Hello,

there is first time for everything. I had OpenSuSE, CentOS 5 and Window$ installed. There were 3 physical hard drives. Bootloader was on OpenSuSE hard disk and I remove it because of media error. After that I installed GRUB onto CentOS hard disk, set up 'grub.conf' according to existing hardware setup. After reboot everything worked great but after I shut down machine and turn it back on after 2 hours I had a GRUB error with only 'GRUB' on the screen.

Then I reinstalled GRUB from install media and GRUB is OK after that and everything boots as I planned. But after next power down same error occurs.

Any ideas ?

Thanks,

Josip

Hi.

My first thought is that the BIOS thinks the boot drive is some other drive ... cheers, makyo

Thanks for the reply,

My thoughts precisely. I also rearranged boot order and after that it started to boot but when I power off it fails again. It just came to me: there is also setting in BIOS called "reset configuration data". I'll try it in next few minutes.

Thanks,

Josip

Nope !

More curiously, after changes in BIOS which are regarding to boot parameters GRUB boots OK but after machine is powered down same problem persists.

Maybe I'll try to remove GRUB and install it again. Can that help ?

Josip

Hi.

My second thought is the battery that keeps the RTC/NVRAM alive. I haven't had a battery die on an Intel box, but I had a PowerPC that lost time, knowledge of boot volume, etc., when the battery ran down. The batteries last for years -- how old is the motherboard in the box? ... cheers, makyo

Check if the drive order during boot time matches the drive order when running. When the machine first boots, it may see (hd0) as the Linux drive, but afterwards the kernel will say it's /dev/sdb. If you can't update grub, because the install drive is the wrong one after booting, this could be the problem. The fix is to edit /boot/grub/device.map which contains a mapping between the (hdN,N) nomenclature of grub and /dev/sdXN of the kernel.

If the computer is over 3 years old, it could be the motherboard battery. But I would expect that to be the problem after a cold reboot and not a warm one, unless it has a short, and not just dead.

I checked it already.

Battery is in excellent condition, this is brand new Intel BadAxe 2 - D975XBX2 board - so I ruled that out. All data in BIOS stays the same as before. No changes because this BIOS has way to check it out so every anomaly is reported at next reboot.

Thanks,

Josip

I never checked that, thanks. I'll try - hope that will solve the problem.

Josip

Solved !

Obviously, hard drive has two kinds of master mode (with and without slave attached). After I disconnected malfunctioned drive with OpenSuSE which was on same cable (it is on PATA interface), drive with CentOS stays in master mode but jumper setting was not set to "stand-alone" mode. I completely forgot that Maxtors sometimes do so.

After I set jumper properly everything worked great !

Thank you all !

Josip