iWolf> dgilmore: Cool. I won't keep anything that can't be replaced on it if people find the need to re-use. 15:18 * mmcgrath notes we should be getting the new dells back soon. 15:19 i guess lmacken is not here 15:19 lmacken: when you get time can you email the list with an update of firewall stuff 15:20 warren: xen boxes are up i guess we can mark that as doen 15:20 done 15:20 warren: did you hear back on a ip for xen3? 15:20 hm... 15:20 * lmacken is here (but on the phone) 15:20 apparently he lost it 15:20 warren: ok 15:20 he... doens't tend to read e-mail carefully enough 15:21 I don't remember what this was for anymore 15:21 warren: well i started building xen3 and used 10.8.34.77 which is the ip after xen2 15:21 warren: it is for tha xen host for xen builder replacing hammer1 15:21 if need be ill change it hammer1's ip 15:22 dgilmore: yeah, I'm going to finish the deployment of pyroman next week. I'll keep the list up to date. 15:22 lmacken: cheers 15:22 dgilmore, ok, so just ask to point xen3 (internal only) to 10.8.34.77? 15:23 warren: yes please 15:23 k 15:23 * warren sending ticket 15:23 warren: lets also ask for xen4 for the new build box on its way 15:23 and a xenbuilder2  to use on it 15:24 -!- daMaestro [n=jon@fedora/damaestro]  has joined #fedora-admin 15:24 10.8.34.77 and 10.8.34.78 I'm guessing 15:24 < daMaestro> any info on this -> https://admin.fedoraproject.org/tickets/customer.pl?Action=CustomerTicketZoom&TicketID=384 ?? 15:24 warren: yeah 15:24 dgilmore, where is the new build box coming from? 15:24 warren: not sure Dell donated it 15:24 cool 15:24 its a dual dual core xen box 15:24 with 4gb ram 15:25 mdomsch said it should be there any day 15:25 s/sen/seon/ 15:25 gahh its got xeon cpus 15:26 xen3.fedora.phx.redhat.com 10.8.34.77 15:26 xen4.fedora.phx.redhat.com 10.8.34.78 15:26 xenbuilder2.fedora.redhat.com (new external IP) 15:26 xenbuilder2.fedora.phx.redhat.com (new internal IP) 15:26 ? 15:26 dgilmore, xeon x86_64? 15:26 warren: as far as i know yeah 15:26 these requests look good? 15:26 daMaestro: you didn't include the stream ;-) 15:27 warren: yeah that looks good 15:27 k 15:27 < daMaestro> mmcgrath, ok.. so i just need to submit the stream url? 15:27 daMaestro: yep 15:27 < daMaestro> dgilmore, ok.. thanks.. i will get it setup 15:27 < daMaestro> we are just in the testing phase now 15:27 < daMaestro> i'm working on getting the metaweblog api working 15:28 xenbuilder2 needs 8887-8889, anything else? 15:28 (same ports open as hammer1)? 15:28 warren: thats it 15:28 warren: same as the other builders 15:28 k 15:28 warren: we should get hammer1 and 3 taken out 15:29 OK. 15:29 Legacy buildsys is setup and ready 15:29 it needs push scripts 15:30 dgilmore, f13 needs to decide what happens there 15:30 its future depends on what legacy wants to do 15:30 warren: :)  i know 15:30 given that he proposed discontinuing legacy during the summit 15:30 < daMaestro> is atom, rss or rdf better? 15:30 I didn't fully agree with it, and the legacy team is debating it on their list 15:30 EPEL buildsys is up. i have succesful builds for EL-4 and EL-5 15:30 < daMaestro> *for you guys 15:30 daMaestro: rss 15:30 * lmacken has to run to another meeting. 15:30 daMaestro: atom has a spec 15:31 < daMaestro> ok 15:31 no updates from my end.. finals are done tomorrow night, and I'm going to start hacking on the update system and get the rest of the firewalls deployed next week 15:31 * lmacken & 15:31 lmacken: thanks and good luck 15:31 dgilmore: thanks 15:31 dgilmore, hammer1 hardware is becoming xen3? 15:31 warren: yes 15:31 k 15:32 we will need to rename consoles. etc 15:32 that was part of my ticket 15:33 warren: :D thanks 15:33 OK anyone  have anything to add? 15:34 floor is open for whatever 15:35 Damnit. 15:35 Ok, this isn't good 15:35 http://www.google.com/search?hl=en&lr=lang_en&q=admin.fedora.redhat.com+%22dump-group%22&btnG=Search 15:36 no thats not 15:36 do we need a robots.txt file on admin 15:37 I'm going to enable authentication again 15:37 the issue is that some stuff breaks when I do that, we'll have to create a fake user for the scripts. 15:38 mmcgrath: :) yeah   accounts to the buildsys was the biggest 15:38 Can you limit by ip instead? 15:38 should be able to in the apache configs 15:38 -!- daMaestro [n=jon@fedora/damaestro]  has quit ["Leaving"] 15:39 Right.  So then the buildsys can get access but not google. 15:39 robots.txt is the only way you'll get that out of google's cache in the future. 15:39 we will need a robots.txt  and allow only from certain ips without auth 15:39 abadger1999: not easily because we're behind a balancer and reverse proxy. 15:39 nod 15:40 hmmm 15:43 anyone got anything else? 15:44 if not ill close the meeting in 60 seconds 15:44 I've got one thing 15:44 mmcgrath: sure 15:44 warren: whats the easiest way for us to communicate with the developers to find out what is most important from infrastructure as far as releases go? 15:45 mmcgrath: yes  that is important 15:45 we won't always get everything done but I'd hate for us to be working on something that isn't important to the project while something tha t was important didn't get done. 15:45 mmcgrath, warren: what can we do so that non RH people can file tickets for stuff we need RH IS to do for us? 15:45 -!- paulobanon [i=keefejoh@gateway/web/cgi-irc/ircatwork.com/x-82df935d0609c1a1] has joined #fedora-admin 15:45 mmcgrath, in the case of fedora merge, f13 needs to communicate effectively with infrastructure. 15:46 and when we make the hire for full-time fedora infrastructure lead, that role becomes accountable to all communication with developers 15:46 howdy 15:46 sounds good. 15:46 hey paulo, we're just wrapping up 15:46 warren: what was the word on that position? 15:46 http://ircatwork.com 15:46 for everyone behind fw 15:46 :D 15:47 dgilmore, mspevack and I are still working on it. 15:47 warren: ok cool keep us in the loop 15:47 there is a huge and burning need for the role to happen 15:47 but no guarantees in anything =( 15:47 yep 15:48 alrighty, thats all I've got dgilmore, feel free to close whenever ;-) 15:48 ok closing meeting in 30 15:48 === mark closed meeting ===