[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: preset menu and disk access

From: Brett Russ
Subject: RE: preset menu and disk access
Date: Mon, 10 Mar 2003 20:40:10 -0500

I was really hoping someone could offer some help on this question.  It's
not really a bug so I can't file a bug report on it but if there is a more
appropriate forum please advise.


-----Original Message-----
Sent: Sunday, March 02, 2003 2:32 PM
To: bug-grub
Subject: preset menu and disk access

I want to use the preset-menu functionality to have grub failover to another
disk if the menu.lst file can't be read from the current disk.

An example of what I am talking about: currently I have 4 IDE disks loaded
and the grub stage2 image and the menu.lst file are on a partition which is
RAID1 across all 4 disks.  However, sometimes when I boot the primary master
(default boot) drive grub can't recognize the partition and therefore can't
find the menu.lst file which would normally tell it to go to another drive.
I've found that this problem is because the primary master drive is out of
sync with the RAID LUN.  In a situation such as this, it would be ideal to
have preset-menu instruct grub to go to (hd1) (or any other drive besides

First, is this possible?

Second, if so, how is the device map communicated to the preset menu?  I was
afraid that this functionality could only be used to set up serial
console/network boot since I didn't see any way to communicate the device
map to GRUB using the preset menu.

Last, if it is not possible, are there any suggestions as to how this
problem could be worked around?  Ideally I would like the ability to do
nested failovers, i.e. try disk 0; if not good then try disk 1; if not good,
keep moving on until something is found.

It seems that the very feature that makes grub so flexible--it's ability to
read a filesystem--also makes it prone to failures due to corrupt/out of
sync filesystems which have nothing to do with the GRUB image and
configuration files themselves.  Is this a case where LILO, with it's direct
sector addressing, would be a better bet for me to avoid such corruption

Thank you,

reply via email to

[Prev in Thread] Current Thread [Next in Thread]