Quote:
|
Originally Posted by JEO
Many BIOSes will not try to boot a drive even if it has grub in the MBR if they detect that no partitions on the drive are set active (parted boot flag on). This has bitten me many times on several systems. One partition on a bootable drive should be set active. For grub in the MBR, any one primary partiton set active will allow it to boot.
|
Thanks. That's a good explanation for it. After all, we've gotta figure that it often happens that a drive is completely repartitioned and only Fedora is installed. In that situation and with the issue you mentioned, it could make a difference.
I just thought of something related to this. I use NTLoader to boot XP, Fedora 10, and Fedora 11 on my desktop. When I installed Fedora 11, I chose to install GRUB in the first sector of the Fedora 11 boot partition like I always do. Anaconda did not change the active partition because NTLoader never stopped working. So I guess I can conclude that the active partition is changed to the Fedora 11 boot partition when GRUB is installed in the master boot record and not when it is installed in the Fedora boot partition.
It's still a new development with Fedora 11. I just did some searching for threads with Fedora 10 fdisk reports that also involved Windows, and the Windows partitions always stayed the active partition.
UPDATE: I'm now not so sure about this...
Quote:
|
Originally Posted by me above
So I guess I can conclude that the active partition is changed to the Fedora 11 boot partition when GRUB is installed in the master boot record and not when it is installed in the Fedora boot partition.
|
A day or so ago I installed Fedora 11 on my Thinkpad with XP, Fedora 10, Debian, and Ubuntu. It replaced Fedora 10. I use the XP boot loader to boot everything. As usual, I chose to install GRUB in the first sector of the Fedora 11 boot partition. When I rebooted, Fedora 11 booted directly. Anaconda had changed the active partition to the Fedora 11 boot partition, and the MS boot code in the master boot record was booting it. It was easy to fix by changing the active partition back to XP (probably the explanation for my forgetting about it from the time before). Until I learn otherwise, I now am of the belief that Fedora 11's Anaconda always changes the active partition. I'm not interested enough to do a bunch of installs to find out with certainty. Besides, it doesn't really matter unless you're using NTLoader like I do.
ANOTHER UPDATE: Okay now it's clear what happened in my two cases. The reason installing Fedora 11 did not bust the Windows boot loader of my desktop is because Fedora 11 was installed on a different hard drive. Anaconda changed the active partition of that drive, but it didn't matter to NTLoader because it was safely located on the first hard drive. And when I installed Fedora 11 on my Thinkpad which has only one hard drive, NTLoader was busted by installing Fedora 11. It was repaired by changing the active partition back to the XP partition. So the final conclusion is that Fedora 11's Anaconda always changes the active partition.