I think that the installer (anaconda?) is broken in respect to
installing grub on ANY raid-1 disk.
I've tried 3 times to use the GUI install to install a simple IDE-based
raid-1 set-up (hda & hdc) and all three installs hung at the grub prompt.
In contrast, installing a raid-1 set-up with 3.4 (or 3.3) via GUI is
successful. However anaconda will ONLY write/install grub onto the IDE
disk of your choice (hda or hdc). Not both, which it should!!!
On my 3.4 installs, I reboot the system to make sure that the raid-1
set-up is successful booting off of your "primary" disk (in this case,
hda). Then I finish it by installing grub on the second disk (hdc),
like this,
# grub
grub> root (hd1,0)
Filesystem type is ext2fs, partition type 0xfd
grub> setup (hd1)
Checking if "/boot/grub/stage1" exists... yes
Checking if "/boot/grub/stage2" exists... yes
Checking if "/boot/grub/e2fs_stage1_5" exists... yes
Running "embed /boot/grub/e2fs_stage1_5 (hd1)"... 16 sectors are
embedded.
succeeded
Running "install /boot/grub/stage1 (hd1) (hd1)1+16 p
(hd1,0)/boot/grub/stage2 /boot/grub/grub.conf"... succeeded
Done.
grub> quit
I've tried to search Redhat's bugzilla about this but I couldn't turn up
anything. Note: I'm not that adept at searching their database.
So, I don't know if this is a Centos issue or a RHEL4 issue. My guess
is that this is a RHEL4 issue since the GUI raid-1 install using
centos-3.4 works sucessfully.
Michael
Anne Possoz wrote:
> John Tan wrote:
>
>
>>Yep, I put /boot on Linux software RAID-1 on all my machines. Dell has a
>>.pdf that explains how to install grub on the other device (actually,
>>it's more comprehensive than this, but that's the relevant part to this
>>discussion). Here's a listmail version of that .pdf:
>>
>>http://lists.us.dell.com/pipermail/linux-poweredge/2003-July/014331.html
>
>
> Interesting reading.
>
> My experience with kickstart installation of grub in the mbr is that
> it doesn't work for CentOS 4, at least using software raid.
>
> The relevant kickstart lines were:
> zerombr yes
> bootloader --location=mbr
> part raid.11 --ondisk=hda --asprimary --start 1826 --end 1950
> part raid.21 --ondisk=hdd --asprimary --start 1826 --end 1950
> raid /boot --fstype=ext2 --level=RAID1 --device md0 raid.11 raid.21
>
> But at reboot, no grub.
>
> So I reboot with "linux rescue" from the CD.
> Using the grub commands (with the ideas from the Dell reference)
> and knowing that my boot directory is in /dev/hda1 (/dev/hdd1)
> grub> device (hd0) /dev/hda
> grub> (hd0,0)
> grub setup (hd0)
>
> This solved the problem for next reboot.
>
> What went wrong? I kept the anaconda.log file from the installation
> (too bad that it is not saved in the /root directory) where I read:
>
> * moving (1) to step bootloadersetup
> * MBR not suitable as boot device; installing to partition
> (...)
> * GRUB command root (hd0,0)
> install /grub/stage1 d (hd0,0) /grub/stage2 p (hd0,0)/grub/grub.conf
>
> So there are 2 questions:
> - why did anaconda refuse to install grub in the mbr?
> - why my system not boating with grub in the partition?
> Maybe the mbr was not zeroed?
>
> If anyone understand more... I am pretty sure there is a bug there.
> I can show my ks.cfg and anaconda.log if usefull.
>
>
> Another concern that I have with kickstart is: how can I erase
> (or reuse) raid1 partitions without erasing others?
>
> Seems that the questions was frequently asked (google) but never
> answered. Any good clue?
>
>
> Thanks a lot for CentOS.
>
> Anne
>
> _______________________________________________
> CentOS mailing list
> CentOS@xxxxxxxxxxx
> http://lists.caosity.org/mailman/listinfo/centos
>
