[00:01] ScottK, pitti: It's not quite that easy, I'm afraid. At the moment it's hard-coded to backports, but that will probably change in the next few months as we refactor for derivative distros. [00:05] slangasek, breaking for dinner. back on in an hour or so... [00:09] skaet: ok, bon appetit :) [00:11] so, bug #753924 causes ph5p-fpm to fail to install every time. Should I make this a zero-day SRU? [00:11] Launchpad bug 753924 in php5 (Ubuntu) "package php5-fpm 5.3.5-1ubuntu6 failed to install/upgrade: подпроцесс установлен сценарий post-installation возвратил код ошибки 1 (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/753924 [00:12] aw n/m looks like zul fixed it a month ago. ;) [00:12] actually.. no.. looks like the fix was ineffective [00:16] SpamapS: new packages should follow the FFe process, yes. bug #753924, I think there's probably room for that post-beta2 if you can have a fix ready to go by then [00:16] Launchpad bug 753924 in php5 (Debian) (and 1 other project) "package php5-fpm 5.3.5-1ubuntu6 failed to install/upgrade: подпроцесс установлен сценарий post-installation возвратил код ошибки 1 (affects: 1) (heat: 6)" [Unknown,Unknown] https://launchpad.net/bugs/753924 [00:16] slangasek: alright. ty on both answers. :) [00:25] Hi, would someone be able to approve eucalyptus - i'd quite like it to be in the next ISO spin. Thanks. [00:26] SpamapS, it seems still quite early to be thinking about zero day SRU's, no? [00:27] Daviey: well the wording of the email about the freezes suggested there are no more days where the archive is opened for anything except fixes "requested by the release team".. but maybe I read it wrong. [00:27] yeah I read it wrong [00:28] 4/14 - 4/21 should be open for important fixes [00:30] https://lists.ubuntu.com/archives/ubuntu-release/2011-April/000190.html [00:35] SpamapS, Fixes for universe (php5-fqm) should still be able to be uploaded until 26th... but perhaps sooner might be better.. :) [00:35] php5-fpm is universe? [00:36] slangasek, looks like the daily's weren't disabled.... [00:36] ooohhhhh [00:36] how can bin packages from php5 be in universe? [00:36] * SpamapS always thought the whole thing goes into main when the source goes in [00:38] SpamapS: binary demotion [00:38] SpamapS: it's common for transitional packages [00:38] php5-fpm is not transitional [00:39] its quite possibly the most popular way to run php at the moment [00:39] SpamapS, Ah, scrub that - the source is in main.. Sorry for that! [00:39] Daviey: but I can see why you'd say that since the binary itself is in universe [00:41] SpamapS: I think you just have to ask if you want it per https://wiki.ubuntu.com/MainInclusionProcess note 2 [01:11] skaet: are they disabled now or do you need me to do it? [01:13] slangasek, need you to please. [01:13] also, xubuntu, mythbuntu didn't seem to build. [01:16] skaet: dailies are disabled and I didn't do it [01:18] checking on xubuntu [01:19] I see xubuntu 20110411.1 alternate built, 20110411 desktop looks like a build failure [01:20] * skaet nods [01:20] yes, xubuntu alternate posted already. [01:20] ah, mirror sync issue at the time of the CD build; retriggering after I check on mythbuntu [01:20] same problem [01:21] ok, kicking off the ISO builds for both - shouldn't need to respin the livefs for this issue [01:21] coolio [01:40] SpamapS: Yes. Yes you do. [01:54] slangasek, can you see if edubuntu dvd is same problem (mirror synch) it appears to have failed as well. [01:56] I don't see an edubuntu build log for today [01:56] did kubuntu dvd already finish? [01:57] kubuntu dvd doesn't have images [01:57] was just checking [01:57] xubuntu, mythbuntu successfully built btw [01:58] order was ubuntu, edubuntu, kubuntu for dvd... [01:58] re xubuntu and mythbuntu - cool. [01:58] * skaet posting [01:59] right, I don't see any pending jobs for edubuntu or kubuntu DVDs, and they're definitely not there; firing them off again [02:02] slangasek, thanks. [02:03] all livecd's now posted (u, ku, xu, myth, ku-mobile) buntu [02:04] all alternates now posted (u, ku, xu, u-server, u-studio) [02:05] slangasek: I just sponsored an xserver-xorg-video-intel upload. It fixes a major corruption problem with KDE on Sandybridge video systems. Two independent testers have verified it on KDE and Sarvatt verified no regressions in Ubuntu. It would be really nice to get it in and it's a quick build too. [02:06] ScottK: I'm afraid I'm on the way out the door so I won't be able to review it for about an hour [02:06] skaet: ^^ does that sound ok to you? if you need me to review when I get back I can [02:06] slangasek: I think that's fine. You're still likely to be the first reviewer to show up. [02:07] It's a +2 lines patch, BTW. [02:07] (from upstream) [02:07] slangasek, if you could that would be appreciated. [02:15] * skaet just checked, and it looks like we're still waiting for python2.7 to build before kicking off the arm images.... [03:12] looks like edubuntu builds are failing with some sort of grub-related problem [03:13] Setting up grub-gfxpayload-lists (0.1) ... [03:13] /usr/sbin/update-grub-gfxpayload: 4: cannot create /boot/grub/gfxblacklist.txt.new: Directory nonexistent [03:14] slangasek: oh, /me hopes it's not in the LTSP chroot build part ... [03:15] http://people.canonical.com/~ubuntu-archive/livefs-build-logs/natty/edubuntu-dvd/20110412/livecd-20110412-i386.out [03:15] not sure how to tell at a glance if it's in chroot-building [03:16] as ltsp is never mentioned in the whole log, it's safe to assume that's really the livefs failing and not the LTSP chroot [03:16] LTSP chroot would have been really weird as we aren't supposed to have grub in there [03:17] * slangasek idly wonders if this intel driver fix also addresses the occasional screen corruption he's seen [03:18] stgraber: ok, well, hmm. why is this only failing on edubuntu? is grub-gfxpayload-lists edubuntu-specific? [03:18] ScottK, skaet: xserver-xorg-video-intel accepted; doesn't reset our waiting for armel thankfully :) [03:19] slangasek: did we get any successful livefs build since cjwatson added grub-gfxpayload-lists as a depend of grub-pc this morning ? [03:19] slangasek, thanks. so the armel waiting on python is still pending? ... [03:20] stgraber: we certainly had a number of image builds done and I didn't notice any mails about livefs build failures... [03:20] skaet: looks like python2.7 is done on armel and publishing this hour [03:22] slangasek, thanks. edubuntu's dvd still in progress too, from the looks of it. [03:23] skaet: no, edubuntu dvd builds failed, see above [03:24] * skaet looking more carefully... sigh. [03:24] hmm, can't think of anything weird we do with grub in Edubuntu ... [03:25] slangasek: what's the magic involved when two packages depend on each other to determine which is the first to get configured ? [03:25] slangasek: I see that grub-pc depends on grub-gfxpayload-lists which depends on grub-pc [03:26] stgraber: dpkg looks first for a package with no postinst [03:26] if both or neither have a postinst, it draws lots [03:26] * stgraber wonders if both grub-pc and grub-gfxpayload-lists have postinst ... [03:30] ok, so grub-pc has a postinst that amongst other things seem to create /boot/grub and grub-gfxpayload-lists' postinst calls /usr/sbin/update-grub-gfxpayload which writes the blacklist into /boot/grub [03:31] yeah. I'm guessing it was an oversight to create this circular dep [03:40] slangasek: while you are around, can you have a quick look at: https://bugs.launchpad.net/ubuntu/+source/pam/+bug/745532/+attachment/2021342/+files/bug745532.diff that's to fix bug 745532 (pam upgrade error when gdm isn't running) ? [03:40] stgraber: Error: Bug #745532 is private. [03:40] Launchpad bug 745532 in pam (Ubuntu Oneiric) (and 2 other projects) "fails to restart (not running) gdm on maverick->natty upgrade (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/745532 [03:40] slangasek: Thanks for taking care of xserver-xorg-video-intel . [03:41] with upstart not returning anything on status, I didn't find any better way than using grep to get the current status :( [03:43] ScottK: no problem [03:43] stgraber: I think better form would be: if $idl status | grep -q stop/waiting; then [03:44] indeed, updating with that [03:46] I'm wondering if invoke-rc.d is wrong here, however [03:47] if 'invoke-rc.d $service reload' is not meant to return an error for a non-running service, we should fix invoke-rc.d [03:48] policy is mute [03:48] so yeah, let's go with your patch, at least for now [03:49] I would probably do s/skipping/no reload needed/, personally [03:51] ok, changed to "no reload needed" and uploaded [04:31] slangasek: Looking at https://launchpad.net/ubuntu/+source/dhcp3/3.1.3-2ubuntu7 I'm thinking dhcp3 source ought to be removed. [04:32] it has been removed from Debian (pre-squeeze release). Want to file a pro forma bug? [04:32] OK. [04:34] slangasek: Bug #758357 [04:34] Launchpad bug 758357 in dhcp3 (Ubuntu) "Please remove dhcp3 source from Natty (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/758357 [04:36] slangasek: Do you think pam should go in before Beta 2? [04:37] ScottK: no strong opinion; it might be something we want in the archive at beta release for upgrades, otherwise it's probably something to accept opportunistically [04:38] OK. It looks fine to me, I guess I'll leave it for if someone needs a publisher run. [05:01] edubuntu dvd posted [05:09] I'm assuming we can remove universe packages up to final universe freeze? [05:10] minus however long it takes to get an archive admin's attention to do it, yes. [05:38] kubuntu dvd built && posted [06:19] unbuntu-netbook built and posted [08:18] Good morning [08:20] skaet: hey [08:20] how bad is it? [08:21] heya pitti, not too bad, so far. [08:21] nice, my double wait-for-package armel builds triggered fine [08:21] armel are still emerging... [08:21] I'll add them to the tracker then [08:21] they all finished building [08:22] edubuntu dvd apparently failed [08:22] I'm not seeing netboot images though. can you check that they've been started? [08:22] skaet: hm, we don't have cronjobs for any netboot "images" as such [08:23] pitti, hmm... they show up in the dailies, so something's building them. [08:24] pitti, probably best to wait until cjwatson's online and talk it through with him. [08:24] other than that, I'm keeping my fingers crossed when I wake up we don't have too many new bugs. ;) [08:24] skaet: well, d-i builds them, yes, but they aren't ISOs [08:25] upgrades aren't iso's either ;) [08:26] * skaet is fuzzy brained though at this point, so will leave you to it, and get some zzz's. [08:26] skaet: so I guess there's some magic to create "beta-2" in http://archive.ubuntu.com/ubuntu/dists/natty/main/installer-amd64/ [08:30] probably... also, edubuntu dvd was kicked off again by slangasek, and there's dvd images in the iso tester. [08:30] ah, so that probably just killed mine, good [08:30] skaet: so we have something to play with now, and I'll review the upload queue for more beta-2 RC bug fixes [08:31] so that we might have another round of CD builds tonight? [08:31] pitti, yeah, review for the opportunistic fixes, and we'll see what the testing shows us. [08:32] on the positive side - I upgraded my 10.10 clean wubi system, and it booted up nicely into 11.04 for the first time through that path, so I was smiling tonight. [08:37] all arm images posted [08:44] pitti, netboot from 11-Apr-2011 19:03 is good to add to the tracker ? [08:47] jibel: I hope so, I'll add it [08:48] pitti, nm I'm on it. [08:48] ah, thanks [09:08] ok, I reviewed/accepted everything which is either obviously safe or urgent for beta2 [09:10] the remaining uploads should wait until after b2 IMHO [10:01] Do we have a guess when the next spin is due? [10:07] Daviey: pretty much "when we need it" [10:07] Daviey: do you need any of the virtinst etc. packages that were accepted an hour ago on the server images? [10:07] pitti: Yeah, i wondered if he knew of something on the horizon that might be enough. [10:07] "he"? [10:08] Daviey: but the current server isos should be fine [10:08] for an initial test, anyway [10:08] pitti: More curious to enable hggdh to QA Eucalyptus. [10:08] It doesn't jusify a respin. [10:37] skaet: netboot images are built by uploads of the debian-installer packages. FYI [10:38] did anyone resolve the grub-gfxpayload-lists issue I see in backscroll? [10:38] are we too late in the game for beta-2 stuff? Or would any accepted uploads require a respin? [10:38] the obvious solution to that seems to be to have grub-gfxpayload-lists.postinst mkdir -p /boot/grub [10:39] I'm happy with the circular dep part of it [10:41] ev: right now we plan to do another respin this evening, and accept important or safe stuff over the day [10:41] yay [10:41] I have safe stuff ;) [10:42] but I'll hold off for a bit just in case any more ubiquity bugs get fixed [10:42] cjwatson: not that I have seen [10:43] uploaded a fix === doko_ is now known as doko [11:09] ta [11:29] pitti, i didnt get any tracker notification for headless armel builds, but they seem to exist on cdimage [11:30] ogra_: they do exist in the tracker as well [11:30] hmm, weird [11:30] * ogra_ checks, i thought i subscribed [11:33] Accepted kubuntu-meta. [11:33] And now I'm off for the day. [11:35] pitti, oops, sorry, i was actually not subscribed [11:35] ScottK: thanks [11:35] ogra_: ah, at least that explains it [12:08] I think http://people.canonical.com/~ubuntu-archive/nbs.html is about as far as we can realistically get [12:08] x264 is FTBFS on powerpc, and I already tried to build lipsia and odin against libdcmtk2-dev, they fail horribly on other things [12:09] so I'm inclined to just declare lipsia/odin broken and remove the libdcmtk1 stuff [12:09] I pinged siretart about whether he's interested in x264 on powerpc, but if not, I'd just declare it broken as well [13:37] Riddell, bug 758614 I haven't found any useful information in the logs. I can replay the installation if needed and provide additional infos. [13:37] Launchpad bug 758614 in ubiquity (Ubuntu Natty) (and 1 other project) "Kubuntu Wubi - Black screen during stage 2 (affects: 1) (heat: 6)" [High,Confirmed] https://launchpad.net/bugs/758614 [14:55] jibel: hmm, I don't even know enough about the wubi process to know what should be going on during that [14:56] Riddell: "stage 2" of wubi is just ubiquity [14:57] Riddell, I don't think it's wubi, it's more like the session not starting before launching the installer. [14:57] with a load of preseeding [15:08] morning all [15:08] happy Tuesday [15:09] :) [15:09] Good morning, skaet [15:09] Hi skaet [15:10] hey skaet [15:11] hi jibel, charlie-tca, any particularily nasty bugs showing up from the testing so far? [15:11] hi seb128 :) [15:18] skaet, nothing exciting. 2 ubiquity bugs with Wubi, translations/strings issues, and a grub failure on amd64+mac [15:19] jibel, nothing exciting == good. ;) [15:20] thanks [15:20] skaet, I've added a report on qa.u.c to tracker the state of the builds the bugs filed during the release. [15:20] http://reports.qa.ubuntu.com/reports/isotesting/natty/opened.html [15:21] * skaet hugs jibel [15:21] thanks! That should make the final cross checks and documentation easier. :) [15:21] the first page 'image list' takes the images available from cdimages and the build logs and match it with the results on the tracker [15:22] so you'll see images that are published on cdimage but not published on the tracker. [15:22] skaet: upgrade Xubuntu maverick to natty pulls in all of gnome [15:22] like natty-server-powerpc.iso for example [15:22] gives me the option to use any ubuntu or xubuntu session at login [15:23] charlie-tca, that's a dependency issue that mrpouit needs to fix. [15:23] I know, but that doesn't make it less nasty [15:38] jibel: where do I download wubi from? [15:39] Riddell, it's on the iso [15:39] mount the iso and wubi.exe is at the root of the iso [15:39] so I have to burn the ISO, boot up windows and run wubi from there? [15:40] Riddell, to go faster mount the iso, and copy wubi.exe and the iso to windows [15:40] ok [15:40] Riddell, then run wubi.exe --isopath=natty-desktop-amd64.iso or the arch you're running. [15:41] or burn a cd [15:41] jibel: I hope you *occasionally* test from a burned CD; the code paths aren't the same when you run from an ISO image on disk [15:42] could somebody review console-setup? in order for that fix to be effective, it will need to land in the archive so that we can incorporate it into the next ubiquity upload [15:43] cjwatson, yes, but once I've validated that the CD is ok I use the isopath method, which is faster than burning a cd for each test. [15:44] ok [16:20] cjwatson: doing now (sorry, had a two-hour mumble debug session) [16:20] cjwatson: accepted [16:21] no problem, thanks [16:21] can someone let migration-assistant through? I'd like to get that into the ubiquity upload as well [16:21] ev: yep, I'm running through the queue now [16:21] yay [16:21] thanks [16:23] jibel, skaet, cjwatson: I think tonight we should regenerate all images to pick up today's bug fixes; do you have a preferred time when to start this? (I thought around 2000 UTC) [16:23] GrueMaster, ogra_: WDYT about the current armel images? do you want another build tonight? [16:23] pitti: what bug fixes affect armel? [16:24] the python-apt one, according to ogra [16:24] GrueMaster: ^- [16:24] GrueMaster: not the gfxboot ones, and not the installer ones; there were some fixes in the desktop, but none which break the installation [16:25] python-apt is bug 758732 [16:25] Launchpad bug 758732 in python-apt (Ubuntu Natty) (and 1 other project) "enable_component() does not enable components for deb-src lines (affects: 1) (heat: 8)" [Medium,Fix released] https://launchpad.net/bugs/758732 [16:25] Well, let me get some testing in so we can catch any criticals. I haven't had fresh working images in a week. [16:25] that bug is very low. [16:26] so if the prebuilt images have a wrong apt sources.list, you might want that [16:26] Not sure why it is marked as medium. [16:26] GrueMaster: yeah, I'm asking as we need to weigh the nontrivial hours of building and mandays of testing against the fixes [16:26] Having no deb-src for universe is not enough of a reason to regenerate. [16:27] Right now I am seeing a bug where oem-config fails to start on the headless images. Higher priority. [16:28] hm, haven't see the report for that one yet [16:28] I should be able to do enough testing between now and 2000UTC to determine what else needs critical fixes before we respin armel. [16:28] Just filed. [16:28] right [16:28] bug 758858 [16:28] Launchpad bug 758858 in ubiquity (Ubuntu) "oem-config fails to start on headless images (affects: 1) (heat: 8)" [High,New] https://launchpad.net/bugs/758858 [16:28] GrueMaster: so we'll rebuild after these get fixes, not time-based [16:29] That would be much better for me. [16:29] 2000 for desktop/alternates/DVDs just seems like a good time to pick up ev's/cjwatson's/mvo's fixed from the day [16:30] As long as I have new images by this time tomorrow, I can get them tested. I'm in US-PST. [16:31] GrueMaster: 758858> can you try booting the first time round with 'debug-oem-config' on the kernel command line, instead of running oem-config-firstboot a second time? [16:31] GrueMaster: the log is full of stuff that indicates that the debconf database is already locked by something else, and I don't know if that's related to the way you did your debug run or not [16:32] I tried that yesterday with the same results. Unfortunately I had to get something else tested and didn't get a chance to file a bug. [16:32] I think we'll get a more useful log if you use debug-oem-config to put it in debug mode right from the start [16:33] GrueMaster: sure, it will still *fail*, but it will produce a more useful log [16:33] I'll try again on a clean image. [16:33] the current logs don't have anything actionable [16:38] ev: ^ [16:38] much appreciated [16:42] cjwatson: I added debug-oem-config and the log doesn't appear any different from launching it manually (other than I have no console using this method). [16:43] Is there a different log to look at? [16:43] /var/log/oem-config.log should be the right one [16:43] can I see it? [16:44] (also, to be perfectly honest, I need ARM people to be debugging installer problems on ARM, as a general rule ...) [16:44] http://paste.ubuntu.com/593172/ [16:44] I can try but success is far from guaranteed [16:44] hmmm [16:45] odd multiplicity of log entries there, but as you say, no different [16:45] I guess I need a way to reproduce this on i386, or else an ARM person needs to debug this [16:45] Looks like each run just appends new output. There should be 3 runs in that log. [16:46] I've pinged ogra. [16:55] I'm off for about 1 h for my Ubuntu Dev Week talk about PyGI [17:01] pitti, i discovered a jasper bug with the netbook image so i would like a rebuild once i uploaded the fix [17:02] ogra_: rebuild is planned,but let's get as many bugs as possible before you trigger the rebuild. [17:02] GrueMaster, indeed [17:04] cjwatson, on the headless image i'm calling a chrooted debconf-communicate from initramfs to preseed the debian-installer/framebuffer value, could it be that i miss some black magic to have it remove the lock ? [17:04] (thats the only idea i have for the above) [17:04] simply exiting should be enough [17:04] hmm, k [17:05] there is nothing else that could touch debconf, weird [17:05] getting an strace might be worthwhile, in case it's oem-config trying to recursively claim the lock itself [17:05] (strace -f -s 1024, at least) [17:05] and also of course check for other running processes [17:16] Is adding a new meta package to a (universe) source package OK at this stage in the cycle? [17:40] skaet: doko: ogra_: pitti: so, do I reject bug 67544, or just mark it incomplete? (we're going to actually need some binaries to bootstrap from that run on ubuntu...) [17:40] Launchpad bug 67544 in fpc (Debian) (and 5 other projects) "Bootstrapping needed for fpc for armel (heat: 13)" [Unknown,Fix released] https://launchpad.net/bugs/67544 [17:40] pitti, before starting the images, have you done "Modify debian-cd/CONF.sh to set OFFICIAL " [17:41] lamont, if you're needing binaries to bootstrap (effectively waiting for input from reporter sort of thing) then incomplete seems like the right state [17:43] can someone let jasper-initramfs in ? diff is at http://paste.ubuntu.com/593206/ i need it on the next netbook image [17:45] skaet: that was already done for beta 1, and wasn't undone [17:46] (I checked yesterday, remembering last time ...) [17:47] cjwatson, :) [17:59] * highvoltage starts upgrade testing for Edubuntu Desktop i386/amd64 [18:18] we're in hard beta freeze now? [18:21] jbicha, yes [18:22] jbicha, is there a specific fix you're trying to get in? [18:22] yes, fixing bug 426215 would make the Ubuntu Help better as we could link directly to Software Center [18:22] Launchpad bug 426215 in software-center (Ubuntu) (and 1 other project) "[UIF exception] apt:package-name isn't handled by the Store when appropriate (affects: 5) (heat: 30)" [Wishlist,Fix released] https://launchpad.net/bugs/426215 [18:27] I can ask again Friday though, right? [18:28] doing the ubiquity release dance now [18:37] ^ if someone could let that through, I'd appreciate it [18:40] jbicha, yes, please ask as soon as we get the beta images published. [18:47] virt-manager's scale option ftw! http://people.ubuntu.com/~jonathan/files/natty/testing.png [18:53] * skaet goes to get some lunch, biab [18:54] cjwatson, strace data attached to bug 758858 ... i see a lot of EBADF errors [18:54] Launchpad bug 758858 in ubiquity (Ubuntu) "oem-config fails to start on headless images (affects: 1) (heat: 8)" [High,New] https://launchpad.net/bugs/758858 [19:18] ev: looking [19:20] ogra: jasper-initramfs accepted; so you want all armel rebuilds for that (also kubuntu)? (<- GrueMaster) [19:21] pitti, for that one i dont need kubuntu, but for the python-apt one [19:21] ogra_: so, full rebuild tonight? [19:21] the jasper-initramfs change is only for omap4 netbook [19:22] yeah, we are still trying to find the cause for bug 758858 though [19:22] Launchpad bug 758858 in ubiquity (Ubuntu) "oem-config fails to start on headless images (affects: 1) (heat: 8)" [High,New] https://launchpad.net/bugs/758858 [19:22] ogra_: right [19:22] ogra_: I'll monitor that then and do rebuilds once that's fixed [19:23] ogra_: if it doesn't get fixed by, say, 2200 UTC, do you still want rebuilds for the other fixes? [19:23] pitti, yeah, theoretically the non headless rebuilds could happen now, lets see what GrueMaster thinks about that [19:24] ogra_: we still need the jasper-initramfs build/publication, no? [19:24] and headless is so fast to test that we could still roll it tomorrow worst case [19:24] *nod* [19:24] yeah, it needs to be on the images [19:24] anyway, I need some food, bbl [19:24] I haven't hit any criticals on the non-headless images yet. Haven't looked at kubuntu though. [19:24] GrueMaster, well, the jasper fix is for unity-2d only (the PPA icon) [19:25] theoretically we wouldnt need to rebuild kubuntu (unless there are heavy bugs you will find) [19:25] understood. Just specifying what *needs* rerolling immediately vs later tonight. [19:26] the plan earlier was to reroll all armel to pick up other bug fixes. I want to test (and hopefully fix) as much as possible before that happens. [19:27] right [20:14] ^^ ecj is a fix for an armel ftbfs, and for ecj1 being unusable due to multiarch; it's a dep of the gcj jdk and of eucalyptus so it impacts dvds and server images [20:14] (but is not beta-critical) [20:15] cjwatson, backing out your fix for bug #746020 makes headless work [20:15] Launchpad bug 746020 in ubiquity (Ubuntu) "ubiquity crashed with ValueError in command(): invalid literal for int() with base 10: '' (affects: 3) (dups: 2) (heat: 28)" [Low,Fix released] https://launchpad.net/bugs/746020 [20:24] skaet: how critical is LTSP for beta2 ? [20:25] stgraber, can you give me the bug number? [20:25] skaet: there's a bug that seems to affect ubuntu alternate 32bit/64bit + edubuntu 32bit/64bit. Not sure where it comes from yet, debugging it now. [20:26] skaet: don't have one yet [20:26] stgraber, how does it manifest? [20:26] patrickmw: can you file one anyway ? [20:26] skaet: install LTSP from alternate or test it in Edubuntu => won't boot [20:26] as in, the thin client won't boot. The server is fine AFAIK [20:27] stgraber, yes. did you determine the best package to file the bug against? [20:27] patrickmw: just file it against ubuntu for now and assign to me please [20:27] stgraber, interesting. Let me have a couple of discussions. Sounds like something we can document around if necessary, but I'll see if I hear a dissenting opinion. [20:27] stgraber, ok [20:28] stgraber, thanks for flagging. [20:28] skaet: I should have the issue reproduced here in 10-15 minutes, then it shouldn't be long to figure out what broke and give you more information. [20:29] stgraber, cool, thanks! [20:30] skaet, root cause for bug 758858 found, but i dont really know how to solve it, seems its caused by a fix for another bug [20:30] Launchpad bug 758858 in ubiquity (Ubuntu Natty) (and 1 other project) "oem-config fails to start on headless images (affects: 1) (heat: 8)" [High,Confirmed] https://launchpad.net/bugs/758858 [20:31] i just did a succsessfull headless install here [20:34] patrickmw: reproduced the bug [20:36] stgraber, you will have a bug id in a few [20:36] and identified the problem [20:37] it's some kernel packaging (I guess) that makes the files 600 instead of 644 [20:37] so now the tftp process can't read the kernel and initrd [20:37] skaet, highvoltage: ^ [20:38] stgraber, ack [20:39] ogra_, glad you know root cause, guess we'll go ahead with builds, and if you can figure a workaround/resolution by tomorrow morning your time, work with pitti to incorporate it. Doesn't look likely we'll get a fix tonight then. [20:39] stgraber, bug 759115, I assume you will be changed the package its assigned to [20:39] Launchpad bug 759115 in ubuntu-meta (Ubuntu) "ltsp thin client kernel error - Could not find kernel image: vmlinuz (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/759115 [20:40] skaet, well, cjwatson wrote the original fix, i would like to get some input from him [20:40] ogra_, understandable. [20:41] skaet, but a full install test of headless takes around 10min (plus another 10 for dd'ing the image to SD) so we can be late with that [20:41] skaet: not totally sure how to fix the ltsp issue ... I could do a hack in LTSP to chmod the files when copying the kernels, but that definitely changed since beta1, so I'd be interested to know why the kernel is now only readable by root [20:42] stgraber, I was about to suggest going onto u-kernel and asking, but I see that bjf has just joined here, so wondering if word has spread ;) [20:42] stgraber: I guess there would be a feature freeze bug for it if it was done on purpose? [20:44] bjf: hey, any idea why kernel and initrd is now 600 instead of 644 ? that's breaking ltsp :) [20:45] stgraber, [ Kees Cook ] [20:45] * [Config] packaging: adjust perms on vmlinuz as well [20:45] from 2.6.38-8.40 [20:45] doh, security ... should have thought of it [20:45] talk to kees [20:45] :) [20:46] eek, I seemed to have forgotten about https://bugs.launchpad.net/ubuntu/+source/edubuntu-artwork/+bug/746028 since the last beta [20:46] Launchpad bug 746028 in edubuntu-artwork (Ubuntu) "Edubuntu Wallpapers are not updated on upgrade to Natty (affects: 1) (heat: 493)" [Wishlist,Confirmed] [20:47] * skaet had noticed that last night, when scrubbing techoverview and was a bit surprised it was still there.... [20:49] skaet: could I still fix that after the beta release is done? [20:50] highvoltage, depends when after. ;) Its localized scope though, so bug fix on friday would be fine. [20:50] didnt you write 21st would be upload deadline ? [20:50] :) [20:51] skaet: ok, I'll prepare and test a fix in the meantime, so if there's a window for uploading, I'll be ready for it [20:52] highvoltage: upload when ready; in the worst case, it'll be in teh archive for upgrading, and in the best case we do a respin and it's in [20:52] I don't see a problem with accepting it [20:52] pitti: thanks! [20:57] pitti: I'll be uploading a new upstream version of LTSP soon to fix the LTSP bug affecting alternate. It's going to be including my fix + a ntpdate race condition fix + translation update + gentoo specifc delta. I'll add that to bug 759115 [20:57] Launchpad bug 759115 in ubuntu-meta (Ubuntu) "ltsp thin client kernel error - Could not find kernel image: vmlinuz (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/759115 [20:58] ogra_: hmm [20:58] :( [20:58] ogra_: I wonder if I put that code in the wrong place [20:58] stgraber: ah, good; so we'll wait for that with the respin of alternates in any case [20:59] cjwatson, we also verified that it worked before the code was added, the 0402 image works, from 03 on it fails ... i find it a bit odd that only the debconf frontend seems to cause it [20:59] ogra_: I have a good guess, just reading code [21:00] k [21:00] yeah, I can fix this [21:00] \o/ [21:00] thanks for that investigation, that's exactly what I needed [21:00] :) [21:01] anything else in ubiquity known to need urgent fixing? [21:02] well, ev just did an upload, i guess he already covered most [21:03] stgraber: http://paste.ubuntu.com/593284/ [21:03] (#ubuntu-devel from five days ago) [21:05] cjwatson: yeah, I don't really like the idea of changing permissions on the kernel less than a month before release ... [21:17] pitti: ok, ltsp is on its way (5.2.7). Let me know if you have any question regarding the changes (and I'll either try to answer or nag the one who commited it :)) [21:21] ogra_: the big pile of EBADF is actually harmless - it's just that Unix has no way to say "close all fds" or "close all fds apart from these ones" [21:21] it's an omission from the interface [21:21] ah [21:22] the best you can do if you want that is to iterate up to some plausible maximum [21:22] jibel, skaet: I see you tested ubuntu upgrades. could you confirm that the ubuntu wallpapers changed to the new ones after upgrade? I suspect that the edubuntu wallpaper bug might be present in ubuntu as well and since the wallpapers are quite similar, it might have gone unnoticed [21:22] well, that strace was truncated anyway [21:22] the second one actually showed some more info [21:23]