LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Kernel (https://www.linuxquestions.org/questions/linux-kernel-70/)
-   -   Kernel on Penguin box (https://www.linuxquestions.org/questions/linux-kernel-70/kernel-on-penguin-box-481111/)

Storm16 09-06-2006 10:12 PM

Kernel on Penguin box
 
I have a Penguin Relion 120 1U server with dual PIII/933 (Coppermine) CPUs, 2GB of ECC RAM and MegaRAID controller with dual 36GB SCSI drives.

Installed Debian, upgraded to sid. This gave me kernel 2.6.15-1-486 on install. Unfortunately, this server is slated to run VMware server, and since Debian is not on VMware's "supported" list, it has to compile the kernel. The gcc installed is 4.1.5, and the kernel was compiled under gcc-4.0.3, so the module will not load.

I attempted to install another canned kernel from the repository, and every time I do, the system gives me

Kernel panic: init not found. try passing init= to the kernel

I have tried 2.6.16-2-686-smp and 2.6.17-2-686. Both give me this error. On install (apt-get install <kernel-version>), it runs initrd, and I see the initrd-img for the appropriate kernel.

I didn't want to have to roll my own kernel for a server, and especially if I spend the time configuring and compiling it and it panics on me.

Does anyone have any recommendations/advice on getting this to boot something more current? (hopefully it was compiled with the same verison of gcc...)

Thanks,
--Storm

HappyTux 09-06-2006 10:48 PM

Quote:

Originally Posted by Storm16
I have a Penguin Relion 120 1U server with dual PIII/933 (Coppermine) CPUs, 2GB of ECC RAM and MegaRAID controller with dual 36GB SCSI drives.

Installed Debian, upgraded to sid. This gave me kernel 2.6.15-1-486 on install. Unfortunately, this server is slated to run VMware server, and since Debian is not on VMware's "supported" list, it has to compile the kernel. The gcc installed is 4.1.5, and the kernel was compiled under gcc-4.0.3, so the module will not load.

I attempted to install another canned kernel from the repository, and every time I do, the system gives me

Kernel panic: init not found. try passing init= to the kernel

I have tried 2.6.16-2-686-smp and 2.6.17-2-686. Both give me this error. On install (apt-get install <kernel-version>), it runs initrd, and I see the initrd-img for the appropriate kernel.

I didn't want to have to roll my own kernel for a server, and especially if I spend the time configuring and compiling it and it panics on me.

Does anyone have any recommendations/advice on getting this to boot something more current? (hopefully it was compiled with the same verison of gcc...)

Thanks,
--Storm


Check in your /boot/grub/menu.lst to make sure that there is an initrd line beneath the kernel line for the newer kernels like for the old one. For the vmware problem apt-get install gcc-4.0 then in the console where you will compile the modules export CC=/usr/bin/gcc-4.0 if this fails then ln -sfn /usr/bin/gcc-4.0 /usr/bin/gcc as root and try again do not forget to change the link back to the gcc-4.1because software will eventually expect it to be there.

Storm16 09-07-2006 09:01 AM

Quote:

Originally Posted by HappyTux
Check in your /boot/grub/menu.lst to make sure that there is an initrd line beneath the kernel line for the newer kernels like for the old one. For the vmware problem apt-get install gcc-4.0 then in the console where you will compile the modules export CC=/usr/bin/gcc-4.0 if this fails then ln -sfn /usr/bin/gcc-4.0 /usr/bin/gcc as root and try again do not forget to change the link back to the gcc-4.1because software will eventually expect it to be there.

Indeed there is:

Code:

title          Debian GNU/Linux, kernel 2.6.16-2-686-smp
root            (hd0,0)
kernel          /boot/vmlinuz-2.6.16-2-686-smp root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.16-2-686-smp
savedefault

title          Debian GNU/Linux, kernel 2.6.15-1-486
root            (hd0,0)
kernel          /boot/vmlinuz-2.6.15-1-486 root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.15-1-486
savedefault

Note that I delisted the recovery modes, as they are the same as the multiuser stanzas.

Regarding the gcc problem, it appears that gcc-4.0.3-6 is actually gcc-4.0.4, so I guess I'm going to either have to pull the gcc-4.0.3 from sarge or etch temporarily.

Well, that didn't work too well.

Code:

monitor:/var/cache/apt/archives# apt-get install gcc-4.0/testing gcc-4.0-base/testing
Reading package lists... Done
Building dependency tree... Done
Selected version 4.0.3-3 (Debian:testing) for gcc-4.0
Selected version 4.0.3-3 (Debian:testing) for gcc-4.0-base
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
  gcc-4.0: Depends: cpp-4.0 (= 4.0.3-3) but it is not going to be installed
E: Broken packages

--Storm

HappyTux 09-07-2006 10:24 AM

Quote:

Originally Posted by Storm16
Indeed there is:

Code:

title          Debian GNU/Linux, kernel 2.6.16-2-686-smp
root            (hd0,0)
kernel          /boot/vmlinuz-2.6.16-2-686-smp root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.16-2-686-smp
savedefault

title          Debian GNU/Linux, kernel 2.6.15-1-486
root            (hd0,0)
kernel          /boot/vmlinuz-2.6.15-1-486 root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.15-1-486
savedefault

Note that I delisted the recovery modes, as they are the same as the multiuser stanzas.

Well that is strange then what happens if you try booting by the grub command prompt use c when the grub splash comes up then type in the use enter key for the following lines.

Code:

root (hd0,0)
kernel          /boot/vmlinuz-2.6.16-2-686-smp root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.16-2-686-smp
boot

Now to save some typing when typing in the kenerl/initrd lines you can use the TAB key for auto-completion just like in a normal shell type in first few letters (eg. /boot/vm) then TAB.

Quote:

Regarding the gcc problem, it appears that gcc-4.0.3-6 is actually gcc-4.0.4, so I guess I'm going to either have to pull the gcc-4.0.3 from sarge or etch temporarily.

Well, that didn't work too well.

Code:

monitor:/var/cache/apt/archives# apt-get install gcc-4.0/testing gcc-4.0-base/testing
Reading package lists... Done
Building dependency tree... Done
Selected version 4.0.3-3 (Debian:testing) for gcc-4.0
Selected version 4.0.3-3 (Debian:testing) for gcc-4.0-base
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
  gcc-4.0: Depends: cpp-4.0 (= 4.0.3-3) but it is not going to be installed
E: Broken packages

--Storm
That should not matter you do not need exact minor versions (eg the .0.3 part) to compile modules for the kernel as there is no abi (should be binary compatible) change in minor versions and if vmware is requiring exact version then it is wacked, just try with the .0.4 version it should work.

Storm16 09-07-2006 10:53 PM

Quote:

Originally Posted by HappyTux
Well that is strange then what happens if you try booting by the grub command prompt use c when the grub splash comes up then type in the use enter key for the following lines.

Code:

root (hd0,0)
kernel          /boot/vmlinuz-2.6.16-2-686-smp root=/dev/sda1 ro
initrd          /boot/initrd.img-2.6.16-2-686-smp
boot

Now to save some typing when typing in the kenerl/initrd lines you can use the TAB key for auto-completion just like in a normal shell type in first few letters (eg. /boot/vm) then TAB.

Um, this caused a panic. I tried taking a picture of it with my daughter's camera phone, but I don't think it has the resolution to be readable. I'll fiddle with it tomorrow...But most of the errors are related to acpi. So I may be rolling my own kernel after all. I don't want to give up one of the processors or the extra optimization from 486 -> 686.

Quote:

That should not matter you do not need exact minor versions (eg the .0.3 part) to compile modules for the kernel as there is no abi (should be binary compatible) change in minor versions and if vmware is requiring exact version then it is wacked, just try with the .0.4 version it should work.
It did. I have an ongoing problem with proprietary drivers at work on a RH4.0 update 2 (CentOS 4.2, actually) box. The drivers will not load on our custom kernel, but it will on the stock kernel. The only difference between the two is the stock one was compiled with gcc-3.4.3 and gcc-3.4.4. So I thought the gcc versions had to be an exact match.

Thanks,
--Storm

HappyTux 09-08-2006 12:46 AM

Quote:

Originally Posted by Storm16
Um, this caused a panic. I tried taking a picture of it with my daughter's camera phone, but I don't think it has the resolution to be readable. I'll fiddle with it tomorrow...But most of the errors are related to acpi. So I may be rolling my own kernel after all. I don't want to give up one of the processors or the extra optimization from 486 -> 686.

Try adding noacpi to the end of the kernel line when trying to boot if it works then put it in the /boot/grub/menu.lst at the end of its line.

Quote:

It did. I have an ongoing problem with proprietary drivers at work on a RH4.0 update 2 (CentOS 4.2, actually) box. The drivers will not load on our custom kernel, but it will on the stock kernel. The only difference between the two is the stock one was compiled with gcc-3.4.3 and gcc-3.4.4. So I thought the gcc versions had to be an exact match.

Thanks,
--Storm
Haven't used any RH in years but that does seem strange it would not work on them could be some kind of compiler bug I would guess but really have no idea on that OS's problem. Good to hear you got it going.


All times are GMT -5. The time now is 06:57 PM.