21:01:30 #startmeeting Spins SIG 21:01:37 wow ;-) 21:01:41 #chair nirik 21:01:48 I have a tentative agenda at http://fedoraproject.org/wiki/Spins_SIG_Meeting_2009-07-27 21:01:59 #link http://fedoraproject.org/wiki/Spins_SIG_Meeting_2009-07-27 21:02:17 #topic Recurring Spins Process 21:02:37 ho, i forget something 21:02:41 who's here? ;-) 21:02:53 present 21:02:59 nirik: thanks for the reminder 21:03:21 * nirik is here somewhat. ;) 21:03:34 ok, on to our first topic then 21:04:05 When I said it was basic I meant it, I've not spent too much attention to it this week but here's a basic flow for recurring spins: https://fedoraproject.org/wiki/Spins_SIG/Drafts/Recurring_Spins_Process 21:05:30 any comments? 21:05:50 When should spins get added to the per release category? 21:06:06 Is that at GA, branch, something different? 21:06:19 when they are accepted as official spins, and when they are +1 at Feature Freeze 21:06:53 i added that just now 21:07:03 * huff here 21:07:10 hi huff ;-) 21:07:54 is there anything else we should think of wrt to recurring spins? 21:08:08 kanarip: looks pretty straight-forward to me :) 21:08:18 well ok then 21:08:32 yeah, seems fine, we can grow it if more is needed. 21:08:38 +1 for the draft, its pleasantly to the point 21:08:44 can i get some +1's and then move it into the official Spins_Process (a little more verbose then this version of course) 21:08:50 Somewhere should document how to adjust the spin web pages for release. 21:09:01 +1 from me, too ;) 21:09:28 brunowolff, Spins_SIG_Wrangling_Documentation? ;-) 21:09:56 +1 for recurring spins process 21:10:18 +1 here, it makes sense to me. 21:10:32 That seems like it would be better as it can get referred to from a few places without duplicating things. 21:10:36 #agreed Recurring Spins Process @ https://fedoraproject.org/wiki/Spins_SIG/Drafts/Recurring_Spins_Process to go into the official Spins_Process 21:10:58 #action kanarip to integrate Recurring Spins Process in Spins_Process 21:11:04 #topic Discontinued Spins Process 21:11:04 #link https://fedoraproject.org/wiki/Discontinued_Spins_Process 21:11:14 brunowolff, take it away sir! 21:11:36 I think the main things are determining when to continue a spin discontinued and 21:11:57 what to do once one has been determined to be discontinued. 21:12:25 right 21:12:32 As for what needs to get done, I think adjusting the spin web page and dealing with the ks file should cover it. 21:12:57 wrt. the former; you hit it right on the spot, although a week may be a little short 21:13:04 I made some guesses for both, though the ks part depends on the ks reorganization to some extent. 21:13:27 A week was minimum. Is a month too long? 21:13:45 wrt. the latter; if the kickstart is declared dead then i guess there's no reason in shipping it to the user in the form of a package either 21:14:12 how about we say, 2 weeks (2 Spins SIG meetings) 21:14:27 My thoughts on the ks is that we would be less likely to mess up history by moving it to an attic directory in the event that 21:14:27 sounds fine to me. 21:14:38 someone wants to resurrect it later. 21:14:49 we don't mess up history anyway since it's git 21:14:56 It also might be more likely to be resurrected if it is still somewhat visible. 21:14:59 it's practically impossible to mess up history ;-) 21:15:40 If people recover the deleted version rather than recreate it as a new file. 21:16:01 I was thinking the odds of that happening were nonneglible, 21:16:41 well we could always ship'em but how do we make it clear it's dead? 21:16:49 put it in a spin-kickstarts-dead package? 21:17:32 I updated the delay to two weeks in the docuement. 21:17:47 ok, that sounds fine to me 21:18:09 i say we just remove unmaintained kickstarts 21:18:14 kanarip: maybe spin-kickstarts-retired ... -dead sounds morbid to me for some reason :/ 21:18:24 I was (perhaps incorrectly) thinking the spin-kickstarts package would have directories reflecting those in git (that don't exist 21:18:39 yet, but that we were thinking of doing). 21:18:39 we can leave the page in Spins_in_Development and link to the most recent version, and it'll exist in branches, but not in master 21:19:26 the spin-kickstarts package (which imho is a bad name to begin with) will have some kind of directory structure, but that's the next topic 21:20:04 ok, +1 for the draft as it is now 21:20:05 If we end up getting a lot of them, we may need a new category. But for now spins_in_development should be OK. 21:20:08 anyone else? 21:20:33 Let me adjust the ks disposition text. 21:21:36 +1 for doc as is. 21:23:05 nirik, maxamillion, sdziallas, how about you? 21:23:40 +1 here... seems fine to me. 21:23:50 they can get the ks back from git if it gets revived later. 21:23:56 ok, that's a majority 21:24:01 +1 too 21:24:06 #agreed Discontinued Spins Process @ https://fedoraproject.org/wiki/Discontinued_Spins_Process 21:24:06 #action brunowolff to integrate Discontinued_Spins_Process in Spins_Process 21:24:07 #topic Review splitting up spin-kickstarts from kanarip 21:24:07 #link http://fpaste.org/paste/20065 21:24:07 'fedora-kickstarts': Official Fedora Spins 21:24:08 'custom-kickstarts': Kickstart files for Custom Spins (not official) 21:24:10 'l10n-kickstarts': Localized version of Fedora Spins and Custom Spins 21:24:23 I can type really really fast as you can see 21:24:24 ;-) 21:26:57 Are these going to be directories in git AND in the spin-kickstarts package or just in git? 21:27:53 This important for includes. 21:27:56 they are going to be separate RPM packages and separate directories in git if it's up to me 21:28:28 But even separate packages could potentially be in the same directory. 21:28:40 Also we need to worry about dependencies. 21:28:44 seperate packages? or subpackages? 21:28:54 I'm ok with it, I think it makes sense to have them separate so that there is a clear definition for what is "official" and what is "an accepted custom spin" 21:29:07 Presumably the l10n package would depend on the fedora-kickstarts package. 21:29:30 But we may need rules for the custom-kickstarts being able to include stuff. 21:30:12 brunowolff, that's not actually a worry, everything requires fedora-kickstarts, l10n- requires all 21:30:24 nirik, "sub" packages but named differently 21:30:34 "spin-kickstarts" is going to have it all 21:30:43 Can a custom spin include an l10n ks? 21:30:46 that sounds fine to me. 21:31:00 brunowolff, i think the official broffice.org spin does so already 21:31:04 I am not saying that it should be able to, but we should document limitations. 21:31:31 the localized versions are sort of a Free for All 21:31:48 there's limitations to what one can do, but they are not actual spins and they are not actual custom spins 21:32:15 either way, these same limitations apply currently 21:32:23 or *lack of limitations* 21:32:36 My question is are custom-spins allowed to depend on l10n spins? That has implcations for rpm dependencies and 21:32:44 wrangler documentation. 21:32:45 shuffling the location on the filesystem and changing the name of the package doesn't change that 21:33:05 well, custom-spins as a package is not going to depend on l10n-spins 21:33:33 Then we should document that custom spins may not include l10n ks files. 21:33:41 it's either a spin, or it's a different l10n version of an existing spin in custom- or fedora- 21:33:46 does that make sense? 21:33:51 kanarip: but wouldn't that break the broffice spin if its already using the l10n as an include? 21:34:05 maxamillion, it's not using the l10n as an include 21:34:24 OK, then I think we are on the same page. 21:34:24 it's a custom spin that just so happens to also do l10n because it's purpose is limited to pt_BR 21:34:42 alright then 21:34:55 But we should add a wrangler note to verify that custom spins do not include l10n ks files. 21:35:02 kanarip: ah, I misunderstood 21:35:05 so the naming makes sense? fedora-kickstarts, custom-kickstarts, l10n-kickstarts, and spin-kickstarts to wrap it all up? 21:35:25 That seems pretty reasonable. 21:35:28 brunowolff: that should be the sigs job to check the spin when it's approved here, IMHO 21:35:37 * nirik nods. Makes sense to me. 21:35:41 brunowolff, the ks itself is actually a review item for the Spins SIG, the wrangler just oversees the process 21:36:00 ok, can i get sufficient +1's? 21:36:03 Right, but the wrangler does the checking on our behalf and it should be on the checklist. 21:36:04 +1 from me ;-) 21:36:19 +1 here 21:36:20 I guess I have it backwards. 21:36:29 +1 for layout proposal 21:36:36 +1 21:37:19 and +1 here ;) - makes sense to me! 21:37:47 #agreed split up the crown jewels (spin-kickstarts package)! 21:37:55 #action kanarip to build and release to rawhide 21:38:13 * huff has to head to class, I will check logs later if yall need any thing send me an email 21:38:22 #topic Spins_in_Development, Spins_Ready_for_Wrangler, Spins_Ready_for_SIG 21:38:25 https://fedoraproject.org/wiki/Category:Spins_Ready_for_Wrangler 21:38:33 this is the first category on our list 21:38:36 the oVirt spin 21:38:38 i guess i can stay for a bit 21:38:53 im still waiting on the pacakges to get into rawhide 21:38:54 huff, that's yours ;-) 21:39:01 as soon as they are ill send an email to the list 21:39:08 the Feature Freeze is tomorrow :/ 21:39:14 i know 21:39:22 but also, how can we test this? 21:39:46 is this intended to go the full nine yards? trademark approval? release by fp.o officially? 21:39:52 or is this a custom- spin? 21:40:02 yes we would like the whole 9yards 21:40:14 is that going to be a problem with the current time frame 21:40:30 im hoping to have everything complete by tomrrow 21:40:35 ok so we need a description of the flow with which QA or anyone else can test if this spin works 21:41:04 i can only assume that since this is a node i also need a management system to control this node, no? 21:41:11 its just going to be a regular spin/livecd image built with our ks file.. is that what you mean 21:41:21 kanarip: it has a stand alone mode 21:41:36 with a limited menu 21:41:43 to run w/out a management serbver 21:41:45 server 21:41:49 but if i spin this, how do i test whether it does the ovirt it's supposed to do? 21:41:58 but the server team is also working ot get the ovirt-server in f12 as well 21:42:17 kanarip: you can boot it and run guest form the console 21:42:23 it has a 21:42:40 "fist boot" style menu to configure networking and run guests 21:42:51 i use first boot lightly 21:43:07 it has a menu to do basic configuration with out the server 21:43:20 you need to write that down in the Scope / Testing section 21:43:25 kanarip: will do 21:44:09 also, technically, i'm afraid that fixing dhclient-script is out of the question 21:45:25 rpm -e --nodeps is also not allowed 21:45:42 kanarip: ok ill look at the we did that dude to an selinux issue however it may have been fixed in a policy updatge 21:45:46 update 21:46:04 i rebmmber seeing a bz about that 21:46:13 removal of kernel drivers needs to be motivated thoroughly 21:46:24 if there's a bz then it's a dependency for you 21:46:45 wow, removing .py files 21:47:19 i think it was resolved and fixed in ana update to selinux-core-policy 21:47:28 Dan is usually pretty good at doing prompt selinux policy updates for problems. 21:47:50 huff, how does using the aos.ks as a basis sound? 21:48:03 yea dhclient script was doing a mv not a cp -a and causing the selinux context to change 21:48:06 if you want minimal to build on, they're after the exact same thing 21:48:47 kanarip: yea we added some more blacklisting to reduce size 21:48:52 but you can't really modify anything rpm puts on the filesystem without a very good reason, that's our rule of thumb 21:49:15 when in doubt, we ask other people but that obviously costs time 21:49:23 which we don't have for f-12 anymore 21:49:43 aos has all o pythong and yum we did not want all that and wanted ot keep it under 256M to fit on small usb key 21:50:20 but you can remove override it with -python* and -yum nowadays 21:50:34 the aos is ~130MB iirc 21:50:55 Isn't yum required for spins? 21:51:22 we've not said it was a requirement but removing stuff from @base raises eyebrows, yes 21:52:55 there was a decision once on selinux that needed to be in enforcing, or permissive with very good reasons 21:53:02 either way; let's move on 21:53:14 https://fedoraproject.org/wiki/Category:Incomplete_Spins 21:53:19 If it is anticipated that people will use the install to disk feature, not having yum is going to be a pain. 21:53:19 most of which are recurring spins 21:53:59 Can we talk about Games Spin this week, since I am going on vacation? 21:54:01 as you can see on http://www.kanarip.com/revisor/ on 20090725 this all built on i386 21:54:09 ok will Ill try to touch it up tonight and send an email to the list where we can continue this conversation via email, i guess a congency plan is to use aos as ase and add our packages and minimual configuration to get a working node 21:54:16 we need to talk about them all they need to be approved by tomorrow 21:55:42 so we need to talk about the recurring spins on that page 21:56:15 kanarip: all spins (including the recurring ones) need to be approved by tomorrow? uh! 21:56:25 https://fedoraproject.org/wiki/AOS_Spin +1 21:56:30 no rocket science there 21:56:54 +1 21:57:01 :) 21:57:31 https://fedoraproject.org/wiki/BrOffice.org_Spin oversized, again, but what else is now in the midst of a release cycle 21:57:58 it's 839 M right now, we need to contact igor about that 21:58:04 #action kanarip to contact igor 21:58:12 That's a lot over for a CD spin. 21:58:22 It might be hard to get back under. 21:58:42 sdziallas, you're in the list too: https://fedoraproject.org/wiki/Education_Spin 21:59:01 sdziallas, also oversized at ~766M 21:59:46 https://fedoraproject.org/wiki/ElectronicLab_Spin gets a +1 from me 21:59:47 kanarip: yup :)... oh well, that's well possible. heh. I've a new concept some time ago in the repo... if no one objects, we'd move to a DVD spin. 22:00:17 sdziallas, there's limited slots for DVD spins and limited slot of CD spins 22:00:37 i think i talked about that with f13 once but i can't remember what the numbers are exactly 22:00:59 kanarip: oh, okay... didn't knew that 22:01:06 brunowolff, your Games Spin is 3.9G on i386, nicely done ;-) 22:01:32 Yeah I didn't add much, just colossus. 22:01:36 KDE spin is slightly oversized but I see no problems there either 22:01:54 I gave up on lzma for this go around, so I haven't been looking to add stuff. 22:01:55 and of course the XFCE spin is undersized, nirik ;-) 22:02:05 I do want to do more work on the test cases. 22:02:20 kanarip: and we added roughly 6 xfce panel plugins as well as about 10 claws mail plugins :) 22:02:26 58 minutes before another compose starts that will not run out of loop devices! 22:02:49 so, does anyone object against moving these spins to Category:Spins_Fedora_12? 22:03:43 +1 on adding to F12. 22:04:05 anyone else? 22:04:18 * nirik is good with that, 22:04:34 +1 (I guess the Edu Spin needs to be discussed re dvd size?) 22:04:49 sdziallas, yes 22:04:57 please aim for a CD size or i'm going to get into trouble 22:05:24 #action kanarip to add spins to spins fedora 12 category 22:05:35 kanarip: heh. that's going to be funny... but okay, will do 22:05:37 kanarip: does somone want to update https://fedoraproject.org/wiki/Releases/12/FeatureList#Customized_Fedora_12_Spins to point to the F12 spins? 22:05:49 poelcat, ok 22:06:10 sdziallas, comment some stuff out while we're discussing what to do if you will 22:06:34 kanarip: sure... 22:13:49 Are we done for today? 22:13:54 kanarip: discussing right now what to take away... 22:14:14 ohw sorry, yeah i think we're done unless someone has something to add? 22:14:22 nirik, want to say something on spin1? 22:14:29 kanarip: when do you need it to be cd-size? (today, right?) 22:16:57 I'll be getting back on the 10th and may or may not make that meeting. 22:17:45 I am hoping to handout a few games spin disks at the board game convention I am intending, so I'll be doing some spin stuff yet this week. 22:20:32 kanarip: sorry, got pulled away on work. 22:20:47 kanarip: I was hoping to look at spin1 and see about setting up a nightly script to make isos... 22:21:08 possibly maxamillion would be interested in helping on that too. 22:21:25 nirik, i'm afraid at some point we'll run into the exact same problems i do; loop devices and the like 22:21:53 kanarip: I suppose we could reboot it when it hits that... 22:22:56 will see if I can come up with something. ;( 22:23:32 lsof might help find process ids. 22:24:12 I tried that once though and didn't have much luck figuring out which processes had loop mounts open. 22:24:45 I wasn't sure that there was something else going on with the loop mounts and just figured it was easier to reboot. 22:27:05 yeah, I have seen it and been unable to track down whats keeping the mount busy. 22:27:11 Oh well, shall we end the meeting? ;) 22:27:19 +1 22:30:50 brunowolff, i can find the process id's i just can't kill them 22:30:59 #endmeeting