[09:36] <psivaa> cjwatson: xnox: http://pastebin.ubuntu.com/5721076/ is seen with the default raring server installations today.
[09:37] <xnox> psivaa: is that on all server jobs?!
[09:37]  * xnox ponders how to subscribe to failures on those.
[09:38] <psivaa> xnox: both i386 and amd64
[09:40] <cjwatson> hmmmm
[09:40] <psivaa> xnox: public jenkins is taking too long to load, but we have not set up email notifications for the smoke tests yet, we could do that
[09:41] <cjwatson> psivaa: new today?
[09:41] <psivaa> cjwatson: yes this failure is new and only with today's images
[09:41] <cjwatson> oh, d-i updated without seeds again
[09:42] <infinity> It sure didn't.
[09:42] <cjwatson> hmm
[09:42] <cjwatson> maybe just really unlucky timing
[09:42] <infinity> Unless something's trying to pull d-i from proposed.
[09:42]  * cjwatson grabs the image to look in more detail
[09:42] <infinity> But I did the seed change at more or less the optimal time (sure, there's a window for a publisher cycle there, or so)
[09:43] <cjwatson> the build is around the time when you *might* have got unlucky, but I'll check
[09:43] <psivaa> there seems to be a new kernel yesterday 3.8.0-19
[09:43] <cjwatson> yeah, I know
[09:43] <cjwatson> I'll look into it
[09:44] <psivaa> cjwatson: ok thank you
[09:44] <cjwatson> psivaa: can I have a link to the full log?
[09:45] <psivaa> cjwatson: i'm trying to load that in the public instance which has not still loaded, ill copy and paste the full log from the private jenkins
[09:45] <infinity> Maybe if we mangle when cdimage dailies build, we can just disallow migration of d-i/linux during suboptimal windows. :P
[09:45] <cjwatson> I have lab access, so a private jenkins link would be fine
[09:46] <infinity> Or maybe my hackish solution to get kernel ABIs completely out of the seeds would solve all of this.
[09:46] <psivaa> cjwatson: http://pastebin.ubuntu.com/5721108/
[09:46] <infinity> I'll have to look at landing that early in S.
[09:47] <cjwatson> hmm, that's d-i with -18 and -18 is on the images.
[09:47] <cjwatson> Apr 19 08:19:23 main-menu[1734]: (process:13782): libkmod: kmod_lookup_alias_from_builtin_file:
[09:47] <cjwatson> Apr 19 08:19:23 main-menu[1734]: (process:13782): could not open builtin file '/lib/modules/3.8.0-18-generic/modules.builtin.bin'
[09:47] <cjwatson> is that a recent kmod change?
[09:48] <infinity> I did upload a new kmod, but that seems to be a red herring, no?
[09:48]  * xnox thought i was fixing builtin file opening.... and it was mostly harmless before?!
[09:48] <infinity> Apr 19 08:19:23 main-menu[1734]: (process:13782): unknown udeb squashfs-modules
[09:48] <infinity> Apr 19 08:19:23 main-menu[1734]: (process:13782): FATAL: Module squashfs not found.
[09:48] <infinity> Those seem to be the more interesting lines.
[09:49] <cjwatson> ah, yeah, that's package-level of course
[09:49] <cjwatson> ok, I'll have to actually try the image and see what's going on there
[09:50] <infinity> Could this just be that the test was run after I NBSed out squashfs-modules-18?
[09:50] <infinity> Though, I assume that should be on the CD, so nevermind.
[09:50] <cjwatson> it may well have been, because squashfs-modules sure isn't on the image
[09:51] <cjwatson> this is odd because various other -modules are
[09:51] <cjwatson> (which means that the sanity check in debian-cd didn't fire)
[09:52] <infinity> Seems like a bunch of them aren't on the image.
[09:52] <cjwatson> aha, and the image build postdated the seed change
[09:52] <cjwatson> ! Allowing d-i kernel versions: ['3.8.0-19-generic']
[09:52] <ubot2> cjwatson: I am only a bot, please don't think I'm intelligent :)
[09:52] <cjwatson> ! Pruned squashfs-modules-3.8.0-18-generic-di from installer
[09:52] <ubot2> cjwatson: I am only a bot, please don't think I'm intelligent :)
[09:52] <cjwatson> oh shut up ubot2
[09:52] <infinity> So, just poor timing.
[09:53] <infinity> Mystery solved.
[09:53] <cjwatson> but then things like
[09:53] <cjwatson> * Chose crypto-modules-3.8.0-18-generic-di out of crypto-modules to satisfy netcfg
[09:53] <cjwatson> and indeed this is probably why:
[09:53] <cjwatson> [09:53] <cjwatson> Fri Apr 19 07:07:54 UTC 2013
[09:53] <infinity> Right, so all the modules that made it are there because of transient deps, not direct seeds.
[09:53] <cjwatson> so it was building alongside something else and hence the mirror was a bit staler than it should have been
[09:54] <cjwatson> cdimage's parallel build support isn't perhaps as absolutely perfect as it should be for this (though it's not clear that it can be)
[09:54] <cjwatson> So the fix is just to respin the ISO part.  I'll do that now
[09:54] <infinity> Is there no kernel in the squashfs for server?
[09:55] <infinity> Ahh, guess not.
[09:55] <cjwatson> nope
[11:05] <cjwatson> psivaa: should be better now?
[11:09] <psivaa> cjwatson: our cron jobs have not yet picked up the new set. Will let you know once they run.
[11:10] <psivaa> cjwatson: thanks for the quick fix though :)
[11:12] <psivaa> cjwatson: The desktop media-info files still say Alpha. I know that does not have any user impact, but is there a plan to change it? Or Is it that todays desktop images wont be the RC?
[11:13] <cjwatson> psivaa: I already changed that earlier today in the code
[11:13] <cjwatson> those desktop images can't be RC
[11:13] <psivaa> cjwatson: ack
[13:45] <xnox> ev: well does it matter who files the RT for signing wubi? =) my RTs usually get stuck forever.... Yours are getting done quicker ;-)
[13:46] <ev> :) you just need to make friends with people in IS/webops
[13:46] <ev> on it
[13:47] <xnox> ev: thanks.
[13:47]  * xnox just makes friends with you ;-)
[13:48] <ev> xnox: ha! :) I don't think the transitive property applies to RT
[13:48] <ev> oh wait, it just did
[13:48]  * xnox hides
[13:51] <ev> RT 60951
[13:56] <xnox> ev: No permission to view ticket  =(
[13:56] <ev> strange
[13:56] <xnox> maybe after it's trianged in correct queues?!
[13:56] <ev> yeah, probs
[13:57] <xnox> anyway added to my short list of RTs I am waiting on.
[14:02] <ev> xnox: if you're blocked on RTs, Steve can raise the priority of them and mention them in his catch up call with IS
[14:02] <cjwatson> yeah, doit, he wants to have something to talk about there :)
[14:02] <ev> xnox: it also helps to know whats in front of you: https://portal.admin.canonical.com/ruins?team=losa
[14:03] <xnox> ev: well, _I_ am not blocked on them. They are just fixes that affect the users of our services: manpages.ubuntu.com, geoip-lookup.ubuntu.com, wubi release.
[14:04] <ev> geoip-lookup is relatively important, I would think
[14:04] <ev> definitely worth mentioning to Steve
[14:05] <xnox> well geoip is #15 with size estimate XS =) (i hope that's ExtraSmall and not ExtrodinarySophisticated)
[14:06] <xnox> and manpages got reopened and are not up on the board yet.
[14:07] <xnox> any opinions on bug 1170150
[14:07] <ubot2> Launchpad bug 1170150 in grub2 (Ubuntu) "vmlinuz/initrd.img symlinks do not point to signed versions on kernel updates of secure boot UEFI machines" [Medium,Confirmed] https://launchpad.net/bugs/1170150
[14:07] <xnox> ?
[14:11] <cjwatson> how is that a grub2 bug?
[14:11] <cjwatson> it doesn't manage the symlinks
[14:11] <xnox> cjwatson: what does manage symlinks? linux-kernel?
[14:11]  * cjwatson reassigns to linux
[14:12] <cjwatson> -> not our problem :)
[14:12] <xnox> love it =)
[14:12] <cjwatson> (also not a release-critical problem or anything since the symlinks are only ever used for the sort of manual use case the reporter suggests)
[14:19] <xnox> ev: when you will be state side would be able to poke Keybuk about doing a libnih release? or e.g. handing the project over to ~upstart-devel team? https://code.launchpad.net/libnih/+activereviews shows the amount of fixes that are in ubuntu but not upstream, and there hasn't been a bugfix release with those.
[14:19] <xnox> we have a few people on RPM-based distros asking for a tarball release with all of those fixes.
[14:20] <xnox> ev: or is it something you don't talk with Keybuk about? =))))
[14:21] <ev> texted him what you said
[14:23] <xnox> ev: thanks.
[14:24] <ev> fingers crossed that it's just something he can quickly cover off on the google bus
[14:24] <xnox> haha.
[14:48] <mpt> xnox, what do you think of the jumpy "Keyboard layout" step?
[14:49] <mpt> http://goo.gl/YwIcT
[14:51] <xnox> mpt: I am currently staring at the spinning logo around the image.
[14:51] <xnox> =)))) one moment please.
[14:55] <xnox> mpt: description sounds ok. But how does one /change/ layouts any any step? (or are we simply not support that like we currently do). E.g. what if I want English(US) everywhere but at the Name field I got a feeling to type "Дмитрий Ледков"? aka setting up a secondary layout?
[14:55] <xnox> or should we always setup english as a secondary layout and say " you can change layout above, or by pressing Control-x Alt-c Alt-butterfly
[14:56] <cjwatson> If you set up a non-Latin layout in console-setup then it is always true that you can press Alt+Shift to toggle to English
[14:56] <mpt> xnox, same as normal, using the text entry menu in the menu bar
[14:56] <cjwatson> (IIRC)
[14:57] <xnox> mpt: ack.
[14:57] <mpt> Which ... actually isn't mentioned at the moment
[14:57] <xnox> cjwatson: I had no idea, and I don't think we tell users that at the keyboard step do we?
[14:57] <cjwatson> I don't think we do
[14:57] <mpt> fixed
[14:57] <cjwatson> But we might as well tell people rather than inventing another set of keys :)
[14:58] <cjwatson> I think we may have told people at one point and it got lost in a redesign
[14:58] <xnox> mpt: is your google doc meant to have a spinner animation as the first figure under Keyboard layout? cause that's what I see here.
[14:58] <mpt> xnox, that happened when I moved it from one part of the document to another. I'm trying to pluck the image out of an older revision.
[14:58]  * xnox thinks "interation" is a better word than "redesign" as really everything is constantly "redesigned" =)))))
[14:59] <xnox> s/interation/iteration/
[15:00] <mpt> Fixed!
[15:00] <cjwatson> Well, I was specifically thinking about before the maverick-redesign branch, but I may well be mistaken
[15:01] <mpt> I've been doing more or less of the design since Ubiquity started, and this is the first I've heard of Alt Shift :-)
[15:03] <xnox> cjwatson: Alt Shift launches Unity HUD for me to search/execute global-menu items
[15:03] <xnox> or if done too quick, does nothing.
[15:05] <xnox> right - I reset to default readd Russian layout and the "key combination to change layouts" has nothing selected.
[15:06]  * xnox goes to file a bug.
[15:07] <cjwatson> Bah, I guess the Unity folks didn't know about it either ...
[15:08] <xnox> cjwatson: askubuntu brings up 2 year old questions with total of 20 000 views on approx. "why alt+shift doesn't change keyboard layout"
[15:09] <xnox> enabling alt-shift does work well enough with HUD/Dash/Apps
[15:09] <xnox> so it's just that shortcut is not enabled by default at all.
[15:12] <cjwatson> It's only enabled by default if you installed with a non-Latin layout
[15:12] <cjwatson> Or used to be anyway ...
[15:20] <xnox> Русский is non-latin =) i will do a test install at the weekend to test this.
[15:21] <cjwatson> Should be, yeah ...
[15:51] <ev> xnox: jacekn is signing wubi now
[16:29] <ev> xnox: signed wubi is in place
[16:41] <xnox> ev: awesome, what's next to "release it" ? I guess next raring image build will simply pick it up?
[16:41] <xnox> and precise daily as well?
[16:41] <ev> correct
[16:41] <xnox> what about just wubi.exe downloads from ubuntu.com?
[16:42] <xnox> would those just work after the image build?
[16:42] <cjwatson> The web team usually make that be a redirect to the one we publish on releases.ubuntu.com
[16:42] <cjwatson> Or a mirror of it
[16:43] <xnox> cjwatson: yeahp, link from ubuntu.com for precise wubi goes to: http://www.mirrorservice.org/sites/releases.ubuntu.com//precise/wubi.exe
[16:43] <xnox> cjwatson: so would you just republish precise wubi there? or do you want me to test the know signed wubi.exe first?
[16:44] <cjwatson> Usually done at the next point release
[16:44] <cjwatson> But if it's a fix for 12.04.2 then we could republish
[16:44] <cjwatson> Rather have it tested first though
[16:45] <xnox> cjwatson: it's a fix for 12.04.2, so we didn't release any wubi with 12.04.2 but we should have, due to kernel name change to -> *.efi
[16:45] <xnox> cjwatson: ok will test as per precise-daily iso tracker wubi test cases & will report results there, such that it's documented.
[16:48] <cjwatson> ah yes, that
[21:32] <mattrae_> hi, during installation it says it will take 53 min to download language packs. i see they are downloading at ~30KB/s not so fast. ive noticed this a couple times recently
[21:35] <mattrae_> this is for 12.04.2 desktop
[21:36] <mattrae_> wondering if anybody else has noticed that downloading the language packs is slow
[22:29] <cjwatson> mattrae_: not something we can really do much about here, I'm afraid ...
[22:29] <cjwatson> I guess the mirror for your country is suboptimal for you or something