Re: Kernel panic - Attempted to kill init!
On 05/16/2015 05:03 AM, Ben Armstrong wrote:
On 15/05/15 06:04 PM, Marc F. Clemente wrote:
I tried re-building in /root, and still no luck. Any other ideas?
su
mkdir /root/live
cd /root/live
lb clean
lb config --distribution sid
lb build
Why did you build sid here and stretch in the previous test you pasted?
We should at least compare apples and apples.
Agreed. Ideally, I would like sid/amd64. But in an effort to get
something working I tried different combinations of distribution and
release.
[ 2.548424] Kernel panic - not syncing: Attempted to kill init!
exitcode=0x00007f00
[ 2.548424]
[ 2.549079] CPU: 0 PID: 1 Comm: init Not tainted 4.0.0-1-amd64 #1
Debian 4.0.2-1
[ 2.549430] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS 1.8.1-20150428_134905-gandalf 04/01/2014
The sid kernel you show here is 4.0.0-1-amd64 which no longer has aufs.
You must use overlayfs (which is not configured by default yet). So
regardless of whether or not this panic relates to that, please stick
with stretch for now (which should be the default).
That's fine. We'll stick with stretch/amd64.
Also, I did not notice until re-examining your first paste and examining
this one that you're using qemu. Please try booting the image on real
hardware instead.
My goal is to get a USB stick running on real hardware. It didn't work
on real hardware. It's difficult to capture the console before the
panic on real hardware, so that's why I was using qemu.
Please also keep a complete log of your build. We don't know what errors
might have occurred in the build that you are overlooking. It is not
sufficient to simply check if in image is created or not. If vital
components are missing from your build, the image will not boot.
I will look at build logs to see if any messages or errors come up.
What's the preferred way to get build logs? Something like "lb build >
logfile.txt" or is there something better?
And finally, you did not indicate the exact version of live-build you
are using. Since you said you're using up-to-date packages on unstable,
I assume it's 5.0~a7-1, but you should always include versions of live-*
components when reporting issues.
I had 5.0~a6-1, but I upgraded to 5.0~a7-1. It was even worse; I
couldn't even get an image created. Commands:
su ; mkdir /root/live ; cd /root/live ; lb clean ; lb config ; lb build
lb build exited with an error:
no such script: binary_grub
So I just downgraded back to 5.0-a6-1. Same commands as above, I can
get an image built. The build logs do not seem to indicate any anomaly.
If I run that image on qemu, I still get a kernel panic, with the
following nugget right before the panic:
[ 2.908847] Freeing unused kernel memory: 216K (ffff8800017ca000 -
ffff880001800000)
/bin/sh: error while loading shared libraries:
/lib/x86_64-linux-gnu/libc.so.6: invalid ELF header
[ 2.977436] Kernel panic - not syncing: Attempted to kill init!
exitcode=0x00007f00
If I move the image to a USB drive and boot on "real" hardware, I get...
a bootable system. This leaves a few unanswered questions. What's
wrong with 5.0~a7-1? What's wrong with qemu and/or the binary image
that they won't run together? N.B. The Debian installer
(--debian-installer netboot) will run just fine under qemu.
Marc
Reply to: