--- Frans Pop <[EMAIL PROTECTED]> wrote: > reassing 276100 grub-installer > retitle 276100 Should take boot sequence defined in > BIOS into account > tags 276100 d-i > thanks > > On Tuesday 12 October 2004 02:35, [EMAIL PROTECTED] > wrote: > > (1) any of the guided partition schemes using > largest > > block of free space wanted to create an active > primary > > partition for the root filesystem after the > extended > > partition. > > I'm guessing that the free space started _after_ the > end of the extended > partition. If the free space had been enclosed > within the existing extended > partition it should have created the new partitions > as logical.
yes, i believe it was completely unallocated space at the end of the drive. so i understand d-i's partitioning methods with your explaination. i have had problems in the past (my pre-debian days) with really unconventional partioning (like a primary after the extended); which is why i shy away from setups like that. i'll just keep manually partitioning for that little extra piece of mind. :) > > (2) first attempt: allowed d-i to install grub on > mbr > > of 'first drive' as it did properly detect the > > existing windows install (1 of them, identified as > > 2000/xp probably because of the xp bootloader). > > however, it must have considered the ide drive as > the > > 'first drive' as that's where it put grub. > > > > second attempt: knowing the above, manually chose > > /dev/sda to install grub to mbr of the sata drive. > it > > installed there as expected, however since d-i saw > the > > ide as the 'first drive' and sata as the second it > > configured grub to boot to hd1,0 for windows and > hd1,5 > > for linux. > > > (3) which caused grub to not boot anything. had to > > manually edit grub paramaters to read hd(0,5) to > boot > > linux. once to console, had to edit > > /boot/grub/menu.lst to read hd(0,0) for windows > and > > hd(0,5) for linux entries. > > It looks like the boot sequence definition in your > BIOS is different from the > sequence in which the disks are detected by d-i. the boot sequence is configured: 1)floppy, 2)cdrom, 3)scsi/sata. then there's a separate configuration for choosing to boot to a pci scsi card or on-board sata (which is set for sata, obviously). > I'll reassign this report to grub-installer, but I'm > not sure if we will be > able to do autodetect this. Maybe we should just add > a question which is the > primairy boot device if different types of disk > controller are detected. in this instance, shouldnt the actual location of the windows installations, existing mbr and xp bootloader, be enough to determine what the boot drive is? i think someone will need to look at whether it's because my sata is jbod and not a raid array. i do not have a second sata drive to test this with a sata raid configuration. i can, however, move the ide drive to secondary ide controller to see if that makes a difference. i'll do that in a few days, in case someone needs additional information from this install as it sits now. i had zero'd the ide drive before partitioning it since the first attempt had put grub on it's mbr. there's no way that it should have been identified as the boot drive. will i run into problems with grub getting misconfigured again if grub or the kernel gets updated? or will those upgrade scripts leave my edits to /boot/grub/menu.lst alone? i thought i saw a kernel upgrade already. will try that before i rearrange the drives to have another go. > > (4) during base-config's tasksel, i chose only > manual > > package selection and got a slew of pre-selected > > packages (over 80, totalling 150+mb installed). > > This is because these packages are priority > 'standard', the definition of > which is that they will be installed by default. is that something new? i haven't experienced that before. can it be avoided or are other packages now going to depend on "standard" packages being installed? thanks for the quick input, frans. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
