[01:08] kees: Or just apt-get install gcc-multilib. [01:08] kees: Which sets up the symlinks to make it all work for -m32 === hughhalf_ is now known as hughhalf === elmo_ is now known as elmo [13:35] apw: currently I am at https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/1371651/comments/58 [13:35] Ubuntu bug 1371651 in plymouth (Ubuntu) "Daily does not boot into graphical interface after installation" [Critical,Fix released] [13:36] elfy, ok then this is something new, as the segv was the clear trigger for the problem i fixed [13:36] ok [13:38] can you file a new bug against, erm lightdm to start with and let us get the bits for your issue together [13:38] apw was just going to ask that :) [13:39] apw: do you want me to do the bug before I've started lightdm ? [13:41] elfy, sounds best yes [13:43] elfy, also when there what does "cat /proc/fb" say [13:44] once booted or before? [13:44] either i think [13:45] ok - well once lightdm starts 0 VESA VGA [13:45] I'll check the other in a minute [13:48] apw: bug 1375805 [13:48] bug 1375805 in lightdm (Ubuntu) "Lightdm fails to start in VirtualBox " [Undecided,New] https://launchpad.net/bugs/1375805 [13:52] elfy, ok can you grab a dmesg from the machine before lightdm starts onto the bug as well [13:52] do my best [13:55] jodh, is there a way to say "status all" to get all jobs ? [13:55] apw: done [13:56] elfy, and an "initctl list" [13:57] yep - anything else while I'm there? [13:57] elfy, not sure, /me is flying blind currently [13:57] heh [13:58] elfy, i wonder if this reproduces with --verbose on the end of the kernel command line [14:00] apw: initctl added - I'll try with --verbose [14:02] apw: just goes straight to tty1 without even trying tty7 [14:04] elfy, i hate debugging upstart interactions [14:04] :) [14:06] apw: ok - so now when I start lightdm (with --verbose and it not even trying tty7) it starts into a sensible sized vbox instance instead of 64x480 [14:06] elfy, erm [14:06] probably no help - but I'm trying to just give you what I see here [14:13] elfy, ok so i guess i need the syslog bits for init: at least from the --verbose boot you did, so i can compare them to mine; which obviously works [14:13] so verbose and do a new syslog? [14:16] yeah the --verbose bits get dumped into syslog, well they did on mine [14:17] apw: added [14:17] elfy, what video options did you select for this VM btw ? [14:18] I don't tend to fiddle much - 3d enabled and 128Mb [14:19] ok so you are different to me, i'll try those [14:19] apw: ok - what do you have - I can try with yours [14:20] whatever the defaults are, something like 12M and none of the tickies on [14:20] ok [14:20] yeah 12m [14:21] nope that works for me as well (128m and 3d( [14:21] elfy, what release is your host [14:22] utopic [14:22] this is a trusty host [14:22] ok - I'll reboot into trusty and see what occurs there - biab [14:23] give me a while to install the vm - biab [14:23] and whatever updates I've got pending - don't boot it often :) [14:51] ** [14:51] ** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [14:51] ** [14:53] apw: ok - so in trusty with vbox v 4.3.16-95972 I get exactly the same fail [14:53] elfy, ack thanks for testing that, i am sure it was a pita [14:53] fairly :) [14:54] not sure it puts us anywhere further forward though [15:17] elfy, ok i think i know approximatly what is happening; you are not getting a graphics card notification [15:17] elfy, which for the moment i am going to put down to a race [15:17] elfy, which means you should hit the fallback, and it seems that that is busted in utopic [15:17] elfy, someone has tried to fix it, and failed badly [15:18] * apw will poke it and get back to you [15:19] elfy, actually could you take the contents of http://paste.ubuntu.com/8466613/ as /etc/init/udev-fallback-graphics.conf [15:19] elfy, and let me know if that at least gets you lightdm [15:32] Hello! [15:32] I'm seeing a major problem on Ubuntu 12.04 with trusty-lts: https://i.imgur.com/EXQm49N.png [15:33] 1 GB of RAM is used by upstart-udev-bridge and 240 MB of RAM is used by "init" [15:33] I was able to run into this problem by running ~200,000 Docker containers. [15:38] jodh, ^^ that sounds like a memory lieak in upstart-udev-bridge or something [15:39] apw: I succeeded in killing the existing vm - reinstalling it now - will do that ^^ asap [15:39] backjlack, is that in one instance of upstart-udev-bridge ? [15:39] apw: yes [15:40] apw: headless ubuntu server with trusty lts kernel, fully updated [15:41] backjlack, ok that does sound like a memory leak or similar in upstart-udev-bridge ... could you file a bug on that [15:41] apw: Ok, what should I file the bug against? [15:42] backjlack, ubuntu-bug upstart [15:43] backjlack: as apw says, please raise a bug so we can investigate. [15:43] "skipabi=tue" well that ain't gonna fly [15:44] hallyn, :) [15:48] apw: posted confirmation to bug - that .conf works here [15:49] hallyn, try skipabi=weds [15:49] elfy, thanks ... thats at least an easy fix [15:49] rtg, :) [15:49] :) [15:51] elfy, now to go and beat on the person who removed it :) [15:51] heh [15:52] I shall toddle of into other less uncharted channels for me now - I'm about if needed :) [15:52] apw: thanks for your help [15:52] elfy, heh yeah thanks to you for your quick testing, that got us there [15:52] welcome as always [15:52] cheers :) [15:54] rtg: why not just shoot for the moon, fri no doubt will have the happiest results. unless we're in one of those countries where kids get a hlaf day every wed [15:57] apw: https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/1375874 [15:57] Ubuntu bug 1375874 in upstart (Ubuntu) "upstart leaks memory in init and upstart-udev-bridge" [Undecided,New] [15:57] apw: You also had https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/1198180, but that was ignored. [15:57] jodh, ^ [15:57] Ubuntu bug 1198180 in upstart (Ubuntu) "possible leak in upstart 1.5" [Medium,Confirmed] [15:59] Should I also investigate 14.04 to make sure it doesn't have the exact same problem? [16:00] backjlack, that owuld be interesting to know for sure [16:04] Ok, that should take only a few minutes to confirm. [16:23] smb, shortly there should be new udev binaries for this fallback thing in https://launchpad.net/~apw/+archive/ubuntu/lp1375805 ... i'll be interested to know if they fix your binary-drivers not giving you lightdm issue too [16:23] apw, okidok [16:36] smb, of cour the build estimates are just static in time [16:37] It will be build by 6pm (though we do not say which day)? [16:38] apw, can you comment on this ? (should be simple) https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1121699/comments/9 [16:38] Ubuntu bug 1121699 in linux (Ubuntu Raring) "aufs CONFIG_AUFS_EXPORT build option should be enabled" [Medium,Fix released] [16:40] rtg, ack, and i still have that perl thing to look at now i have "fixed" this vbox "kernel" issue [16:41] apw, the hyper-v dudes are complaining about crash dump still, but it appears to be a matter of a larger reservation. [16:42] smb, ^ didn't we talk about using a smaller initrd for those? ==neeed or something ? [16:42] smb, as i suspect we are holding like half the ram in these instances [16:42] apw, Though the size was also just made bigger... [16:43] smb, "just" == ? [16:44] apw, would have to install it again.... though maybe arges remembers it from his head (having to use it) [16:44] smb, ok trying to work out if the hyper-v folk tested before or after "just" [16:45] apw, iirc it would be = something (something maybe 256M [16:45] yeah i assumed they changed it from 128 to 256 or somethign [16:45] but more it is the when of it ... [16:45] though how big is am m1.small, that isn't going to leave anything for the actual instance :) [16:46] Somehwen between P and T but I know that is vaque [16:46] ## [16:46] ## Ubuntu Kernel Team Meeting - in 15 minutes - #ubuntu-meeting [16:46] ## [16:46] smalles you get would be micros (~ 600M). I think m1.small is 3.5G already [16:47] apw, Hrm... and m1 seems to be "old fashion" anyway http://aws.amazon.com/de/ec2/instance-types/ [16:48] ok... m1.small down there , so 1.7 G [16:50] smb, that feels like they have gotten a whole heap bigger than i remember [16:51] apw, Yeah. m1.small is now considered something similar than micro... at least one could think that [16:51] Is there any way at all to get a libc with dbgsym? [16:55] smb: what happened? [16:56] arges, Nothing immediate. Just apw was asking about what mem size we nowadays use for kcrash [16:57] smb: i think i had a bug to bump it [16:57] it would be nice if we made it have an auto parameter [16:58] Yeah, at least for the small ram size it would not work in the past. So I think we just went to something for all... just cannot remember what value [16:58] 384M [16:59] oh ok. quite a bit for small systems but we probably not expect those being around that much [16:59] smb: we can specify ranges too [17:00] Yeah, I know. Problem was the initrd by default being too big [17:01] ## [17:01] ## Meeting starting now [17:01] ## === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues October 7th, 2014 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! If the question is should I file a bug for something, likely you can assume yes. [17:21] Hi, I'm trying to detect if a target kernel is a ubuntu kernel [17:21] I noticed the new UTS_UBUNTU_RELEASE_ABI symbol which is great [17:22] but I need to include generated/utsrelease.h to have it [17:22] which might not exist [17:22] why might it not exist [17:22] I don't see it on a vanilla kernel [17:25] hum actually I see it, but I'm wondering if this file always exists [17:31] i believe that is a purfectly standard header indeed, i am not sure you need to nclude generated though [17:31] isn't there a real wrapper for that ? [17:35] include/linux/vermagic.h:#include [17:35] ah great [18:03] Is there anything I should do for https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/1375874 ? [18:03] Ubuntu bug 1198180 in upstart (Ubuntu) "duplicate for #1375874 possible leak in upstart 1.5" [Medium,Confirmed] === ivan\_ is now known as ivan\ [20:24] The ddebs repo is the slowest thing I've ever seen. [20:25] good luck debugging the Ubuntu kernel to whomever is stupid/brave (choose the one you like most) enough to try. === hughhalf_ is now known as hughhalf