Skip to contentSkip to conversation

PROBLEM: cannot boot 2.6.15-rc6 on Opteron machine

DISCUSSION Erez Zilber 5 participants 10 messages Dec 29, 2005 – Dec 30, 2005

lore
Share & tools
lore
#0 lore raw

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and built it on RHAS 4:
make
make modules
make modules_install
make install

When I did that on an Opteron machine, after rebooting and selecting the 2.6.15-rc6 entry in the grub menu, I get the following:

root (hd0,0)
Filesystem type is ext2fs, partition type 0x83
kernel /vmlinuz-2.6.15-rc6 ro root=/dev/VolGroup00/LogVol00 rhgb quiet console=ttyS0,115200 console=tty0 [Linux-bzimage, setup=0x1400, size=0x1dfa8c]
initrd /initrd--2.6.15-rc6.img
[Linux-initrd @ 0x37f20000, 0xcff3f bytes]

Uncompressing Linux... Ok, booting the kernel.

Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(0,0)

I made sure that ext2 is compiled with the kernel (not as a module).

When I do the same on an emt64 machine, everything works ok. Any idea?

Thanks

Signature · 11 lines
____________________________________________________________

Erez Zilber | 972-9-971-7689

Software Engineer, Storage Team

Voltaire – _The Grid Backbone_

__

www.voltaire.com <http://www.voltaire.com/>
#1 lore raw
Quoted 3 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and

2.6.15-rc needs a newer udev version...

#2 lore raw
Quoted 14 lines · Arjan van de Ven
Quoted 5 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and

2.6.15-rc needs a newer udev version...

Redhat AS4 comes with udev 039. I couldn't find a newer rpm for this distribution, so I cannot replace the current udev rpm. I've downloaded the latest udev sources from http://www.us.kernel.org/pub/linux/utils/kernel/hotplug/ and installed it but I still get the same errors.

Erez

#3 lore raw
Quoted 11 lines · Erez Zilber
Quoted 9 lines · Arjan van de Ven
Quoted 5 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and

Quoted 7 lines
Quoted 5 lines

2.6.15-rc needs a newer udev version...

Redhat AS4 comes with udev 039. I couldn't find a newer rpm for this distribution, so I cannot replace the current udev rpm. I've downloaded the latest udev sources from http://www.us.kernel.org/pub/linux/utils/kernel/hotplug/ and installed it but I still get the same errors.

after you get a recent enough udev you probably need to recreate a new initrd (and pray that you don't need a newer mkinitrd to match the newer udev...) since udev is also put inside the initrd.

To be fair, enterprise linux distributions aren't actually designed to run brand spanking new kernels, and the vendors aren't likely to update userland tools to make newer kernels work either (unlike, say, distributions like Fedora Core or SuSE or .. or .. or .. )

#4 lore raw
Quoted 30 lines · Arjan van de Ven
Quoted 16 lines · Erez Zilber
Quoted 12 lines · Arjan van de Ven
Quoted 7 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and

Quoted 10 lines
Quoted 7 lines

2.6.15-rc needs a newer udev version...

Redhat AS4 comes with udev 039. I couldn't find a newer rpm for this distribution, so I cannot replace the current udev rpm. I've downloaded the latest udev sources from http://www.us.kernel.org/pub/linux/utils/kernel/hotplug/ and installed it but I still get the same errors.

Quoted 12 lines

after you get a recent enough udev you probably need to recreate a new initrd (and pray that you don't need a newer mkinitrd to match the newer udev...) since udev is also put inside the initrd.

To be fair, enterprise linux distributions aren't actually designed to run brand spanking new kernels, and the vendors aren't likely to update userland tools to make newer kernels work either (unlike, say, distributions like Fedora Core or SuSE or .. or .. or .. )

So, I understand that, currently (until I get a a recent enough udev), 2.6.15 cannot be used with RHAS-4 on Opteron machines. Anyway, thanks for your help.

Erez

#5 lore raw
Quoted 4 lines · Erez Zilber
Quoted 1 line

So, I understand that, currently (until I get a a recent enough udev), 2.6.15 cannot be used with RHAS-4 on Opteron machines. Anyway, thanks for your help.

I think that's too strong. udev setups are fragile and seem to be inventented by the devil to make kernel updates hell, but if you just compile the needed drivers and file systems for root statically into the kernel it should work. For network devices and other drivers not needed for root you can either load them manually early in a rc.d file or also compile them in.

Typically when udev goes wrong it just doesn't autoload the modules, but still processes the events when devices appear.

-Andi

#6 lore raw
In reply to Erez Zilber
Quoted 8 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and built it on RHAS 4:

When I did that on an Opteron machine, after rebooting and selecting the 2.6.15-rc6 entry in the grub menu, I get the following:

Quoted 6 lines

Uncompressing Linux... Ok, booting the kernel.

Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(0,0)

I made sure that ext2 is compiled with the kernel (not as a module).

When I do the same on an emt64 machine, everything works ok. Any idea?

Check on your /boot/grub.config. The above error message is typically seen when the root= parameter points to a symbolic disk label instead of a specific disk such as /dev/hda1. I'm not sure what the procedure for updating where the label points (possibly a mkinitrd) because I normally change the grub.conf to point to a specific disk.

#7 lore raw
Quoted 15 lines · Thomas Molina
Quoted 13 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and built it on RHAS 4:

When I did that on an Opteron machine, after rebooting and selecting the 2.6.15-rc6 entry in the grub menu, I get the following:

Uncompressing Linux... Ok, booting the kernel.

Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(0,0)

I made sure that ext2 is compiled with the kernel (not as a module).

Quoted 7 lines
Quoted 1 line

When I do the same on an emt64 machine, everything works ok. Any idea?

Check on your /boot/grub.config. The above error message is typically seen when the root= parameter points to a symbolic disk label instead of a specific disk such as /dev/hda1. I'm not sure what the procedure for updating where the label points (possibly a mkinitrd) because I normally change the grub.conf to point to a specific disk.

there is no such procedure, because the disk labels are... ON THE DISK. And the initrd reads them from all the disks at boot time to find the one needed. This means that if your disk changes name (for example because of a scsi bus order change or because of a different order you load the device drivers... or even if you forget to compile the sata drivers and suddenly the disk goes from /dev/sda to /dev/hda).... things just remain working

#8 lore raw
Quoted 37 lines · Arjan van de Ven
Quoted 15 lines · Thomas Molina
Quoted 11 lines · Erez Zilber

Hi,

I've downloaded kernel 2.6.15-rc6 (had the same problem with rc7) and built it on RHAS 4:

When I did that on an Opteron machine, after rebooting and selecting the 2.6.15-rc6 entry in the grub menu, I get the following:

Uncompressing Linux... Ok, booting the kernel.

Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(0,0)
Quoted 8 lines
Quoted 5 lines

I made sure that ext2 is compiled with the kernel (not as a module).

When I do the same on an emt64 machine, everything works ok. Any idea?

Check on your /boot/grub.config. The above error message is typically seen when the root= parameter points to a symbolic disk label instead of a specific disk such as /dev/hda1. I'm not sure what the procedure for updating where the label points (possibly a mkinitrd) because I normally change the grub.conf to point to a specific disk.

there is no such procedure, because the disk labels are... ON THE DISK. And the initrd reads them from all the disks at boot time to find the one needed. This means that if your disk changes name (for example because of a scsi bus order change or because of a different order you load the device drivers... or even if you forget to compile the sata drivers and suddenly the disk goes from /dev/sda to /dev/hda).... things just remain working

Quoted 7 lines

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Compile the kernel with support for your hardware built in, i'm assuming you eighter build the controller as a module or didn't built it at all

#9 lore raw
Quoted 8 lines
Quoted 8 lines

there is no such procedure, because the disk labels are... ON THE DISK. And the initrd reads them from all the disks at boot time to find the one needed. This means that if your disk changes name (for example because of a scsi bus order change or because of a different order you load the device drivers... or even if you forget to compile the sata drivers and suddenly the disk goes from /dev/sda to /dev/hda).... things just remain working

Interesting. I've been compiling my own kernels for quite some time and have never had a boot device end up anywhere except where I thought it should. This includes during the 2.5 craziness when parts of ide support was rewritten at least twice.

Quoted 3 lines
Quoted 1 line

Compile the kernel with support for your hardware built in, i'm assuming you eighter build the controller as a module or didn't built it at all

Certainly this is the best advice. Not only support for the hardware, but also ext3 journalling as well as ext2 support. I think it may be possible to end up with ext2 builtin but ext3 journalling as modular.