Felix Miata composed on 2015-10-23 15:08 (UTC-0400): > Felix Miata composed on 2015-10-21 18:02 (UTC-0400):
>> All my many installations are seriously multiboot, with typically no less >> than 10 distro installations within a HD's first year of use, and upwards of >> 30 partitions as HD warranty becomes a distant memory. >> In order to Keep It (multiboot) Stupidly Simple, most booting here is via >> Grub 0.97.x. So as a first attempt to revert from grub2 to grub-legacy I've >> been trying with a 64 bit intel CPU wheezy installation on host msi85. I did >> apt-get purge of os-prober and everything I could identify as related to >> grub2, rebooted, then tried apt-get install grub-legacy. That pulled in >> grub-common, claiming to have installed both. But, nothing reached >> /boot/grub/. > ... >> Enabling use as bootloader on the / partition can be a simple matter >> of creating a menu.lst and running setup from the grub shell. Thus, I rarely >> make use of any grub scripts personally managing multiboot configuration. > ... >> When I first tried to run grub from root shell in wheezy, first command >> attempted, as suggested by grub's man page, is find /boot/grub/stage1. This >> listed the first 3 instances, (hd0,0), (hd0,1), (hd0,7), then hung. I aborted >> with Ctrl-C, then rebooted after disconnecting hd1, and tried again, reaching >> same hang. Anyone know whether this hang would be a bug, known or previously >> unknown, or some pre-configuration or configuration required in a debian that >> isn't necessary in an rpm distro? I tried searching on bugs.debian.org, but >> as a 15 year user of various Bugzillas, I'm not making much sense of how the >> search by keyword(s) process is supposed to work on >> https://www.debian.org/Bugs/. I get an error message "An error occurred. >> Error was: You have to choose something to select by " but with no sign of >> what mandatory select list I'm missing. >> I tried running setup for wheezy from a grub> shell in openSUSE 13.1, but >> attempting to chainload boot from it, it only does >> chainloader (hd0,14)+1 >> GRUB Loading stage2.. >> and locks up the machine. >> I tried repeating setup from LMDE2 Betsy like I tried in openSUSE, but it >> hangs sooner than it did in wheezy, simply by hitting <return> after typing >> find /boot/grub/stage1. I tried to chroot to wheezy from lmde to run grub >> shell, but it hangs after showing instance 3 (hd0,7) same as run booting >> wheezy directly. > After perusing /usr/sbin/grub-install, as suggested by first thread response, > I tried: > # grub-install /dev/sda15 > The entire result is: > Searching for GRUB installation directory ... found: /boot/grub > _ > IOW, grub-install hangs in like manner to running setup from the grub> > prompt. Ctrl-Fn works. And after waiting about 5 minutes hoping some loop > might timeout, Ctrl-C succeeded to abort the script. Nothing shows up in > dmesg or /var/log/* to provide any recognizable clue why the failure. The > script did succeed to cp /usr/lib/grub/x86_64-pc/* to /boot/grub/, as all > *stage* files there now acquired a current date/time stamp (291,076 bytes in > 9 files), as did /boot/grub/default. > So, I booted a different host, and tried roughly the same overall process on > Jessie: > 1-purge grub2* > 2-install grub-legacy > 3-copy a device.map from a co-installed distro to /boot/grub/ > 4-grub-install /dev/sda21 > 5-# grub --> grub> find /boot/grub/stage1 > 6-create /boot/grub/menu.lst > Step 4 completed, but with an error message about some sort of stage1 read > failure. > Step 5 hung, instantly, so after waiting a bit, I aborted with Ctrl-C. Again, > no apparent clues are to be found in dmesg or /var/log/* > I attempted to chainload to Jessie from the main bootloader. All that results > is a grub> prompt from which a kernel command results in "error: can't find > command `kernel', and an initrd command results in > alloc magic is broken at 0xdd901170: 0. > Aborted. Press any key to exit. > No boot device available - > strike F1 to retry boot, F2 for setup utility > Press F5 to run onboard diagnostics > F1 triggered reboot. Again from Jessie boot, # grub --> grub> find > /boot/grub/stage1 hangs, so I can't help but thing the grub-legacy debs or > some deb(s) it/they depend on on the mirrors is broken. > Next I tried booting openSUSE 13.1, starting grub shell, and setting up > (hd0,20). It claimed to succeed. In attempting chainload to it and running > find /boot/grub/stage1, it outputs only (hd0,2), then hard locks the machine. > Retrying but instead executing 'root (0,20)' succeeds, only to be followed by > failure of any kernel to be found by any name offered. configfile > (hd0,20)/boot/grub/menu.lst is not found, while chainloader (hd0,2)+1 -> boot > produces expected result. Apparently lack of EXT4 support in grub-legacy on Debian is multifaceted. By first moving the installation to an EXT3 partition, then skipping the grub shell confirmation step (find /boot/grub/stage1), both root and setup can perform their functions correctly, installing grub to the PBR, to which chainloading works as expected. Now if only I could figure out how to get /boot/message to be recognized and used after 'aptitude install gfxboot'. :-p -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/
