View previous topic :: View next topic |
Author |
Message |
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Mon Aug 11, 2003 11:12 am Post subject: The new release is BROKEN - as in BUSTED-DOESN"T WORK-H |
|
|
Spent the entire weekend installing Gentoo - multiple times. Using genkernel as shown in the install document. Same result three times straight: "KERNEL PANIC: Attempted to kill Init" somewhere while scanning USB devices.
Please note that there are a number of problems with the new install guide. For example under Lilo - you define the label to be "jumped to" to be "linux", but label the linux section with "gk_linux".
I am so pissed - just wasted better than 20 hours of my weekend.... |
|
Back to top |
|
|
scriptkiddie l33t
Joined: 30 Mar 2003 Posts: 955
|
Posted: Mon Aug 11, 2003 1:08 pm Post subject: |
|
|
well.. instead of complaining, can you post some of your problems, with detailed descriptions? I'd be glad to help you.. I have installed Gentoo 1.4 on everything from a p2 300mhz machine to a P4 3.2ghz machine, plus a mac.. and all of my installs went fine. |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Mon Aug 11, 2003 1:27 pm Post subject: |
|
|
Started at stage1. Used genkernel after downloading gentoo-sources. Emerged hotplug, as described in manual, continued on to the 'reboot" step. Lilo hosed. NOTE: The install guide on the cd has "messages" entry, but there is no corresponding file in the bootloader. Stopped using install.txt from the CD, and went to the current install guide here. 8 hours or so shot....
Start over. Follow the NEW install guide exactly. through the reboot step. Lilo hosed, there is no "linux" entry - the install guide labels the linux boot as "gk_linux". Boot disk bad. Start over. another 8 or so hours shot.
Start over. Use stage 3 to attempt to save some time. Follow the new install guide exactly, but use "gk_linux" as the jump to label when setting up LILO. Goto reboot step. Machine boots, starts through the hotplug step, gets to USB devices, goes to "KERNEL PANIC: attempt to kill init". Caps lock and numlock keys flash.
Re-installed from beginning. Used openmosix-sources. Did the genkernel thing. Compile dies during SCSI module compile.
Start over. Try using GRUB bootloader instead of lilo. Did the genkernel thing with gentoo-sources. Reboot. "KERNEL PANIC: attempt to kill init" during hotplug detection of USB devices.
Better than 20 hours of download and install time shot. |
|
Back to top |
|
|
scriptkiddie l33t
Joined: 30 Mar 2003 Posts: 955
|
Posted: Mon Aug 11, 2003 1:49 pm Post subject: |
|
|
interesting.. why don't you try just compiling the kernel your self with gentoo-sources |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Mon Aug 11, 2003 1:58 pm Post subject: |
|
|
If I can find the energy to do this over again tonight, I'll try it without using genkernel, (and skip the hotplug stuff too) - the more I think about it, the more I'm convinced that there is a problem with initrd stuff. The only reason that I could imagine getting a kernel panic, due to trying to kill init, would be that a login shell doesn't get started, and you just "kind of fall off the end" of the rc.d and init.d stuff...
Last edited by rsbriggs on Mon Aug 11, 2003 2:00 pm; edited 1 time in total |
|
Back to top |
|
|
carbon Guru
Joined: 27 Jun 2003 Posts: 455 Location: New York
|
Posted: Mon Aug 11, 2003 1:59 pm Post subject: |
|
|
if problems happen with compiling so often, maybe you want to check your cpu's cooling, get the dust off, keep it cool and nice.
and try again. _________________ I do what I want, and that's what I do.
GNU World Domination
Carbon |
|
Back to top |
|
|
tyancey n00b
Joined: 20 May 2003 Posts: 35 Location: Maryland
|
Posted: Mon Aug 11, 2003 2:00 pm Post subject: |
|
|
Thats fine but lets try to fix the GenKernel panic problems. He is not the only one having the problem. I for one have problems with the whole bootsplash and progress bar. The GenKernel compile does all this for you (if it gets past the panic). I would like the install to work as advertised. Does anyone have a solution? |
|
Back to top |
|
|
wmgoree Apprentice
Joined: 08 Aug 2003 Posts: 246 Location: Alexandria, VA
|
Posted: Mon Aug 11, 2003 2:03 pm Post subject: |
|
|
rsbriggs wrote: |
Better than 20 hours of download and install time shot. |
Hmm... you seem to have isolated the problems to two main areas: genkernel and the bootloader. Two thoughts:
1. emerge your sources of choice (try gentoo or vanilla if you're getting problems) and compile your kernel with
(apologies if you know how to do this already; just trying to cover all bases)
cd /usr/src/linux
make menuconfig
make dep
make clean bzImage modules modules_install
The menuconfig part might help you isolate your kernel problems; genkernel isn't too fine-grained.
2. Rather than starting over each time (particularly with a bootloader or kernel problem), why not just boot from the installation CD and pick up where you left off? That is, if you're getting a kernel panic, boot off the CD, mount your partitions, do the chroot as described in the install manual, recompile the kernel, and keep going. Or, with the LILO problem, boot, mount, chroot, and mess around with the lilo config file. That way you don't have to spend 8 hours; it's probably only a couple of parts of the installation that you need to change. One of the cool things about the Gentoo install process is that it's an OS, not an application; if you want to take a break or try again later you don't lose the progress you already made. This, btw, is why I would recommend *against* the install guide's suggestion of removing the stage .tar.bz2 file you downloaded until you know your installation is how you want it, particularly if you have a slow connection. _________________ vi? *snicker* it doesn't even include a mail reader... |
|
Back to top |
|
|
bgzee Tux's lil' helper
Joined: 11 Aug 2003 Posts: 116
|
Posted: Mon Aug 11, 2003 2:04 pm Post subject: |
|
|
FYI, if you reboot after installing, you dont have to completely reinstall again if it doesnt work. you just need to remount your file systems, chroot and env-update and your back in. |
|
Back to top |
|
|
mennodeij n00b
Joined: 08 Aug 2003 Posts: 7
|
Posted: Mon Aug 11, 2003 2:57 pm Post subject: |
|
|
1. What does your /etc/fstab say? Does it say /dev/ROOT /dev/BOOT /dev/SWAP??
this should be changed to the real device names i.e. /dev/hda1 or something else
2. If you use lilo, don't use the append="root=/dev/ram0 init=/linuxrc" but instead of /dev/ram0 use your real root device
3. Try the ac-kernel-sources, they made my troubles, combined with the above, go away. I had exactly the same attempted to kill init thingie. It blows |
|
Back to top |
|
|
scriptkiddie l33t
Joined: 30 Mar 2003 Posts: 955
|
Posted: Mon Aug 11, 2003 3:01 pm Post subject: |
|
|
use grub... that might get rid of some of your toubles |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Mon Aug 11, 2003 3:04 pm Post subject: |
|
|
I'll try going through the entire procedure again tonight, but I'm going to build the kernel manually instead of using genkernel. Assuming *THAT* boots correctly, I'll re-build the kernel with genkernel instead and change the boot loader to point to the initrd- and kernel- images and see what happens....
There are several reasons that I prefer to start over from a stage-1 install, mostly based on bad past experiences with mixing several kernel compiles on the same system.
I now see about 7 posts on the subject of people receiving that same kernel panic error message. Does that say to anyone that there might be a problem here ????? |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Mon Aug 11, 2003 3:13 pm Post subject: |
|
|
mennodeij wrote: | 1. What does your /etc/fstab say? Does it say /dev/ROOT /dev/BOOT /dev/SWAP??
this should be changed to the real device names i.e. /dev/hda1 or something else
2. If you use lilo, don't use the append="root=/dev/ram0 init=/linuxrc" but instead of /dev/ram0 use your real root device
3. Try the ac-kernel-sources, they made my troubles, combined with the above, go away. I had exactly the same attempted to kill init thingie. It blows |
No - I had edited /etc/fstab - they are correct. My only concern is that the root and boot partition are about 60 GB into a 120 GB drive - my first couple of partitions are dedicated to Windows Advanced Server 2003 on that box, which is a 3.2 Ghz-HT Dell Server.
I prefer lilo and dislike grub, which is why I use lilo.
I'm not certain I understand the rationale behind number 2 - Are you saying that the installation manual is incorrect for suggesting it? I thought it figured into the entire initrd/kernel separation. ( I wouldn't use the line if I were booting from bzImage, of course. )
I would wonder why ac-kernel-sources works, and gentoo-sources doesn't. Something is major FUBAR if that is the case.... |
|
Back to top |
|
|
bgzee Tux's lil' helper
Joined: 11 Aug 2003 Posts: 116
|
Posted: Mon Aug 11, 2003 3:41 pm Post subject: |
|
|
I get this error too. It is after some usb detection, saying it found a usb mass storage device. I dont use usb, and dont have a usb storage device but for some reason it seems to think i do and then gives me the kernel panic. I even tried disabling usb in the kernel config and recompiled and still no luck. I think this may be a m/b chipset related issue, because i thnk everyone who has posted about this problem says they have a via chipset. |
|
Back to top |
|
|
zipdog n00b
Joined: 06 Aug 2003 Posts: 10 Location: Left Coast, USA
|
Posted: Mon Aug 11, 2003 3:57 pm Post subject: |
|
|
get ready to invest at least another 20 hrs after you've got your booting issues worked out. I'm sitting here in front of attempt number 4, compilind kde. holy crap this takes a long time. started my compile yesterday, 20 hours ago actually, and it's still going.
i quit and set use flags again and instead of 108 packages it's now on 17 of 29, which i started at 0430 this morning.
i'm sure it will be worthhhhh it. it damn well better be.
hehe |
|
Back to top |
|
|
AutoBot l33t
Joined: 22 Apr 2002 Posts: 968 Location: Usually Out
|
Posted: Mon Aug 11, 2003 4:32 pm Post subject: |
|
|
This sounds like a hotplug issue to me, manually configure your kernel and dont emerge hotplug. _________________ This message self destructed a long time ago. |
|
Back to top |
|
|
pilla Bodhisattva
Joined: 07 Aug 2002 Posts: 7729 Location: Underworld
|
Posted: Mon Aug 11, 2003 7:48 pm Post subject: |
|
|
See the front page for www.gentoo.org, some ISOs were shipped with a faulty genkernel. _________________ "I'm just very selective about the reality I choose to accept." -- Calvin |
|
Back to top |
|
|
cccomeau n00b
Joined: 11 Aug 2003 Posts: 5
|
Posted: Mon Aug 11, 2003 8:42 pm Post subject: |
|
|
Bloody Bastard wrote: | See the front page for www.gentoo.org, some ISOs were shipped with a faulty genkernel. |
I used the updated LiveCD for x86 with the fixed genkernel utility and still got a kernel panic after USB detection during the boot process.
I'll try replacing append="root=/dev/ram0 init=/linuxrc" with append="root=/dev/hda3 init=/linuxrc" in /etc/linux.conf when I get home tonight. I have no idea what that line does, but it's worth a try! |
|
Back to top |
|
|
bgzee Tux's lil' helper
Joined: 11 Aug 2003 Posts: 116
|
Posted: Mon Aug 11, 2003 9:46 pm Post subject: |
|
|
Yeah, those updated didnt affect me anyway, I used the athlon-xp live cd.
good call cccomeau, tell me how that works out for you. I might just get rid of that append line altogether. |
|
Back to top |
|
|
cccomeau n00b
Joined: 11 Aug 2003 Posts: 5
|
Posted: Tue Aug 12, 2003 3:12 am Post subject: It worked!!! |
|
|
cccomeau (myself) wrote: | I'll try replacing append="root=/dev/ram0 init=/linuxrc" with append="root=/dev/hda3 init=/linuxrc" in /etc/linux.conf when I get home tonight. I have no idea what that line does, but it's worth a try! |
It worked!!! Make sure you replace hda3 with the actual block device you are using. If you followed the installation instructions step by step, it should be hda3. I'll post the content of my lilo.conf file along with comments at work tomorrow morning. I'm typing this using Lynx, which makes it difficult. If you want to try this now, note that you DO NOT need to start the installation process all over to get to your lilo.conf file. Just boot off the LiveCd, mount your filesystems (Code listing 7.2 in the installation instructions), enter the chroot (Code listing 8.6), edit lilo.conf (nano -w /etc/lilo.conf), run LILO (/sbin/lilo), unmount your partitions and reboot (Code listing 26.1). I hope this is helpful. As I said, I'll post more details tomorrow morning. Cheers! |
|
Back to top |
|
|
cccomeau n00b
Joined: 11 Aug 2003 Posts: 5
|
Posted: Tue Aug 12, 2003 2:31 pm Post subject: Re: It worked!!! |
|
|
cccomeau (myself) wrote: | I'll post the content of my lilo.conf file along with comments at work tomorrow morning. |
As promised, here's my commented /etc/lilo.conf file. I find that the part about LILO in the installation instructions is not very clear and it took me a bit of fooling around to get it right. Hopefully it will work for you too!
To recap, I was getting "Kernel panic: Attempted to kill init!" after usb detection during the boot process. It seems that quite a few people have that same problem. I used gentoo-sources and genkernel.
#/etc/lilo.conf begins
boot=/dev/hda
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
lba32
#This should be the same as the "label" of your
#default OS of choice below.
default=gk_linux
#My Gentoo partition
#After running genkernel, you should have seen something like
#"Please specify /boot/kernel-2.4.20-gentoo-r5 and
#/boot/initrd-2.4.20-gentoo-r5 when customizing your boot loader
#configuragion file." This is where you do that. Of course, what
#you have to type in beside "image" and "init" below depends on
#which kernel you chose. If you didn't write it down, look in /boot/.
image=/boot/kernel-2.4.20-gentoo-r5
#This should be the same as "default" above
label=gk_linux
root=/dev/hda3
#Please see comment about "image" above.
initrd=/boot/initrd-2.4.20-gentoo-r5
#Replace hda3 with the actual block device you are using. If you
#followed the installation instructions step by step, it should be hda3.
#This is what was causing the kernel panic.
append="root=/dev/hda3 init=/linuxrc"
#My Window$ partition (which I never use)
other=/dev/hdb1
label=dos
#/etc/lilo.conf ends
Good luck! |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Tue Aug 12, 2003 2:47 pm Post subject: |
|
|
Bloody Bastard wrote: | See the front page for www.gentoo.org, some ISOs were shipped with a faulty genkernel. |
Nope - updated that, too, and still got the same problem.... Haven't tried the other fix - will get to that this evening.... |
|
Back to top |
|
|
progster Apprentice
Joined: 16 Jul 2002 Posts: 271
|
Posted: Tue Aug 12, 2003 4:20 pm Post subject: |
|
|
rsbriggs wrote: | Started at stage1. Used genkernel after downloading gentoo-sources. Emerged hotplug, as described in manual, continued on to the 'reboot" step. Lilo hosed. NOTE: The install guide on the cd has "messages" entry, but there is no corresponding file in the bootloader. Stopped using install.txt from the CD, and went to the current install guide here. 8 hours or so shot....
Start over. Follow the NEW install guide exactly. through the reboot step. Lilo hosed, there is no "linux" entry - the install guide labels the linux boot as "gk_linux". Boot disk bad. Start over. another 8 or so hours shot.
Start over. Use stage 3 to attempt to save some time. Follow the new install guide exactly, but use "gk_linux" as the jump to label when setting up LILO. Goto reboot step. Machine boots, starts through the hotplug step, gets to USB devices, goes to "KERNEL PANIC: attempt to kill init". Caps lock and numlock keys flash.
Re-installed from beginning. Used openmosix-sources. Did the genkernel thing. Compile dies during SCSI module compile.
Start over. Try using GRUB bootloader instead of lilo. Did the genkernel thing with gentoo-sources. Reboot. "KERNEL PANIC: attempt to kill init" during hotplug detection of USB devices.
Better than 20 hours of download and install time shot. |
You do realise you don't have to start all over again, just because your bootloader fails? The same for a badly setup kernel, you just chroot into your box (using the gentoo cd) and start from the gentoo-sources part. Recompiling kernel + boatload shouldn't take more than 1 hour (and that's allot if you ask me).
~Progster |
|
Back to top |
|
|
rsbriggs n00b
Joined: 30 Jul 2003 Posts: 45
|
Posted: Tue Aug 12, 2003 6:10 pm Post subject: |
|
|
Quote: | #Replace hda3 with the actual block device you are using. If you
#followed the installation instructions step by step, it should be hda3.
#This is what was causing the kernel panic.
append="root=/dev/hda3 init=/linuxrc" |
Hmmm. While this may "fix" the kernel panic problem, it seems to defeat the entire concept of using an initial ram disk (initrd). This couldn't be used to network boot a system, since a network boot device doesn't have a hard drive, and HAS to depend on the ram disk.
If you are going to do that, then there isn't any point in using a separate initrd and kernel image. Might just as well use the consolidated bzImage...
This leads me to believe that there is something broken with rc.d (like it fails to execute a login bash shell, or a config file somewhere that SHOULD be mounting / onto the ram disk but doesn't.
Well, the nice thing about using LILO is, you can make entries and different labels for multiple kernels compiled in different ways and try them, and not have to overwrite a kernel that you know works. If the latest kernel doesn't work, just reboot from one of the previous ones that do... |
|
Back to top |
|
|
|