From Fedora Project Wiki

Fedora Release Engineering Meeting :: Monday 2008-03-17

F9 Beta Status

  • anaconda regressions over the weekend
  • url method is broken--traceback when getting bugurl fixed with updates.img
  • pushing hard for a Thursday release
  • space issues with PPC compose fitting on media
  • if we have to slip beta, we slip for anaconda issues and let rawhide move on

IRC Transcript

-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering Meeting - Roll Call14:07
* jwb is 1/2 here14:07
* poelcat here14:08
f13ping: jeremy notting jwb spot wwoods rdieter lmacken poelcat warren14:08
* jwb is 1/2 here14:08
f13jeremy is in class, spot is MIA14:08
* wwoods here14:09
* warren here, although might lose my laptop 14:11
* jwb adjourns meeting14:14
jwbseriously, i think we have lots of stuff to do.  summary:14:16
jwb1) anaconda has issues14:16
jwb2) we have space issues on ppc that aren't really ppc related growth14:16
wwoodsthere's an updates.img for that, I think14:16
jwbwwoods, sure14:16
jwb3) OOMs during install14:16
wwoodsf13: do you have a URL for the updates.img? or a bug number that contains that image/URL?14:17
warrenthere are remaining mkinitrd on software raid issues14:17
warrenI'm trying to track that one down, but managed to crash anaconda in the process14:17
wwoodsnote to self: actually write code for 'grabupdates' anaconda flag14:18
* jwb gets 40414:19
f13jwb: yeah, it's an internal host14:20
* f13 wonders why is not letting me log in14:21
f13oh well14:21
-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering Meeting - Beta Status14:21
f13today, not so good.  We've had some anaconda regressions over the weekend14:22
f13url method, busted.  traceback when getting bugurl (see updates.img discussion above)14:22
f13Trying to get an install to finish so that we can see if encrypted stuff works with the new mkinitrd, but running into OOM issues.14:22
f13but I'm still going to push hard for a Thursday release14:23
warrenany idea what exactly is using all memory?14:23
f13Today I'm oging to pre-stage some content on the netapp for beta, basically the exploaded trees and some isos from today.  They won't be available to the mirrors, but it'll save us tonns of time once we actually have gold content14:24
f13warren: no, it happened while I was gone at lunch.  I suspect either a runaway %post or bad yum caching or something else.14:24
wwoodsf13: installing from media? or using the updates.img?14:25
f13with updates.img14:26
f13media needs updates.img too14:26
f13hrm, this is strange14:27
f13free shows 1052216 for swap, but 490548 for mem14:27
f13I thought this machine had a gig, not 512 meg14:27
jwbstill OOM with 512MiB is evil as well14:27
warrenf13, AGP grabbed 512MB aperture?14:27
f13warren: 512 would be a /big/ grab14:28
jwbf13, also didn't you say you installed the PS3 over the weekend?14:28
f13jwb: yes, I did, text install14:28
jwband this is not a text install?14:28
wwoodsmirrored updates.img as
f13I couldn't ssh to for some reason14:29
f13so anyway, I'm going to keep banging on that, while uploading content to the netapp hidden away14:30
f13ideally we'd have the golden package set as tomorrow's rawhide, or else we slip14:30
f13and I will be sad panda14:30
wwoods(someday we'll have code in anaconda to optionally autofetch$TIMESTAMP.img)14:30
wwoodsyeah, my parents decided to come visit this week14:31
wwoodsso if we have to extend the freeze another week.. bluh14:31
f13well, I think if we have to slip, we're going to be slipping for anaconda stuff, and we can let rawhide move on14:32
f13we'll just pull anaconda builds for beta testing14:32
jwbf13, i'm not seeing tons of stuff that we can drop when looking at the livecd configs.  but i'm a kickstart idiot14:33
warrenf13, anaconda and mkinitrd possibly14:33
f13warren: yes, and booty from peter for EFI grub hotness14:35
f13Still haven't been able to do a full media install, or nfs iso install14:36
warrenf13, are we backing ourselves into a corner where beta is delayed for another 2 weeks while rawhide is well past it? =)14:36
f13want to try that later today14:36
f13warren: we're already in that position14:36
jwbf13, skvidal is awesome and added a --size option to repodiff for me.  showing nothing hugely significant for individual packages14:39
f13so maybe we just have to drop something14:39
jwbrdieter, what is kipi-plugins?14:39
jwbf13, yeah maybe :(14:40
f13maybe drop some of the games?14:40
jwbf13, was thinking something like that.  let me run this again on my compose from today and i'll email rel-eng with some results/suggestions14:40
rdieterjwb: addons to kipi-aware apps, graphic/photo managers mostly, like digikam, kphotoalbum14:40
jwbit grew lots14:41
rdieterjwb: since, when?14:41
jwb243K/144K ppc64/ppc14:41
jwb"lots" is a relative term at the moment14:41
rdieterkipi-plugins are definitely optional functionality, so could be a good candidate for chopping block14:42
jwbrdieter, ok.  other than that, most of the additional space is coming from l10n packages for KDE14:43
jwbabout 75MiB worth14:43
jwbat least so far anyway14:43
rdieterthose be huge, kdelibs-apidocs is a monster too (but hopefully not landing on any media)14:43
f13ok, so other than ppc size issues, where we may have to trim some packages out of hte manifest, is there anything else beta releated to talk about?14:45
jwbneed to test installs on multilib machines14:45
jwbyum policy was updated14:45
jwbwwoods, f13: how often have you done x86_64 installs?14:46
f13I'm trying them every day14:46
f13I have one rebooting right now14:46
f13but it was software raid install so it might fail14:47
f13or not...14:47
jwbok, was just curious since that's the only other arch that the policy would impact14:47
wwoodsall the time, yeah14:47
f13for a default install, my x86_64 still has 65 i386 packages on it14:48
f13I thought we wouldn't get that with the new yum default.14:48
jwbskvidal, ?14:48
skvidalf13: dependencies still pull in somethings14:48
skvidalmultilib_policy only impacts what will be selected during an 'install'14:49
skvidalit does impact what gets pulled in due to deps14:49
f13skvidal: this is a clean install...14:49
f13theoretically nothing should depend on i386 content...14:49
skvidalthat first word is what matters14:49
f13oh fun..14:50
skvidaldo you have the depsolve output?14:50
f13removeing \*.i386 does bring stuff out14:50
f13skvidal: anaconda doesn't save depsolve output anywhere14:50
skvidalit brings out x86_64 stuff?14:50
skvidalf13: then do it in an installroot14:50
skvidalbut I'll wager dollars to donuts it's getting pulled in as a dep14:50
f13looks like PyOpenGL.noarch, gnome-games.x86_64, pygtkglext.x86_64, totem*.x86_64, xorg-x11-drivers.x86-6414:50
f13will debug farther14:51
f13skvidal: I blame YOU!14:52
f13lets kick this over to #fedora-devel14:52
skvidalf13: excellent14:52
* poelcat fills the silence14:56
poelcatused to determine ship/noship of beta?14:57
wwoodsused for all releases.14:57
wwoodsbeta being non-final, "SHOULD" items can be ignored14:57
* poelcat saw 2007-nov last change... wanted to make sure14:57
wwoodssuggested updates are welcome14:58
wwoodsbut nobody's suggested anything to me14:58
poelcatlooks good to me... i just don't see how we can do it all to be ready on thurs14:59
poelcatin light of current situation15:00
poelcatwwoods: as i understand it, we can't even start testing yet.  Am I missing something?15:03
poelcattesting == verify to release critiera15:04
wwoodswe have an updates.img15:05
wwoodsand we've tested half of the stuff on there already15:05
wwoodschanges in mkinitrd are not going to affect VNC15:05
wwoodsso that testing is still valid15:05
f13poelcat: we have to test, to see what's broken15:11
f13we can't very well wait until nothing is broken to ... test15:11
f13ok, ending the meeting so that work can continue.15:17
-!- f13 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 for meeting schedule15:17

Generated by 2.3 by Marius Gedminas - find it at!