=== zequence_ is now known as zequence | ||
apw | Hauke (N,BFTL), it does not seem like we have any obvious way currently indeed. i might be persuaded such a thing is reasonable. | 08:10 |
---|---|---|
jpds | Could somoene take a look at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1325260 ? | 09:05 |
ubot5 | Ubuntu bug 1325260 in linux (Ubuntu) "CompuLab Intense PC2 does not shutdown correctly on 14.04" [Low,Triaged] | 09:05 |
jpds | I've found the one line patch that fixes the issue. | 09:05 |
apw | jpds, ok | 09:05 |
xnox | "E: 10mount: mount: unknown filesystem type 'overlayfs'" hm, wtf..... | 10:28 |
brendand | xnox, can you reproduce that? | 10:36 |
apw | xnox, what kernel | 10:36 |
apw | xnox, cat /proc/version_signature | 10:37 |
mlankhorst | isn't it /proc/version, or uname -a ? :p | 10:37 |
xnox | apw: hm, looks like i'm running your kernel from the kernel ppa. Let me upgrade to 3.15-4 | 10:38 |
xnox | Linux version 3.15.0-031500rc5-generic (apw@gomeisa) | 10:38 |
xnox | rc5 sounds old, given that rc8 is out the door. | 10:38 |
brendand | xnox, aren't you using an ideapad yoga 13 too? | 10:39 |
apw | xnox, missing extras perhaps ? | 10:39 |
xnox | brendand: i am! | 10:39 |
brendand | xnox, updating the kernel usually means i have to rebuild wifi drivers | 10:39 |
brendand | xnox, so i avoid it | 10:39 |
xnox | brendand: and i have a bunch of things going wrong with it with 3.15 kernel, but it's slowly getting better. | 10:39 |
xnox | brendand: i have dkms module for that & it's no longer actually needed, as that driver got merged into staging in 3.14 and is enabled in our config | 10:40 |
brendand | xnox, oooh. cool | 10:40 |
xnox | brendand: well, it generates kernel oopses on suspend&resume though =( | 10:41 |
apw | xnox, getting better, thats no good, i need to inject some more "if (xnox) BUG" code clearly | 10:42 |
xnox | apw: michael fry punished me enough for last weeks breakages already | 10:44 |
apw | heh likely true | 10:44 |
apw | xnox, oh you are runing mainline kernels, yeah they do not have anything non-upstream in them | 10:45 |
xnox | apw: 3.15-4 is no good (locks up after graphical login, before compiz/unity get a chance to start) | 10:54 |
apw | xnox, lovely | 10:54 |
xnox | apw: and loads of oopses if i have any usb3 activity | 10:54 |
apw | xnox, what sort of lockup, can you ssh in etc | 10:54 |
xnox | apw: didn't check ssh, but graphics were unresponsive. | 10:55 |
apw | xnox, you bought an absolute turkey of a machine there didn't you, with hindsite buying a lenovo and having it spraypainted yellow sounds a more sensible plan | 10:55 |
xnox | apw: back on 3.13 kernel cause have "work" todo (aka break the archive) | 10:55 |
apw | heh, well have a look aback at syslog and see if it recorded anything before you wacked it with a stick | 10:55 |
xnox | apw: sabdfl liked it =) | 10:56 |
xnox | apw: yeap http://paste.ubuntu.com/7572245/ search for "cut here" same usb error over and over again. | 10:57 |
xnox | not sure if that's upstream rc4 kernel or 3.15-4 though =( | 10:58 |
xnox | /home/apw -> that's rc4 | 10:58 |
xnox | worth reporting at all? | 10:58 |
apw | i'd say not, unless you get it with -4 | 10:58 |
brendand | xnox, maybe it's worth noting that i didn't have any problem last week creating an armhf chroot | 11:23 |
brendand | xnox, i wonder was it broken in an update | 11:23 |
apw | jpds, ok test kernels are available in your bug | 14:24 |
jpds | apw: Ah, cool, though I won't have access to the hardware until at least next week.. | 14:26 |
* jpds tries to find someone who can turn the machine on. | 14:26 | |
manjo | rtg, can you enable trusty arm chroot on tangerine ? | 16:04 |
rtg | manjo, no support for arm chroots anymore. use the cross compiler in the amd64 chroot | 16:05 |
rtg | dpkg-buildpackage -B -aarmhf | 16:05 |
manjo | ah ok | 16:05 |
rtg | actually, dpkg-buildpackage -d -B -aarmhf | 16:06 |
manjo | dpkg-buildpackage -d -B -aarmhf -us -uc | 16:06 |
rtg | yeah, that works too | 16:06 |
=== hatch__ is now known as hatch | ||
manjo | rtg, dpkg-buildpackage -nc is supposed to recompile .. how do I make it not restart config ? | 17:42 |
manjo | rtg, coz I copied in a .config in build/ | 17:42 |
manjo | rtg, I don't want dpkg to ask me any config qs .. I want it to skip that ... I thought -nc would do it .. but it still is asking me Qs | 17:43 |
rtg | manjo, -nc is "do not clean". if you're usinga local .config, then perhaps you want 'make deb-pkg' instead. | 17:49 |
manjo | rtg, how doees that work with arm cross compile ... Ihave not used it | 17:52 |
rtg | manjo, make deb-pkg ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- | 17:55 |
manjo | rtg, sorry got kicked by irc .. how does make deb-pkg work with cross compile ? | 18:07 |
rtg | manjo, make deb-pkg ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- | 18:07 |
manjo | ok let me try that | 18:08 |
sconklin_ | arges: I know you're in sprint recovery but I'm willing to discuss the trusty net ns bug any time :-) We're unable to reproduce it anywhere outside our production environment, where we can't get a crash dump | 19:03 |
sconklin_ | maybe it's time for "debug by kprint" | 19:04 |
rtg | sconklin_, he's not back until Wednesday IIRC | 19:06 |
sconklin_ | ok, nice. Did he tack on some time off? | 19:07 |
rtg | yup | 19:07 |
sconklin_ | cool. | 19:07 |
stgraber | sconklin_: I spent half a day or so last week during the sprint to try and trigger this without much success... I've reproduced it a couple of times by accident but even with linux-coredump installed I can't get a core... My attempt at creating hundreds of netns, flooding them with millions of packets all doing round trips through netfilter + killing them at random didn't trigger any panic either... | 19:16 |
sconklin_ | stgraber: it's so frustrating that we can trivially reproduce this 100% of the time in our operating environment | 19:17 |
stgraber | sconklin_: what kind of load do you have in that prod environment? here I noticed it'd tend to happen a lot more when my machine was crazy busy than when it was mostly idle (it blew up in my face twice when doing kernel builds while playing with lxc...) | 19:18 |
sconklin_ | checking, stand by | 19:19 |
sconklin_ | stgraber: very lightly loaded. It's an instance that's only running our supervisory app and a single lxc container | 19:22 |
stgraber | sconklin_: interesting | 19:26 |
sconklin_ | stgraber: rsampaio_ just popped in here, he's been doing the testing and is a lot more familiar than I am with how it all works (or doesn't) | 19:27 |
stgraber | sconklin_, rsampaio_: do you also get the panic when killing a container or do you somehow get it during standard operation? | 19:27 |
sconklin_ | stgraber: and when I said "production environment" what I meant was "a test environment identical to our production environment" | 19:28 |
rsampaio_ | stgraber, sconklin_, the panic happens after we kill the container, few seconds later | 19:28 |
rsampaio_ | I was able to get the lockdep files before the panic happens | 19:29 |
stgraber | ok, so that matches what I've been seeing here the rare few times it happened to me | 19:29 |
rsampaio_ | I wish I could get a crashdump from the ec2 instance, it is very easy to reproduce in our environment | 19:30 |
rsampaio_ | I've tried the kvm image from Chris running for a couple hours but no panic | 19:31 |
stgraber | and I guess your EC2 instance depends on other instances so you can't just copy the fs to a local VM and reproduce the panic there? | 19:31 |
rsampaio_ | yeah there are a few other instances involved in the process | 19:34 |
balerion | hi anyone there ?? | 21:32 |
balerion | i need help with the kernels which control the display ( as in on screen/monitors ) | 21:34 |
balerion | can anyone help ?? | 21:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!