Subject: Re: mirrored root w/ raidframe - howto install bootblock?
To: Hubert Feyrer <hubert@feyrer.de>
From: Brian Buhrow <buhrow@lothlorien.nfbcal.org>
List: port-sparc
Date: 05/14/2006 20:55:01
Hello. It's my understanding that you need to tell the boot sector
boot code where the start of ffs is when it's not at the beginning of the
disk or if the installboot program can't find it. Once the boot block
knows where ffs is on your disk, it can find the /boot file in the ffs
filesystem and load it for the second stage boot process. Block 64 is the
first block of ffs in your raid filesystem, the first 63 blocks being taken
up with raid header and raid label data. If you want to give the first
block of the boot file itself in the ffs filesystem, then use fsdb to get
the first block number relative to the beginning of the ffs filesystem
itself, and then add 64 to that to get the correct offset from the disk.
Remember, if your filesystem block size is not 512, then I'm pretty sure
you need to convert the fsdb data into 512 byte blocks. For example, if
your disk ffs block size is 4096, then you need to multiply the fsdb block
number by 8 to get the correct 512-block offset from the filesystem start.
Hope that helps.
-Brian
On May 14, 11:08pm, Hubert Feyrer wrote:
} Subject: Re: mirrored root w/ raidframe - howto install bootblock?
} On Sun, 14 May 2006, Brian Buhrow wrote:
} > Hello. I think you can use block number 64, because that's where the
} > contents of the raid filesystem begin. That should install, and the boot
} > block should then be able to read /boot from the ffs filesystem in the
} > raid, which begings at block 64.
}
} Can you please explain that '64' to me?
} As far as I understand, I need to find the block that /boot from raid0a is
} placed at, relative to the start of the physical disk. fsdb(8) told me the
} file is in blocks 1560 and up relative to raid0a, so I cannot imagine how
} this fits together so that the first stage bootloader that I'd like to
} put onto sd0a finds it at block #64? Or is my picture plain wrong?
}
}
} - Hubert
>-- End of excerpt from Hubert Feyrer