Feisty doesn't start

I think I had a similar problem, assuming mdadm is the problem. At boot press escape to get into the boot loader, GRUB, and then try booting into what's labelled 'recovery mode' to get more details.

If mdadm and mounting the disks *is* the problem, a solution may be to edit the boot sequence (from GRUB) adding " break=mount " to end of the line about the kernel. Then during boot, you are presented with a shell, leaving the mounting is left up to you. My problem was my system was trying to mount my disk twice, so all I had to do was wait a few seconds until the other task had mounted the disk, and type exit at the shell.

see this bug perhaps
https://bugs.launchpad.net/bugs/103177

Ok I do what you told to do:
Then during boot, you are presented with a shell, leaving the mounting is left up to you.
But I got the exact same error:
mdadm: No devices listed in conf file were found

I don't know why the mdadm is on my system, I have only one SATA disk 80gb.
Anyone with an idea? please
Thx for the help

Can you give us some more details on your system? Are you using LVM (logical volume management) or any other kind multi-disk (MD) raid? You won't be, unless you specifically set them up during install; Ubuntu doesn't use them by default. (I use LVM, but not for my root partition)

1. Boot into recovery mode (with break=mount), and look for messages/errors relating to disks (eg. unable to mount /dev/hda1, device not found) . At the shell, wait until messages have ceased then calmly type 'exit'. Does that work? That's what worked for me [I get the "mdadm: No devices listed in conf file were found" message every time I boot, but the boot continues fine]. If it is successful, you will get to single user mode (a root shell), at which typing 'exit' again will take you runlevel 3, X, where you like to be.

OR
2. If you aren't using LVM or a multi-disk RAID then you don't need to be running mdadm. Boot using the liveCD.

Follow this guide.
http://www.ubuntugeek.com/how-to-fix-broken-ubuntu-feisty-fawn.html
 It allows you to get into your current installation from the liveCD (by running chroot). You may need /dev/hda1 or /dev/hdb1 where they have /dev/sda1

Once you have chroot'ed run this:
sudo apt-get remove mdadm

to remove mdadm

Hi again. I'm certain all you need to do is remove mdadm. (I did it and it got rid of that pesky message for me)

Alluding to this guide http://www.ubuntugeek.com/how-to-fix-broken-ubuntu-feisty-fawn.html

1. Boot from the liveCD.
2. sudo mkdir /media/feisty

3. Mount your drive using the following command
sudo mount /dev/sda2 /media/feisty

(replace sda2 with the location of the drive, try hda2, hdb1 and things)

4. sudo chroot /media/feisty su

5. apt-get remove mdadm

Restart without the liveCD. Should start successfully.

If you chose the wrong drive at stage 3, chroot will probably fail. in that case run
3b sudo umount /media/feisty
and
3c sudo mount /dev/*hda2* (or whichever drive) /media/feisty

I try What you told me to do and the mdadm error disappears. But now when I try to boot it stays in:
Loading root files system.
I got this message foe the kernel 2.6.17 and 2.6.20. The only one that works all the time was the 2.6.15.
If you can help fixing that problem it will be great because I can't use the nvidia drivers on feisty with the kernel version 2.6.15

Thx for all the help.

Ok, I try to reinstall the mdadm. I think the problem is there: when it's installing it's supposedly generate the file /etc/mdadm/mdadm.conf.
But the problem is that during the installation this is what it told me:
...
update-iniframfs: Generating /boot/initrd.img-2.6.20-15-386
W: mdadm: unchecked configuration file: /etc/mdadm/mdadm.conf
W: mdadm: please read /usr/share/doc/mdadm/README.upgrading-2.5.3.gz .
W: mdadm: no arrays defined in configuration file.
W: mdadm: falling back to emergency procedure in initramfs.
...
And it's shows the same message with all the kernels. But the only one wich works is the 2.6.15. Can someone told me how I can fix this error in the creation of the mdadm.conf file?
Thanks, Nestor

BTW I almost forget. The only thing wich appears on my mdadm.conf file is:

DEVICE partitions
MAILADDR root

I hope this will help. Thanks

Nestor,
    Can I make a wild guess here, and assume that this system was upgraded from Edgy, and after the first reboot is when the issue occurred?

If so, there were some changes rolled into mdadm that require a more extensive mdadm.conf, and you need to go read this file on your system: /usr/share/doc/mdadm/README.upgrading-2.5.3.gz

You can display the file with the command "zmore /usr/share/doc/mdadm/README.upgrading-2.5.3.gz"

Hope this helps....

Nestor,

     Sorry for not reading the entire thread above. The problem with your mdadm.conf is that it does not have any UUIDs for the array(s) in it. Try this:

1) change to the mdadm directory with 'cd /etc/mdadm'
2) back up your mdadmin.conf by running 'cp mdadm.conf mdadm.conf.orig'
3) run '/usr/share/mdadm/mkconf > mdadm.conf' to generate a new mdadm.conf,
4) run 'rm -f /var/lib/mdadm/CONF-UNCHECKED' to confirm that you have check the mdadm.conf
5) run 'update-initramfs -u -k all' to get the new mdadm.conf into your kernal initrds, so the kernal knows where to look for the arrays on boot.

If that doesn't work, edit the mdadm.conf, and change the "DEVICE" line so that instead of saying "partitions", it lists the explicit devices that comprise the array (e.g. "DEVICE /dev/hda2 /dev/hdb2") , and run step 5 again.

This question was expired because it remained in the 'Needs information' state without activity for the last 15 days.