Hello,
I think I was right in guessing that the umount commands I was wondering in the last message should be given after exiting from the chrooted shell (/dev/sda1 file system).
Booting is not successful. The SATA disk is recognized but the root file system cannot be mounted. I get error 6 and a kernel panic soon after that.
As I do not understand the way lilo & initrd work, I have to ask: where is initrd supposed to be read from? Could it be possible that reading initrd is not possible simply because it is located in the EXT3 filesystem on a SATA partition?
Would it be possible to write the initrd on a cd and read it from there? I have 2 dvd drives on the system so this not an obstacle.
Again lost,
HD detection issue.
I think a custom kernel is needed
Hello, Ross & others!
I've been asking about this SATA problem from my friends using Linux. First, it seems my assumption about the boot failure being caused by the nonreadability of initrd seems to be wrong. I reviewed lilo.conf before I wrote a new boot sector with lilo.
As said - I am still getting the same errors:
Mounting ext3 root filesystem fails with error 6
pivotroot fails with error code 2
Would a custom kernel be an option?
Sincerely,
I've been asking about this SATA problem from my friends using Linux. First, it seems my assumption about the boot failure being caused by the nonreadability of initrd seems to be wrong. I reviewed lilo.conf before I wrote a new boot sector with lilo.
As said - I am still getting the same errors:
Mounting ext3 root filesystem fails with error 6
pivotroot fails with error code 2
Would a custom kernel be an option?
Sincerely,
I wish I could be more help, but I would be just guessing at a solution. You may be right that a different kernel with builtin (not module) support for your harware may be the answer.
Here is a link to a linux SATA site: http://www.linuxmafia.com/faq/Hardware/sata.html
Perhaps you will find something helpful there. Please let us know what success you have.
Ross
Here is a link to a linux SATA site: http://www.linuxmafia.com/faq/Hardware/sata.html
Perhaps you will find something helpful there. Please let us know what success you have.
Ross
SATA works when chipset support is compiled into the kernel
Hello,
I have made progress: I have managed to make an installation on an IDE-HDD so that the SATA-HDD partitions are automounted with no trouble.
The trick to do this is to compile a new kernel including the SATA support and the appropriate low-level SATA chipset module into the kernel. The tmb kernel that Ross has used is otherwise good (no wonder - it's made in Finland but it has the SATA support as modules.
They do not seem to work as modules.
Is it possible to make an own LiveCD from this installation I've made now?
Keep on trying - either your forehead or the wall will crack....
I have made progress: I have managed to make an installation on an IDE-HDD so that the SATA-HDD partitions are automounted with no trouble.
The trick to do this is to compile a new kernel including the SATA support and the appropriate low-level SATA chipset module into the kernel. The tmb kernel that Ross has used is otherwise good (no wonder - it's made in Finland but it has the SATA support as modules.
They do not seem to work as modules.
Is it possible to make an own LiveCD from this installation I've made now?
Keep on trying - either your forehead or the wall will crack....
Probably. There are some things in the tmb kernel that are needed, such as squashfs support, and an elusive patch that allows the long boot command lines that result from bootcodes to work. If you built the new kernel from the tmb kernel source, and just changed some config options, it will most likely work just fine.Is it possible to make an own LiveCD from this installation I've made now?
You're right, nothing works like dogged persistence.
Ross