login
Search
I'm using Debian sarge. Migrating from kernel 2.4.27 to 2.6.10
After solving problem with VSF and root partition, I get this error:
-----------------------
pivot_root: No such file or directory
/sbin/init: 426:cannot open /dev/console: No such file
Kernel panic - not syncing: Attempted to kill init!
-----------------------
I also created initrd.img-2.6.10 with mkinitrd. Ext3 & Jbd support included into kernel. Devfs is enabled on boot. Machine is laptop.
I can't find what 426 belongs to
Any suggestions?
hmm...
devfs is not longer supported and it's going to be removed in the next kernel releases... so get do "apt-get install udev"...
not what the poster wants now
I don't, this is a helpful answer to someone, whose machine doesn't boot. Even if the kernel people marked devfs for removal, at the moment it works. Have you read the question, or are you a robot dumbly searching for the string 'devfs' and posting standard comments?
Bullshit SIR!
Yes... he got a panic, while booting... cause of /dev/console... doesn't exist (maybe "MAKEDEV console" could do the job...)
pivot_root: No such file or directory
/sbin/init: 426:cannot open /dev/console: No such file
Kernel panic - not syncing: Attempted to kill init!
so... my question to you(strcmp is unsafe function, you should rename your nick to strncmp instead) is: "did you read it?"
nick
Btw, what is insecure about strcmp()? strcpy() and strcat() I understand, but with strcmp() (which only _reads_ data) the worst thing you can do is a SEGV (which is very difficult, because the memory is full of 0-bytes everywhere for padding, alignment, high bytes of small valued integers; and any program-supplied string argument counts as an implicit n because of its own string length) or a slowdown, if you manage to get it comparing very long strings all the time (but the same obstacles apply). And the string already has to reside in memory, so you probable got the problem at the point, where you read it in, strcmp() just uncovers it.
On the other hand with strncmp() you only compare _part_ of the strings in this case, which leaves the rest unchecked.
Debian 426 errro solution
Hi ,
I also got the same error and i feel before you start the system to debian you shoudl check whether the installation drive and current drive of booting are same or not.
for me i installed in /dev/hda1 and thus GRUB kept the image of bootign as root=/dev/hda1 and then by mistake i connected the drive
to secondary master that is /dev/hdc1 and GRUB tried finding kernal at /dev/hdc1 which was not present because OS is installed in /dev/hda1.
So better is that on GRUB you shoudl go to edit and the change the second line to root=/dev/hda1 and then boot . It shoudl boot after that or reconnect hard drive back to the ide-connector where it was actually connected at the time of installation.
Kshounish
I installed udev, reinstaled
I installed udev, reinstaled my kernel-image, but it didn't help
aftereffect?
After pivot_root failed, nothing in the file system is, where it belongs, so it is unsurprising, /dev/console (maybe the whole /dev) does not exist (because were expected to be found in the new_root not pivoted to). I don't know the initrd code, but I would assume, it loads the disk controller and file system code, mounts the device it has been told to be the root file system at some place, and pivot_root to this place. One of these steps may have failed, maybe your initrd.img is corrupt?
You used the --initrd parameter of make-kpkg to create your initrd along with the kernel-.deb? I never had unsolvable problems with the packaging scrips from the package 'kernel-package' . Iirc the cramfs-patch needed for the '--initrd'-initrd is part of the kernel.org-sources and you don't have to apply it/use the debian packaged sources (out of 'unstable', but being kernel.org + some fixes, i.e. more stable) (but the latter is a good idea anyway to get the security fixes in time).
I used kernel-source-2.6.10_2
I used kernel-source-2.6.10_2.6.10-4_all.deb package. I think, cramfs-patch is included here.
Now I'm trying to compile kernel with --initrd parametr: make-kpkg --initrd kernel_image (as you adviced)
Also one more strange thing: My root filesystem is ext3. I read, if ext3 & Jbc included into kernel, not by modules, I don't need initrd. I Included them into, but still need to 'mkinird ...' and add path to menu.lst (in grub) to solve problem with VFS (can't mount root partition).
compiling and installing with
compiling and installing with --initrd didn't solve the problem. Error is the same..
I mentioned ONE CHANGE. Now error is 431 (in past it was 426)
/initrd exists?
In trying to find out, what the number (probably a line number) means, and if the initrd executes a shell script or so, I read initrd(4). There is a rough description, what initrds are about and what they do while booting, containing the sentence "If the normal root file has directory /initrd, device /dev/ram0 is moved from / to /initrd", which is done via pivot_root, I suppose. Does a directory /initrd exist?
Sorry I can't tell you why without initrd it is even worse... But as I understand it, both methods fail at the first moment they depend on the contents of your root partition. Is something wrong with the partition table? Are the disks/disk controllers swapped for some reason? Can live CDs (e.g. knoppix with kernel2.6 boot option) mount your partitions?
Have you tried the debian ready made kernel images (with all modules) to rule out you forgot one important config option?
A comment to make-kpkg, I always pass a --append-to-version -my-version-string, because this way I am sure the kernel uses an individual /lib/modules/-subdirectory and doesn't try to load another kernel's modules...
It may be because of SATA disc
If you have a SATA disc, you should compile your kernel with proper drivers [sata drivers are under SCSI low level drivers section]. This should work.
I installed Sarge 2 weeks ago and I got the same error when i upgraded my kernel... I had the same problem with slackware, yet the error message was different. Anyway with NVIDIA driver and UPnP, everything returned to normal [I had to modify lilo.conf too since my root device changed to sda from hde]. I think I didnt have that SATA problem with 2.4 kernel because of NVIDIA's ide interface but i dont know why it didnt work with 2.6.10].
using sda breaks cdrom on hdc
I have a sata hard drive and a cdrom on ide. After i complete the changes in the /etc/fstab and /boot/grub/menu.lst to point to sda instead of hda. The kernel loads up but i cannot find a way to mount my cdrom drive.
Tried to use makedev to get the cdrom to be created as a device but that does not make any changes. Also there is no ref to the cdrom device in /dev/.udevdb either. So tried to use the one in /dev/.static/dev and that gives
error
If someone can point the way to get the kernel to recognise my cd drive. Til then i will have to work on the old config.
udev
Hi,
this is old-known problem (possibly) with udev on fedroa systems, when user doesn't use initrd. Solving is described in fedoraforum.org
You need move sode devices from /dev to /etc/udem/devices ... I don't remember correct path in etc directory now ;-)
I hohe I help you...
H.
When I changed bootloader (fr
When I changed bootloader (from Grub 0.94 to Grub 0.95) the problem mentioned above went away.
i'm using grub 0.95 + cvs****
i'm using grub 0.95 + cvs******
heh.. i can't find the answe
heh..
i can't find the answer on FF too :)
I had the same problem (with
I had the same problem (with error 426). As it turns out, somehow the root= line (in grub's menu.lst) for all of my debian-managed kernels was using /dev/hda instead of /dev/sda. This meant that pivot_root was trying to mount an ide cdrom instead of my scsi hard drive.
I have no idea why the device would have changed (as I've always used /dev/sda as my root, even from the initial debian install), and it's not something that I spotted quickly since it is just one character different.
I had the same problem
i am booting off a 16mb flash disk and my / partition is on a RAID5 running on a SATA 3ware 8506-8 controller. i ran into this same issue and was able to get around it.
i had another kernel that booted on the macahine. so i just added the correct module to /etc/mkinitrd/modules in my case 3w-xxxx then rerun mkinird and reboot.
After moving my Debian 3.1 Gu
After moving my Debian 3.1 Guest from VMware 2 to VMware 5.5 I encountered this same problem with this pivot panic, which was curious since my move this also meant a change from /dev/hda1 to /dev/sda1 as my Root Partition!!! In my case I traced the cause of the "break" to the "kopt" line within the Comments of GRUB's Config File, such that adding/remove any Kernels resets the Root Device back to /dev/hda1... Which under VMware 5.5 breaks the Guest since the newer VMware Releases now use SCSI Emulation for Disk Access within the Linux and NT Based Windows Guests!!!
Try to set root=/dev/sdaX ins
Try to set root=/dev/sdaX instead of root=/dev/hdaX in /etc/lilo.conf and run 'lilo'. Worked for me.
Might Also Try
I'm a Linux newbie. Had this same problem after changing the motherboard on my Debian box.
Found that the original GRUB config was trying to load root as hdc1, yet when the system went thru it's init, the IDE HD was showing as /hda1.
So, went into Grub, and changed to boot line to read hda (instead of hdc) and poof.. The system booted all the way to a login prompt.
Now, I'm sure there are other things I'll need to retune, but atleast now I can log in.
May be the answer is "libata"
May be the answer is "libata" module. Please try the patch.
binary at :
http://www.kernel.org/pub/linux/kernel/people/jgarzik/libata/
from :
http://wiki.osuosl.org/display/LNX/Debian+on+Dell+Servers?focusedComment...
doubt it's SATA
I am having the same problem with regular ata HD on a panasonic toughbook. i'll keep ya posted if i get it working
Hi there, did you got an a
Hi there,
did you got an answer??? I am using no SATA too and get this error, thus changing to sda did'nt work with me!
I tried to find some answer in several tutorials, but had no luck ;-(
Thanks!
I also had to change the root
I also had to change the root= line from "hda" to "sda" to get it to work. References to "hda" in /etc/fstab to "hda" also had to be changed to "sda".
Thanks .. fixed it.
I was seeing same problem as initial poster, and am using SATA, and switching to SDA fixed it when using the stock 2.6.8-2. And this let acpid work so my power button works. Thanks google, kerneltrap and anon source for solution! :)
i had the same problem, but i
i had the same problem, but it occured after my box crashed. and that was 2 hours after i set it up, so i was quite upset. :)
the point is i didn't upgrade anything and my kernel wouldn't boot. i also only had the netinstall cd, so no safedisks for me either. reinstalled now, if it happens again at least i know where to look for fixes. thanks.
*** ****** IT HAPPENED AGAIN!
GOD DAMNIT IT HAPPENED AGAIN!
could it be that booting windows screwed something up? i mean im really getting pissed off here, no other image would boot either.
and i don't have the /dev/sda problem, my disks are on /dev/hda.
any other ideas to what might be wrong? the box crashed when i was in windows by the way.
oh and btw my kernel is actua
oh and btw my kernel is actually 2.4.31! same error though.
RAID-1 or alike needs to be in initrd
This hint solved this problem for me:
See: http://www.aumund.org/node/1092
On Tue, 14 Jun 2005, Daniel Benden wrote:
> Solved: it is because of megaraid rename.
>
> mkinitrd is just including module megaraid, not megaraid_mm nor
> megaraid_mbox
>
> So add both modules to /etc/mkinitrd/modules and it is running again
> after mkinitrd rerun.
Kernel Upgrade 2.4 to 2.6 + SATA = Kernel Panic
Hi every body
Recently i have installed a debian version on my computeur
Asus p4p800se with a sata hard drive and udma harddrive
i have installed the last version of debian on my Udma disk
all is work
I upgrade my kernel to 2.6, all work again
But, no SATA
Ok, I re-compil my kernel with the sata option in SCSI menu
and... when I trying to boot, i have this error
VFS: Cannot open root device "hda7" or unknown-block(0,0)
Please append a correct "root=" boot option
Kernel panic - not syncing: VFS : Unable to mount root fs on unknown-block(0,0)
I think it's because, my root parition is not, now, in the hda7 device (because sata driver mix all device).
But... How to know, which device of my root debian (before hda7)
it's now call.
I trying to modify the /boot/grub/menu.lst file for this kernel
(even, kernel 2.4) ==>
Title Debian 2.4
root (hd0,6)
kernel /boot/vmlinuz-2.6.12.4 root=/dev/hda7 ro
to
root (hd0,6)
kernel /boot/vmlinuz-2.6.12.4 root=/dev/sda7 ro
or
root (sda,6)
kernel /boot/vmlinuz-2.6.12.4 root=/dev/sda7 ro
But, all not work....
PLEASE HELP
My hdd configuration
(Primary-Master) - UATA drive
(Primary-Slave) - Nothing
(Secondary-Master)- UATA DVD ROM
(Secondary-Slave) - Nothing
(third -Master 'SATA') - Hard Drive SATA
(Fourth -Master 'SATA') - Nothing
Primary master ====>
(win-XP[principal] ----- linux [logical]/ ---- Linux Swp[logical] --- Data1[logical] --- Data2[logical])
Third Master 'SATA' ==>
(Data1[logical] --- Data2[logical])
Problem fixed
Hello,
I had the same problem discussed above.
After several different tries, I used this in grub
title Debian GNU/Linux, kernel 2.6.8-2
root (hd0,0)
kernel /vmlinuz-2.6.8-2 root=/dev/sda3 ro
savedefault
boot
( 3rd partition is my root )
But you must!! change settings in fstab, too.
After booting normal the SATA HD is recognized as 1st SCSI Device
My System : Dell Poweredge 750 with Debian 3.1
P.S.: I installed the Package "udev", but don't know if it is used here
Always do this, i never arriv
Always do this, i never arrive to the fstab.
Now it's done, the probleme was the kernel compilation
I missing an scsi option.
It's done :)
Missing Kernel option
Hi,
do you remember which scsi option was the missing one?
Would be great when I could also solve that problem...
Regards,
Michael
I think found a path to solve this...
It worked for me:
-Turn off every fancy stuff on the config - this will save time compiling
-Read your documentation (RTMF)** and manage to compile the proper sata drivers in the kernel (no modules)- the same for file systems and scsi support.
-Get rid of initrd
-Try to boot - this will give you a hint of what sdX is the disk you are trying to boot. I had hdg (my second hd) and ended up with sdc - the flash-card drive in my HP-2200 got sda, believe it?
In grub, hit "e" over the kernel you are trying to run and change hdX to sdX, press "enter" and "b"
Fix your menu.lst accordingly and put the stuff you want back to the kernel
there you go...
Hope it helps.
Ricardo
**I spent the sunday compiling the wrong driver... it seems to be all about that...
Some updates on my previous...
When you turn off almost everything on the config file you will end up not knowing exactly what to turn back on... so expand the kernel three again and start over, with some background this time...
Disabling almost everything is still a good idea because it will give you the chance to see what happens when the scsi cards are probed => sdX values - not a bunch of USB mappings...
More on RTFM (sorry for the typo on the previous post): My bios says Promise Raid Controler - BUT- manual says ICH - what is in fact loaded, ICH5
finally, check "Y" to compres
finally, check "Y" to compressed rom file system, solved the problem.
*finally, check "Y" to compre
*finally, check "Y" to compressed rom file system, solved the problem.*
The only solution that worked for me. Thanks!
Software Raid Solution
I ran into the same error upon upgrading a debian sarge 2.4 kernel to a 2.6 kernel. Additionally I had setup software raid which makes use of the mdadm utilities (multiple device administation).
This issue would seem to be caused by my use of sata drives which are emulated as scsi devices by the 2.6 kernel, but detected as ide devices by the older 2.4 kernel.
It is trivial to change the /etc/mdadm/mdamd.conf file to reflect the correct devices but this only seems to affect the second stage of the linux boot process. The first stage (initrd - initial ram disk) has to be configured also. Unfortunately I could only run mkinitrd from the older 2.4 kernel and this caused the resulting initrd file to have regular ide drives setup (/dev/hdXX).
Solution:
While running the 2.4 kernel make a note of the major and minor numbers of the /dev/sdX block nodes.
#ls -l /dev/sd*
Backup and uncram the /boot/initrd.img-2.6.X file
#sudo cp /boot/initrd.img.2.6.X /boot/initrd.img.2.6.x.bak
#sudo /usr/sbin/cramfsck -x /tmp/initrd /boot/initrd.img.2.6.X
Create the scsi device nodes:
#sudo mknod /tmp/initrd/dev/sdXX b MAJOR MINOR
* use major and minor numbers noted from earlier step for each device
Edit the script file to reflect the new scsi devices
#sudo vim /tmp/initrd/script
change the line:
mdam -A /dev/mdX -R -u xxxxxxxx:xxxxxxxx:xxxxxxxx:xxxxxxxx /dev/hdXX /dev/hdXX
to:
mdam -A /dev/mdX -R -u xxxxxxxx:xxxxxxxx:xxxxxxxx:xxxxxxxx /dev/sdXX /dev/sdXX
Re-cram the edited initrd back to the boot directory:
#sudo /usr/sbin/mkcramfs /tmp/initrd /boot/initrd.img.2.6.X
Issues.
These changes allowed my system to boot to the 2.6 kernel, however currently the CD-ROM is not configured correctly and I am unable to mount it. I have not looked into this as it is not critically important but any sugestions would be appreciated.
I hope this information is helpful.
any solution for the cdrom ?
any solution for the cdrom ?
pivot_root error in debian on SATA disk
I have recently installed debian sarge 3.1r2 on /dev/sda and getting the following error when I try to boot debian. I have fedora in /dev/hda which is working without any issue.
-----------------------
pivot_root: No such file or directory
/sbin/init: 426:cannot open /dev/console: No such file
Kernel panic - not syncing: Attempted to kill init!
-----------------------
My grub.list contents:
---------------------
title Debian GNU/Linux, kernel 2.6.8-2-386
root (hd1,6)
kernel /vmlinuz-2.6.8-2-386 root=/dev/sda5 ro
initrd /initrd.img-2.6.8-2-386
savedefault
boot
------------------
I also checked /etc/fstab having the correct partition details, what could be the reason, any help? Is there any way to see detail error message for the kernel panic?
Thanks.
menu.lst and fstab (problem with SATA and kernel)
I have the same problem:
Kernel panic: Attempt to kill init!
/sbin/init: 432: cannot open dev/console: No such file
And I change my menu.lst from hde3 to sda3
And in fstab from hdeX to sdX (X=1,2,3,4,...)
And it works!
Seems Kernel 2.4 calls hdX to SATA hard disks and 2.6 calls sdX
bye!
soluton for cdrom
You can add to boot string: hdX=cdrom, where X is letter of your ide iface.
exactly what i needed, thanx
exactly what i needed, thanx :)
Kernel Panic
I did an upgrade from 2.4 to 2.6 and I had that error too. Changing the hd* to sd* fix partially the problem as I still would have errors of type ATA2 EMask...
After fiddling with this problem a few hours, I found that my ATA BIOS option on my 865PENeo2 motherboard was set to P-ATA & S-ATA. I have set it to S-ATA only and everything started working fine.
Kernel panic: Attempted to kill init
Even I had an earlier kernel 2.4, and I thought of recompiling it with kernel 2.6.6. Upon recompiling first I got the error for VFS, which resolved after running "mkinitrd". This was because the menu.lst file at /boot/grub did not have the initrd entry as with my earlier kernel 2.4. But then I got the error "Kernel panic:Attempted to kill init"
and this is how it worked for me. In the earlier compilation I did not use the option "--initrd" in the "make-kpkg" command. So this time I thought of recompiling the kernel again with the "--initrd" option. And now the initrd entry was automatically done in the menu.lst during the installation. I rebooted, and the system worked fine. Try this one.