"Kernel panic - not syncing: No init found."

Hello. I have a question about building a new kernel. I just downloaded the appropriate kernel 2.6.12 today and trying to compile it, but running into this error that is in the Subject: "Kernel panic - not syncing: No init found."

I think I am doing everything correct. I am using grub as a bootloader.

Here is the output of my 'df -k' command (the important parts of it anyways):

Filesystem Mounted On
------------- ---------------
/dev/hdc2 /
/dev/hdc1 /boot
none /dev/shm


Don't ask me why its not /dev/hda, because I don't know. But it works for the other kernel.

My grub.conf looks like this:

***************************
default=0
timeout=10
splashimage=(hd0,0)/grub/splash.xpm.gz
title Red Hat Linux (2.4.20-8)
root (hd0,0)
kernel /vmlinuz-2.4.20-8 ro root=LABEL=/
initrd /initrd-2.4.20-8.img

title Red Hat Linux (2.6.12)
root (hd0,0)
kernel /vmlinuz-2.6.12 ro root=LABEL=/
initrd /initrd-2.6.12
*****************************************

The first kernel boots fine. I'm pretty new at this stuff, so I just basically left everything the same for the second kernel. I have the vmlinuz and initrd file in the right place, which is /boot. I'm not sure why it has root(hd0,0) on it, because when I do 'df -k' i'm using the hdc?? But it seems to not matter for the first kernel. I have no idea what is wrong. I don't know how I could have messed up the options in the kernel configuration.

Any ideas? It seems to be having trouble loading a few things. Oh and I also tried it with the option of "root=/dev/hdc2" instead of "root=LABEL=/".

Any help would be greatly appreciated! Thanks....

/dev/shm is a tmpfs partition, which suggests to me that your system is configured for a two-stage bootloader (hence the initrd bits). Firstly, make sure that you have compiled in support (not as a module) for the tmpfs filesystem, as you won't get very far booting off /dev/shm without it.

You should also make sure that you have compiled in support (not as a module) for the filesystem type of your /boot, /, /etc and /dev filesystems, otherwise you also won't get very far.

I think hd(0,0) refers to the location of the bootloader rather than the /boot partition.

Also, how far does the kernel get in booting before reporting this error?

Originally posted by rjlee /dev/shm is a tmpfs partition, which suggests to me that your system is configured for a two-stage bootloader (hence the initrd bits). Firstly, make sure that you have compiled in support (not as a module) for the tmpfs filesystem, as you won't get very far booting off /dev/shm without it. You should also make sure that you have compiled in support (not as a module) for the filesystem type of your /boot, /, /etc and /dev filesystems, otherwise you also won't get very far. I think hd(0,0) refers to the location of the bootloader rather than the /boot partition. Also, how far does the kernel get in booting before reporting this error?

1. I am not sure what you mean by compiled in support for the tmpfs. I would assume I would have to go back to 'make menuconfig' and load up my .config file and then go to filesystems and find tmpfs, correct?

2. I do not know what you mean by 'compiled in support for the filesystem type of your /boot, /, /etc.....'. If you mean by making sure I support ext3, ext2, etc....I did. I pretty much selected to support every type of FS there was. I can't remember if it was a module or not though. I doubt it.

3. The last few lines of the boot process are these (Filled with errors):
***********
Mounting /proc filesystem
Creaeting block devices
Creating root device
Mouting root filesystem
mount: error 19 mounting ext3
pivotroot: pivot_root(/sysroot,/sysroot/initrd) failed: 2
umount /initrd/proc failed: 2
Mounted devfs on /dev
Freeing unused kernel memory: 272k freed
Kernel panic - not syncing: No init found. Try passing init= option to kernel.
<6>input: AT Translated Set 2 keyboard on isa0060/serio0
input: ImPS/2 generic Wheel Mouse on isa0060/serio1
*************************

After this it just freezes. Looks like had problem mounting ext3 filesystem. I know I included it....maybe as a module. I'll go back and check. Any other suggestions though?

Thanks a lot for your help.

I just recompiled the kernel and included support for tmpfs and ext3 as part of the kernel, not as a module. I am getting the exact same errors as before.

Anyone help please?

I get the same problem with kernels configured with identical configs. My setup is:

ext3 over MD-RAID1 over Symbios SCSI controller.

ALL 3 modules and their dependencies are in my initrd images but the kernel panics are not being able to mount root.

Yet 2.6.9 boots up fine. I currently have 2.6.11, 2.6.11.12, 2.6.12 and 2.6.12.1 all ready to try out shortly to see when the problem was introduced.

FWIW it looks like one part of the bootup code is no longer waiting for another part to complete. For example my SCSI detection usually happens, then MD detection then finally root is mounted. However what I'm seeing it that the SCSI detection is happening but in the background, like the next part of the boots is not going to wait for it.

So when MD looks around for valid devices to build arrays for possible root mount, it can not see any, as the SCSI controller detection has not finished yet.

Confirmed 2.6.11.12 boots fine for me.

When booting 2.6.12 or 2.6.12.1 the initrd module loading appears to be happening in parallel.

I can see ext3 not finding the symbols from jbd (a dependant module) because ext3 is trying to load before jbd has registered its symbols.

I can see no disc detection occur, because the kernel is panic'ed due to no root before the SCSI driver has managed to scan the bus.

I can see no MD raid detection because this is happening before the SCSI driver has scanned the bus and registered the disk drives.

etc....

I'm sure if I compile EXT3 + RAID1 + SYM53C8XX into the kernel I will have a system that can mount root, however this problem is probably more widespread a concern because if modprobe is returning before having loaded the module it requested you will see hit and miss errors on bootup as tools that run after the module loading (like firewall rules after iptables loading) will be trying to configure the firewall before the firewall module is in the kernel.

From looking thru the LKML archive I can't see any discussion about an intended change to the module loading design so can only think its an introduced bug. The 2.6.11.12 -> 2.6.12 patch is a big one.

Fixed...

With help from LKML the details: http://www.ussg.iu.edu/hypermail/lin...06.3/0354.html

Original pickup of problem: http://www.ussg.iu.edu/hypermail/lin...06.0/1556.html


This works for me:

wget http://download.fedora.redhat.com/pu...-058-1.src.rpm
rpmbuild --rebuild udev-058-1.src.rpm
rpm -Uvh /usr/src/redhat/RPMS/i386/udev-058-1.i386.rpm

wget http://download.fedora.redhat.com/pu...2.17-1.src.rpm
rpmbuild --rebuild mkinitrd-4.2.17-1.src.rpm
rpm -Uvh /usr/src/redhat/RPMS/i386/mkinitrd-4.2.17-1.i386.rpm