Hi,

I have big problems booting grub from HDD partition boot sector.

Disks - BIOS layout:
SATA 270GB
SATA 750GB
SATA RAID1 Volume (ICH9R) 36GB
IDE Master 500Gb #my target disk

Target disk Partitions(only those I use):
hda2 1Gb - primary(past 1024 cylinders), mounts on /boot ext3 - here I install grub
hda6 4Gb - logical, linux swap
hda9 28Gb - logical, mounts on / ext3

I install grub into partition boot sector and chainload to it from other bootloader whitch resides in 270Gb SATA disk MBR. This is not first test:
Slackware 12.2 + LILO - boots this way
Centos 5.2 + GRUB - boots this way
Debian 5.0 - boots from GRUB floppy (see below).


So, Installed Debian 5.0 and grub legacy using debian installer. Was unable to boot - got "GRUB _" with no reaction to any input.
So I made GRUB floppy and tried to boot debian - booted with no problems.
Next I Installed GRUB from GRUB shell booted from floppy, commands I used (there was /grub/stage1 and /grub/stage2 files on hda2):
root (hd3,1)
setup (hd3,1) #tried "setup (hd3,1) --force-lba" also

GRUB installed successfully except stage1_5(it should be that way as I remember). I don't use menu.lst so I should get GRUB shell after booting. So I tried to boot from hda2 and again got "GRUB _" with no reaction to any input.

As I mentioned before Centos 5.2 + GRUB installs and BOOTS!! just fine on exactly the same setup. I noticed few not standard things about centos setup(according to me): It allows to install grub only to partition of first disk(according to BIOS order) - so I had to reorder disks temporary. Also it installs successfully only if I supply BIOS disk order mentioned above. So it works.

What I'm doing wrong - why I'm unable to get grub shell after manual install ?
And most interesting question:
What centos installer does so special that it works ???


thanks in advance
Vilius

Hi Vilius

first, i have much less knowledge about OS/Linux then you!

I installed first OpenSuSE11.0 and later CentOS5.2 (both use the same /boot which is on a separate partition), had a "similar" situation with setup (hdx,x) which also was succesfully but with additional info "something wasn't ideal but no tragedy" and if i remember right i solved it later with editing the menu.lst of SuSE

I think-agree that after reboot you should have GRUB with GRUBshell as you described. (buffer-cache become consitent with quit-reboot)

I read, that if the /boot-partition is a separate one with a filesystem different than the filesystem of the /-partition (e.g. ext2 and ext3) so we should mount it after having done chroot before and so GRUB can recognize it, otherwise it could cause problems for GRUB in stage1_5, where he will hang himself. (is only a uncertain idea about!)

I'm sorry for my english. Best regards
kinimod

I read following too somewhere:

What you should be careful about is buffer cache. grub makes use of raw devices instead of filesystems that your operating systems serve, so there exists a potential problem that some cache inconsistency may corrupt your filesystems. What we recommend is:
If you can unmount drives to which GRUB may write any amount of data, unmount them before running grub. If a drive cannot be unmounted but can be mounted with the read-only flag, mount it in read-only mode. That should be secure.
If a drive must be mounted with the read-write flag, make sure that no activity is being done on it while the command grub is running. Reboot your operating system as soon as possible. This is probably not required if you follow the rules above, but reboot is the most secure way.
In addition, enter the command quit when you finish the installation. That is very important because quit makes the buffer cache consistent. Do not push <C-c>.

Whats is the other boot loader and how is the chainloading done?

Hi again

if you don't have done yet read following chapters:
(4.1.3.1 / 4.1.5) you are using grub, not lilo
4.6 / 4.6.1 / 4.6.2 do you use a own backed Kernel / initrd-Generator?
4.6.3
here (x86):
http://www.debian.org/releases/stabl....html#recovery


(4.7) use Grub version 0.97 or higher
ftp://alpha.gnu.org/gnu/grub/


maybe you want to try this:
start Debian via Grub diskette
look for Kernel+initrd (datas-entries)
reboot
start CentOS
make sure (in menu.lst) that the Kernel+initrd entry of Debian is right and the right one
restart

Hi,

I've solved that already:
http://www.linuxquestions.org/questi...-right-708919/

V