[SOLVED] Dual boot Ubuntu

Soji Antony · 03-02-2011, 05:34 AM

Hi
I am trying to install a dual boot system. I have installed RHEL 5.4 (sda1) first and then installed Ubuntu 9.10 (sda3). Now only Ubuntu is booting. Before installing Ubuntu the boot flag (* ) was set to first partition. This is my fdisk –l o/p: [ Before installing Ubuntu]

Code:

Device  Boot  Start  End  Blocks   id    System
/dev/sda1      *     1     765  6144831  83     Linux
/dev/sda2   	     766   1020 2048287+ 82     Swap

If I am not wrong os from partition marked as active[*] in MBR will be loaded into memory. But when I install Ubuntu the boot flag is still set to sda1 though Ubuntu is installed in sda3. This is my fdisk –l o/p after installing Ubuntu.

Code:

Device  Boot  Start  End  Blocks   id    System
/dev/sda1     *      1     765  6144831  83    Linux
/dev/sda2   	     766   1020 2048287+ 82    Swap
/dev/sda3            1021  2088 8578710  83    Linux

Q1: Will MBR rewritten each time when we install a new os?
Q2. Why MBR boots sda3 [Ubuntu] even though first partition is marked as active?
Q3. Please tell me what happens to MBR that makes first os unbootable after installing a second os?
Thanks in advance….

rizzy · 03-02-2011, 08:24 AM

Q1: Yes if you answer yes to "write grub to MBR?" question that appears on virtually all linux installers. Answer no and it won't write to MBR. But you'll need to run update-grub after installation to detect new OS if you chose not to write to MBR.
Q2: because Ubuntu wrote to MBR the partition table and where to boot from. grub will use that info to look for sda3 to boot from.
Q3: It got overwritten by ubuntu installer.

Question for you: why you use RHEL? Do you pay for support and updates? if not you are better off using CentOS

Soadyheid · 03-02-2011, 09:06 AM

I've just done exactly what you've been doing but with Centos 5. (100% Red Hat Code apparently)

Because I loaded Ubuntu second I've got Grub2 installed which doesn't appear to like Centos. This may be because Centos has arranged itsself in two partitions; /boot and a LVM partition. The latter may be due to the fact Centos is installed on a three disk hardware RAID 5 (HP Smartarray 600, I think)

I managed to fix it last night by running up Ubuntu and editing the Grub 40-Custom file which is read when

Code:

update-grub

is run. In /etc/grub.d I think, check the Ubuntu page mentioned below.

I mounted the Centos /boot "disk" and checked the entry in /boot/grub/grub.cfg (that's the one on the /boot "disk", not the Ubuntu /boot!) . Copied the Centos entry and pasted it into the 40-Custom file which I edited using nano.

You have to modify the entry as Centos uses Grub1 and the entry is a bit different. I googled "grub2 Ubuntu" and came up with a really good Ubuntu howto page. It'll also explain where the files; 40_custom, etc, live.

There's one line I added to the 40_custom file that mentions a floppy and a disk UUID... I used GParted to find the UUID No to allow me to insert this line and change the UUID. Note the pasted in line which is changed from "kernel" to "Linux" has a /dev/sdxn pointer rather han a UUID like the Ubuntu entry... Leave it with the /dev... entry.

I then ran

Code:

update-grub

which found my other OS'; Ubuntu 10.10, Ubuntu 10.10 Server, Fedora 15 and NO Centos

This is because you're manually entering it!

Rebooted the system and picked Centos from the Grub menu... Tah Dah!

All appeared OK.

If you can make any sense of my inane ramblings, I hope you get your system to work!

Play Bonny!

Soji Antony · 03-04-2011, 05:47 AM

Hi Rizzy , Soadyheid

Thank you so much for clearing my doubt ...

chalcogenide · 05-15-2012, 06:04 PM

Quote:

Originally Posted by Soadyheid

I've just done exactly what you've been doing but with Centos 5. (100% Red Hat Code apparently)

Because I loaded Ubuntu second I've got Grub2 installed which doesn't appear to like Centos. This may be because Centos has arranged itsself in two partitions; /boot and a LVM partition. The latter may be due to the fact Centos is installed on a three disk hardware RAID 5 (HP Smartarray 600, I think)

I managed to fix it last night by running up Ubuntu and editing the Grub 40-Custom file which is read when

Code:

update-grub

is run. In /etc/grub.d I think, check the Ubuntu page mentioned below.

I mounted the Centos /boot "disk" and checked the entry in /boot/grub/grub.cfg (that's the one on the /boot "disk", not the Ubuntu /boot!) . Copied the Centos entry and pasted it into the 40-Custom file which I edited using nano.

You have to modify the entry as Centos uses Grub1 and the entry is a bit different. I googled "grub2 Ubuntu" and came up with a really good Ubuntu howto page. It'll also explain where the files; 40_custom, etc, live.

There's one line I added to the 40_custom file that mentions a floppy and a disk UUID... I used GParted to find the UUID No to allow me to insert this line and change the UUID. Note the pasted in line which is changed from "kernel" to "Linux" has a /dev/sdxn pointer rather han a UUID like the Ubuntu entry... Leave it with the /dev... entry.

I then ran

Code:

update-grub

which found my other OS'; Ubuntu 10.10, Ubuntu 10.10 Server, Fedora 15 and NO Centos

This is because you're manually entering it!

Rebooted the system and picked Centos from the Grub menu... Tah Dah!

All appeared OK.

If you can make any sense of my inane ramblings, I hope you get your system to work!

Play Bonny!

Could you post your entry into 40_custom? We are wrestling with a similar problem.

Thanks

Soadyheid · 05-16-2012, 06:52 PM

Wow! This was over a year ago!!!

I'm afraid it's a bit late just now (Nearly 01:00am) so, providing I remember, I'll have to run up my Centos box and have a look. I need to sleep now.

Play Bonny!

chalcogenide · 05-17-2012, 09:38 AM

Soadyheid:

I abandoned raiding the Centos installation and everything worked. In fact, grub2 even saw the bootable image on the second drive.

Thanks