Boot Failed: wrong # of devices in RAID set

[francis9]

Hi,

I installed Fedora 12 and performed the normal updates. Now I can't reboot and get the following console error message.

ERROR: via: wrong # of devices in RAID set "via_cbcff jdief" [1/2] on /dev/sda
ERROR: removing inconsistent RAID set "via_cbcff jdief"
ERROR: no RAID set found
No root device found
Boot has failed, sleeping forever.

Rebooting multiple times does not help as in other posts.

Any suggestions?

francis9

[steelaworkn]

How did you set up RAID?

[francis9]

I mostly took the default Fedora 12 installation. While I generally know what RAID is, I wouldn't know how to set up or configure it, and am not even aware of what the default set-up would be. The only thing I did special on the installation is set up a 2nd disk partition for /home (on the one internal drive that I have). Any ideas what happened?
Thanks
-francis9

[scott32746]

Hello,
Boot up into rescue mode
then post output from
# fdisk -l
# cat /proc/mdstat
# mdadm --detail /dev/md0
# cat /etc/mdadm.conf
# cat /etc/grub.conf

[JEO]

Press a key early in the boot sequence to pull up the grub boot menu, add the keyword "nodmraid" to the kernel command line and see if it boots. You may have unwanted old raid metadata on one of your drives. The dmraid command can remove old metadata too.

[francis9]

Hi,
I used a fedora 12 live CD as my "rescue" disk and tried the utilities suggest by scott32746 to get some info from the problem hard disk, but they did not give any useful information. I don't think that they could see the disk, so they had no information to report. Any other suggestions on how I can enable the use these utilities?

Also, what key do I use to access the grub boot menu during startup as suggested by JEO? All I know is that the delete key takes me to the bios menu, but don't see how to access a command line.

Thanks
-francis9


Boot up into rescue mode
then post output from
# fdisk -l
# cat /proc/mdstat
# mdadm --detail /dev/md0
# cat /etc/mdadm.conf
# cat /etc/grub.conf

---------- Post added at 05:13 PM CDT ---------- Previous post was at 05:13 PM CDT ----------

Hi,
I used a fedora 12 live CD as my "rescue" disk and tried the utilities suggest by scott32746 to get some info from the problem hard disk, but they did not give any useful information. I don't think that they could see the disk, so they had no information to report. Any other suggestions on how I can enable the use these utilities?

Also, what key do I use to access the grub boot menu during startup as suggested by JEO? All I know is that the delete key takes me to the bios menu, but don't see how to access a command line.

Thanks
-francis9


Boot up into rescue mode
then post output from
# fdisk -l
# cat /proc/mdstat
# mdadm --detail /dev/md0
# cat /etc/mdadm.conf
# cat /etc/grub.conf

[scott32746]

If you are not getting info then I would think the only thing you can do is reinstall. I would read up on software and LVM if you are using it.
I would download the full DVD and install from that.

[JEO]

If the utilities can't see the disk that probably means that you have old raid metadata on the drives. The partitions on the source drives for a dmraid are hidden at boot time, unless you add the nodmraid option.

Since hitting a key to enter the grub boot menu is not working for you, boot the livecd but at the boot menu for the cd press tab and there you can add the nodmraid keyword. Your hard disk should then be visible after the boot completes and you can modify /boot/grub/grub.conf on it.

[steelaworkn]

When you set up RAID in Fedora, it has to be done with a little planning. So, figure out which drives you want to be in your raid array. They are usually identified by their branding and serial numbers. Another thing is make sure you are not using a 3rd party raid software like Intel Matrix Storage manager. If you are, make sure you release the disks on boot-up. So, you will also need to know which RAID you will want to use.

Install the live cd and go to set-up your system. The key partitions are /boot, /home, /swap, and /root. Take a look at the Fedora Documentation on RAID (http://docs.fedoraproject.org/ ) Fedora does not want to install the /boot partition on the RAID array. But you can trick it by partitioning out a disk just 300 meg to 1000meg for the /boot. Then take the remaining amount of that disk and select it to join your other disks to be configured for RAID.

Also, take a look at my thread ( http://forums.fedoraforum.org/showthread.php?t=240082 ) . It is not that informative but gives a little idea of what can be done.

Good luck.

[francis9]

I was able to add nodmraid to the fedora live CD kernel command line, but still unable to see the disk. There was no /boot/grub/grub.conf file to modify. Anything else to try to see how it got configured?

I'm not opposed to a reinstall, but don't know what I'd do different to avoid getting in this same situation.

thanks
francis9

[steelaworkn]

I was able to add nodmraid to the fedora live CD kernel command line, but still unable to see the disk. There was no /boot/grub/grub.conf file to modify. Anything else to try to see how it got configured?

I'm not opposed to a reinstall, but don't know what I'd do different to avoid getting in this same situation.

thanks
francis9

I'd say, if I can speak freely, that if you don't really know how you got RAID on your system, then, really, there's no other choice but to start over. You need to read the official Fedora documentation. If you use LVM to partition your RAID array, you will be fine.

[JEO]

That sure is odd. The only guess I have is that you initially installed and then had a drive fail. Or when you updated, the newer kernel does not work with one of your drives. Either way, the grub default timeout being set to 0 seconds is horrible for these situations. If you do reinstall, the very first thing to do is set a nonzero timeout in grub.conf.

[francis9]

Hi,
I was finally able to access the grub menu during boot off of the disk and add the "nodmraid" option. Then I could boot up normally without the RAID error! I did notice that I had the option of two different kernel versions in grub, and don't understand why since I should only have one OS on the disk.
Next, I was able to run the utilities and get some info on the disk that follows below. So I'm more optimistic that I can avoid a re-install, but recognize that may still be in the cards. Is there anything in this output that could help?
Finally, I added "nodmraid" to the /boot/grub/grub.conf file and am now able to boot up normally off the disk.
Based on the output below, is there anything else I need to clean up this problem?
Thanks
-francis9

=============begin output==================
[root@aksel charles]# df -k
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/mapper/vg_aksel-lv_root
75828328 4669564 67306888 7% /
tmpfs 383424 272 383152 1% /dev/shm
/dev/sda1 198337 39484 148613 21% /boot
/dev/mapper/vg_aksel-lv_home
403173232 253792 382439440 1% /home
[root@aksel charles]# fdisk -l

Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x44861b19

Device Boot Start End Blocks Id System
/dev/sda1 * 1 26 204800 83 Linux
Partition 1 does not end on cylinder boundary.
/dev/sda2 26 60801 488178728+ 8e Linux LVM
Partition 2 does not end on cylinder boundary.

Disk /dev/dm-0: 419.4 GB, 419430400000 bytes
255 heads, 63 sectors/track, 50992 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

Disk /dev/dm-0 doesn't contain a valid partition table

Disk /dev/dm-1: 78.9 GB, 78886469632 bytes
255 heads, 63 sectors/track, 9590 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

Disk /dev/dm-1 doesn't contain a valid partition table

Disk /dev/dm-2: 1577 MB, 1577058304 bytes
255 heads, 63 sectors/track, 191 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00000000

Disk /dev/dm-2 doesn't contain a valid partition table
[root@aksel charles]# cat /proc/mdstat
Personalities :
unused devices: <none>
[root@aksel charles]# mdadm --detail /dev/md0
mdadm: cannot open /dev/md0: No such file or directory
[root@aksel charles]# cat /etc/mdadm.conf
cat: /etc/mdadm.conf: No such file or directory
[root@aksel charles]# cat /etc/grub.conf
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/mapper/vg_aksel-lv_root
# initrd /initrd-[generic-]version.img
#boot=/dev/sda
default=0
timeout=0
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title Fedora (2.6.32.9-70.fc12.i686.PAE)
root (hd0,0)
kernel /vmlinuz-2.6.32.9-70.fc12.i686.PAE ro root=/dev/mapper/vg_aksel-lv_root LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=us rhgb quiet
initrd /initramfs-2.6.32.9-70.fc12.i686.PAE.img
title Fedora (2.6.31.5-127.fc12.i686.PAE)
root (hd0,0)
kernel /vmlinuz-2.6.31.5-127.fc12.i686.PAE ro root=/dev/mapper/vg_aksel-lv_root LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=us rhgb quiet
initrd /initramfs-2.6.31.5-127.fc12.i686.PAE.img
[root@aksel charles]#

=======================end output=============================

[JEO]

The second kernel was added at the time you performed the "normal updates" as in your first post in this thread. Normally the last 3 kernels are saved and so if you have a problem with a new kernel you can revert to a previous one.

If you want to be able to boot without the nodmraid option then you can use the dmraid command to erase the unwanted metadata.

su -
dmraid -r

Should list what drive devices have metadata on it. It looks like you have via format metadata so the getting the /dev filename from the above, to erase the small metadata sectors:


dmraid -rE -f via /dev/sda (but make sure the device name matches the one above)

[francis9]

Hi,
I used dmraid as suggested to identify the device, then again to erase the metadata. dmraid -r a second time indicated no raid device, so this looked promising. Then, I removed the nodmraid option from grub.conf and the machine boots normally now without the error.
I'm totally amazed at the response I got from my first thread here and would have never been able to solve this problem on my own. Thanks for the great support.
-francis9