If you want to be 100% certain then you must test the installation by simply disconnecting each disk in turn and booting your system to the grub prompt... - you dont even have to boot the machine beyond the grub prompt (unless you want to break/rebuild the raid as a further test). Either install technique should do the job but I am not convinced that Yast will install MBR code to a second hard disk. My personal position is that I like to use a command prompt for such important work and usually install manually. Hope this helps a little.... On 2 October 2011 07:27, Michael Fischer <michael@visv.net> wrote:
On Sat, Oct 01, dwgallien wrote:
I think this implies that grub is installed on your RAID 1 (I assume you are talking about a raid 1 /boot partition using the first partition on each disk). This doesn't tell you anything about whether grub has updated the MBR.
To install grub on both MBR's I would type (as root):
grub grub> root (hd0,0) grub> setup (hd0) grub> root (hd1,0) grub> setup (hd1) grub> quit
I would be interested in what others have to say on this and whether there is a way of specifying a "time out" so if the system can not boot from hd0 it attempts to boot from hd1? (my guess is this would be bios dependant)
On 30 September 2011 21:59, Michael Fischer <michael@visv.net> wrote:
Being one of the folks who tries to do the Subject, I recently ran across an old thread on the list which confirmed most of my experience with it, but mentioned that Yast (in the installer mode) might not be putting grub, etc. on both disks' MBR or /boot. After some web-digging, I tried the following:
grub> find /boot/grub/stage1 (hd0,0) (hd1,0)
Can someone competent in the matter confirm for me that the above does indeed mean that both disks in the array have had a working grub written to both disks, and therefore, in the event of the failure of one drive, the other is indeed bootable into a working system?
If not, why not and what should I do to "fix" the matter.
TIA.
Michael --
I've used this setup for a long time on my RAID 1 array.
Michael, your syntax puts grub in the boot sector of the first partition on the first and second boot disks. Philip's syntax puts grub in the MBR of those disks. Note that if grub is being installed from the running OS (as opposed to the grub shell with no OS running) it will use /boot/grub/device.map to determine which of the disks correspond to hd0 and hd1, which may or may not be the same as the bios boot device sequence. Assuming hd0 and hd1 are the bios first and second boot devices, if booting from the first disk fails then the bios will call the sector on the second disk. The array is immaterial to this process.
So... can you confirm that the grub output I posted means that I'm covered for a bootable drive in case either one (and only one at a time) goes bad on me? Or do I need to follow either Philip's grub commands (above) or yours (below) to insure such a setup?
Thanks much for the extensive explanations.
YaST's boot manager module can be instructed to install grub to both the partition boot sector and the master boot record. All YaST actually does is write out the file /etc/grub.conf with syntax similar to the above and then call grub in batch mode, the grub shell then does the actual install.
As an aside, a twist on the above is to install grub in the partition boot sector, mark that partition active, and then install the "generic" boot code to the MBR (so YaST uses grub to do the the former and the latter it does itself). All the bios does in every case is hand-off to the strap code in the MBR of the first active partition in the table. If grub is there, it has a pointer to the /boot partition and using the file system locates the kernel. If the generic code is in the MBR, that code calls the strap in the boot sector in the first primary partition which is marked active (or "bootable").
Michael -- Michael Fischer michael@visv.net -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-- "Clouds are not spheres, mountains are not cones, coastlines are not circles, and bark is not smooth, nor does lightning travel in a straight line." BENOIT MANDELBROT 1924 — 2010 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org