RHEL4 boot fails on DL360 G4

[David Cooke]

I just got an HP DL360 G4. This is the first HP server for our group. It has a RAID1+0 volume on the onboard 6i controller. It also has a 6400 controller with a RAID6 volume defined.

I used the RHEL4 CD that came with the system to install to the RAID1+0 volume on the 6i controller. The install went just as smoothly as other RHEL4-like installs I've done on other systems. When the install completed I removed the CD as instructed and hit enter to reboot. While booting up it gets to:
Attempting Boot From Hard Drive (C:)

That's it. Its done. I've verified that the 6i controller is first in the boot order. I can boot the RHEL4 CD1 into rescue mode and see the installed system. It just won't boot.

Any ideas?

[Colin Topliss]

Can you boot the system from the CD (ie on SuSE Linux I get an option to boot the installed system - this is different to booting directly off of the HDD)?

If that works, then its the boot manager which is at fault.

I used to get a problem on old Dell systems with this - it would fail to write the boot manager (LILO in those days), and I'd have to boot from the CD and manually update the MBR.

Col

[David Cooke]

I'm not using SUSE but translating that to what I know of RHEL4...
I booted rescue mode from CD1 then chroot to the installed image. Then grub-install /dev/cciss/c1d0. Then reboot.
Now it gets a little further. It says its booting linux then claims there is no such partition. The grub.conf file looks OK to me but this evidence seems to say grub isn't playing nice with the volume on the 6i controller.
Does anyone use grub and boot from the 6i controller? What's your grub.conf file look like?

[Craig Gilmore]

Here's a grub.conf file from RHEL4 that I installed last week.

GRUB Boot Loader
==================

default=1
fallback=2
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title Red Hat Enterprise Linux AS (2.6.9-34.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-34.EL ro root=/dev/VolGroup00/00lv rhgb quiet
initrd /initrd-2.6.9-34.EL.img
title Red Hat Enterprise Linux AS (2.6.9-34.ELsmp)
root (hd0,0)
kernel /vmlinuz-2.6.9-34.ELsmp ro root=/dev/VolGroup00/00lv rhgb quiet
initrd /initrd-2.6.9-34.ELsmp.img
title Red Hat Enterprise Linux AS (2.6.9-5.ELsmp)
root (hd0,0)
kernel /vmlinuz-2.6.9-5.ELsmp ro root=/dev/VolGroup00/00lv rhgb quiet
initrd /initrd-2.6.9-5.ELsmp.img
title Red Hat Enterprise Linux AS-up (2.6.9-5.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-5.EL ro root=/dev/VolGroup00/00lv rhgb quiet
initrd /initrd-2.6.9-5.EL.img

------------------------------

And an fdisk listing,

DISK PARTITIONS
===============


Disk /dev/cciss/c0d0: 72.8 GB, 72833679360 bytes
255 heads, 63 sectors/track, 8854 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/cciss/c0d0p1 * 1 25 200781 83 Linux
/dev/cciss/c0d0p2 26 2114 16779892+ 82 Linux swap
/dev/cciss/c0d0p3 2115 8854 54139050 8e Linux LVM

Disk /dev/cciss/c0d1: 72.8 GB, 72833679360 bytes
255 heads, 63 sectors/track, 8854 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/cciss/c0d1p1 * 1 2089 16779861 82 Linux swap
/dev/cciss/c0d1p2 2090 8854 54339862+ 8e Linux LVM

----------------------------

I think the problem is the system having issues finding the proper controller/partition for booting the system. Possibly getting the 6i and the 6400 controller confused?

[Jeroen Peereboom]

Can you configure the RAID controller from the BIOS?

JP

[David Cooke]

Craig - I think I agree with you. My grub.conf looks similar except it shows (hd1,0). I guess it conciders the volume on the 6400 controller to be (hd0,0). The 6i controller is first in the BIOS boot order so I'm not sure this makes sense. (I have an identical system purchased at the same time. I installed Windows Server 2003. It worked fine. Grrrr.) So something is confused about which controller/volume is which. I think I'm gonna go delete the RAID6 volume on the 6400 controller and see what happens.

Jeroen - Yes. I can configure both the 6i controller and the 6400 controller from their BIOS. At the system BIOS I can configure controller boot order. The 6i and 6400 can each be set to 1 or 3. The IDE controller can only be set to 2. I have the 6i set to 1 and the 6400 set to 3.

[David Riley_2]

I may have run into similar issues in the past with DL380 and QLogic cards. Once during a kernel upgrade and again during an OS upgrade from 2.1 to 3.0. I don't have an RHEL4 boot handy, but check if device.map is still in /boot/grub and if it even shows the onboard controller. I would boot and just end up at the grub prompt. Turned out on boot the system was looking for the QLogic to boot from, not the onboard where the OS was installed. If I remember right, the workaround was to remove the QLogic cards, do the grub install, verify it booted correctly before adding back the HBAs. Hope some of that helps.

[David Cooke]

I've got it booting. Initialy I configured both the 6i cotroller and the 6400 controller with RAID volumes and then did the RHEL4 install which fails to boot.

6i 6400
device: c1d0 c0d0
grub: (hd1,0) (hd0,0)

i.e: grub.conf was created using (hd1,0) to boot from the 6i controller.

I deleted the RAID volume on the 6400 controller. Still couldn't boot even after running grub-install in a rescue mode boot again. So I did a complete reinstall again. It boots! OK, what's different. Looking at partition devices and grub.conf again I see this:

6i 6400
device: c1d0 c0d0
grub: (hd0,0)

Now that I've read David Riley's reply, I take a peak at device.map and see the above confirmed.

It looks like whatever software sets up the grub configuration doesn't handle 2 RAID controllers that both have a RAID volume configured very well.

I rebooted and configured the 6400 controller to have a RAID6 volume again and the system still boots fine. I guess all is well now.

Thanks everyone.

[Andrew Cowan]

This is a common problem with RAID arrays such as the Promise SATA and many companies provide a Grub fix diskette that you run at the end of the installation process.

You should run this by doing a Ctrl-Alt-F2 at the congratulations screen, immediately before the final reboot.

[David Cooke]

Closing thread. Summerized in last message.