[SOLVED] unable determine major/minor number of root device 'dev/sda6'

nomorewindows · 2012-03-03 14:52:39

It's probably time to run your mkinitcpio -p linux.

miky76 · 2012-03-03 16:01:04

I think I do not have that file

cd /sbin 
mkinitcpio -p linux

the result is command not found

ls -l mkinitcpio

result:
lrwxrwxrwx 1 root root 19 Jan 13 01:58 mkinitcpio -> /usr/bin/mkinitcpio

ls -l /usr/bin/mkinitcpio

result: No such file or directory

I'm sure I didn't deleted it on pourpose, do you know how install it again?

nomorewindows · 2012-03-03 16:54:37

Does your pacman still work? pacman -S mkinitcpio
Hopefully your boot OS is close enough to what you had so that when mkinitcpio runs, it will be pulling the modules from the boot OS. Otherwise, just pacman -S linux and it will have to rebuild it all over again, and may install mkinitcpio with it. You could do a pacman -Q mkinitcpio to see if it is even installed, and if it is pacman -Qk mkinitcpio to see what files might be missing.

Last edited by nomorewindows (2012-03-03 16:56:37)

miky76 · 2012-03-04 23:11:37

those the results:

pacman -S mkinitcpio

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
error: could not opn file /var/lib/pacman/sync/core.db: Failed to open '/var/lib/pacman/sync/core.db'
error: could not opn file /var/lib/pacman/sync/extra.db: Failed to open '/var/lib/pacman/sync/extra.db'
error: could not opn file /var/lib/pacman/sync/community.db: Failed to open '/var/lib/pacman/sync/community.db'

error: target not found: mkinitcpio

pacman -Q mkinitcpio

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
mkinitcpio 0.7.2-1

pacman -Qk mkinitcpio

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
warning: mkinitcpio: /bin/lsinitcpio (No such file or directory)
warning: mkinitcpio: /lib/initcpio/install/dsdt (No such file or directory)
warning: mkinitcpio: /usr/share/man/ (No such file or directory)
warning: mkinitcpio: /usr/share/man/man5/ (No such file or directory)
warning: mkinitcpio: /usr/share/man/man5/mkinitcpio.5.gz (No such file or directory)

mkinitcpio: 52 total files, 5 missing files

what this error message means?

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized

I googled, but with no luck

Last edited by miky76 (2012-03-04 23:12:30)

miky76 · 2012-03-04 23:56:15

I've checked the pacman.conf and in "PGP signature checking"

SigLevel = Never

miky76 · 2012-03-06 20:57:24

maybe something useful, I've found in my /var/cache/pacman/pkg/ those:

mkinitcpio-0.8.0-3-any.pkg.tar.xz
mkinitcpio-0.8.2-1-any.pkg.tar.gz

linux-3.1.5-1-x86_64.pkg.tar.xz
linux-3.2.5-1-x86_64.pkg.tar.xz
linux-api-headers-3.1.4-1-x86_64.pkg.tar.xz
linux-api-headers-3.1.6-1-x86_64.pkg.tar.xz
linux-firmware-20111101-any.pkg.tar.xz
linux-firmware-20120205-any.pkg.tar.xz

can it be useful?

there are also many many other packages there

nomorewindows · 2012-03-08 13:30:07

It sounds like pacman isn't upgraded, and the possibility that your /var isn't mounted. It's a little hard to tell what's happening right ATM.

miky76 · 2012-03-08 13:39:38

mmm I'm seriously thinking to create new partitions in another disk and install from zero a new Arch, the annoying thing is that, after this, I have to reinstall all over again, but seem faster

In the current broken OS, I used to have 2 disks in RAID 0, now, of course, I cannot access to the datas, do you think I can configure the new OS, easily for see the datas in the RAID again?

nomorewindows · 2012-03-08 14:12:50

It sounds like one of the disks in your RAID array has probably failed. In the case of RAID 0, you'll have to figure out which is working and continue using it. Practically RAID 0 is not RAID at all. If the convention for RAID is still the same, it is referred to as /dev/mdxxxxxxx. Do something for me, do fdisk -l again, your output didn't display everything we needed to see (preferrably outside of chroot and leave off /dev/sda). Although installing to a new disk won't destroy anything on your existing disk, I wouldn't suggest going to that trouble until we know for sure that your disks are bad.

miky76 · 2012-03-08 21:02:45

Hi,
first of all, I gave you a wrong information, my disks are in raid 1
and you can see them in the fstab as

/dev/data/databin    /data ext4  defaults 0 1

the OS the one is broken, is not in this disks, but start in a 4GB USB key.

for what I see, I do not have any hint that one of the 2 disks is not working anymore, but maybe it can be.

so, what I can do is:

- keep trying to fix this OS
- or buy a new (real) hard disk and use this for install a brand new ArchLinux OS, reinstall everything and try to mount the raid again in this new one (actually I care only of the datas inside)

what do you suggest?

in the meantime you think, i've rebooted the linux box (electricity went off) so I have to start everything again (but maybe can be a good exercise for see if I mistook something)

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

I've rebooted the system with the booting USB in

loadkeys it
mkdir /mnt/oldarch

cd /mnt/oldarch
mount -t proc proc proc/
mount -t sysfs sys sys/
mount -o bind /dev dev/
mount -o bind /boot boot/
mount -o bind /var var/
mount -o bind /usr usr/

mount /dev/sdd1 boot/

cp -L /etc/resolv.conf /mnt/oldarch/etc/resolv.conf

chroot . /bin/bash

vi etc/fstab

/dev/data/databin    /data ext4  defaults 0 1
/dev/sda1      /boot    ext2  defaults 0 1
/dev/sda6     /      ext4   defaults 0 1
/dev/sda5    swap  swap     defaults 0 0
/dev/sdb1   /var/cache/pacman/pkg   ext4     defaults 0 0
tmpfs   /tmp   tmpfs  nodev,nosuid  0  0

/etc/rc.d/network restart

ping 173.194.35.177  

the ping is working so the network is on

/etc/rc.d/sshd restart

the message say FAIL

Starting Secure Shell Daemon  [BUSY]
ssh-keygen: illegal option -- A
usage: ssh-keygen [options]
Options:
  here is written a long list of options

I plug another 4GB USB key that is seen as sde

mount /dev/sde1 /var/cache/pacman/pkg

ls -l /dev/sde1 /var/cache/pacman/pkg

I see inside all the list of packages I've dowloaded when the system crashed

pacman -S mkinitcpio

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
error: could not opn file /var/lib/pacman/sync/core.db: Failed to open '/var/lib/pacman/sync/core.db'
error: could not opn file /var/lib/pacman/sync/extra.db: Failed to open '/var/lib/pacman/sync/extra.db'
error: could not opn file /var/lib/pacman/sync/community.db: Failed to open '/var/lib/pacman/sync/community.db'

error: target not found: mkinitcpio

pacman -S pacman

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
error: could not opn file /var/lib/pacman/sync/core.db: Failed to open '/var/lib/pacman/sync/core.db'
error: could not opn file /var/lib/pacman/sync/extra.db: Failed to open '/var/lib/pacman/sync/extra.db'
error: could not opn file /var/lib/pacman/sync/community.db: Failed to open '/var/lib/pacman/sync/community.db'

error: target not found: pacman

nomorewindows · 2012-03-09 01:04:47

miky76 wrote:

Hi,
first of all, I gave you a wrong information, my disks are in raid 1
and you can see them in the fstab as
/dev/data/databin    /data ext4  defaults 0 1
the OS the one is broken, is not in this disks, but start in a 4GB USB key.
for what I see, I do not have any hint that one of the 2 disks is not working anymore, but maybe it can be.
so, what I can do is:
- keep trying to fix this OS
- or buy a new (real) hard disk and use this for install a brand new ArchLinux OS, reinstall everything and try to mount the raid again in this new one (actually I care only of the datas inside)
what do you suggest?
in the meantime you think, i've rebooted the linux box (electricity went off) so I have to start everything again (but maybe can be a good exercise for see if I mistook something)
-----------------------------------------------------------------------------------
I've rebooted the system with the booting USB in
loadkeys it
mkdir /mnt/oldarch

cd /mnt/oldarch
mount -t proc proc proc/
mount -t sysfs sys sys/
mount -o bind /dev dev/
mount -o bind /boot boot/
mount -o bind /var var/
mount -o bind /usr usr/

mount /dev/sdd1 boot/

cp -L /etc/resolv.conf /mnt/oldarch/etc/resolv.conf

chroot . /bin/bash

vi etc/fstab

/dev/data/databin    /data ext4  defaults 0 1
/dev/sda1      /boot    ext2  defaults 0 1
/dev/sda6     /      ext4   defaults 0 1
/dev/sda5    swap  swap     defaults 0 0
/dev/sdb1   /var/cache/pacman/pkg   ext4     defaults 0 0
tmpfs   /tmp   tmpfs  nodev,nosuid  0  0

/etc/rc.d/network restart

ping 173.194.35.177  

the ping is working so the network is on

/etc/rc.d/sshd restart
the message say FAIL
Starting Secure Shell Daemon  [BUSY]
ssh-keygen: illegal option -- A
usage: ssh-keygen [options]
Options:
  here is written a long list of options
I plug another 4GB USB key that is seen as sde
mount /dev/sde1 /var/cache/pacman/pkg

ls -l /dev/sde1 /var/cache/pacman/pkg
I see inside all the list of packages I've dowloaded when the system crashed
pacman -S mkinitcpio

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
error: could not opn file /var/lib/pacman/sync/core.db: Failed to open '/var/lib/pacman/sync/core.db'
error: could not opn file /var/lib/pacman/sync/extra.db: Failed to open '/var/lib/pacman/sync/extra.db'
error: could not opn file /var/lib/pacman/sync/community.db: Failed to open '/var/lib/pacman/sync/community.db'

error: target not found: mkinitcpio
pacman -S pacman

warning: config file /etc/pacman.con, line 50: directive 'SigLevel' in section 'options' not recognized
error: could not opn file /var/lib/pacman/sync/core.db: Failed to open '/var/lib/pacman/sync/core.db'
error: could not opn file /var/lib/pacman/sync/extra.db: Failed to open '/var/lib/pacman/sync/extra.db'
error: could not opn file /var/lib/pacman/sync/community.db: Failed to open '/var/lib/pacman/sync/community.db'

error: target not found: pacman

All you've done with these lines is reproduce whatever errors in the process or scheme of things might be lurking.

These three lines aren't right:
mount -o bind /boot boot/
mount -o bind /var var/
mount -o bind /usr usr/
This mounts your boot media over something -- I'm not sure what.

Your ssh session should be started outside of the chroot session (and may require copying the configuration from the original into the boot OS to get it to work)
Your /dev/sde1 I believe is your /dev/sdb1 according to your /etc/fstab, not your 4GB key.
Your crashed OS should have a pacman database, which seems to me to be your boot OS.

Your boot mounting for your crashed OS should be under the same place as where you are mounting your other partitions, so that /dev/sdd6 /mnt/arch (or wherever) is your first, and /dev/sdd1 /mnt/arch/boot is where the boot is mounted. And then your second drive should mount /dev/sde1 /mnt/arch/var/cache/pacman/pkg. Then add the other three lines that are correct (proc, sys, dev), which enable you to rebuild your kernel. Then you can chroot /mnt/arch (or wherever). From there you might be able to mount the line with /dev/data/databin, it should just go on it's own if everything is correct just by typing 'mount /data'

miky76 · 2012-03-09 14:39:38

nomorewindows wrote:

All you've done with these lines is reproduce whatever errors in the process or scheme of things might be lurking.
These three lines aren't right:
mount -o bind /boot boot/
mount -o bind /var var/
mount -o bind /usr usr/
This mounts your boot media over something -- I'm not sure what.

what is the correct way to mount them?

nomorewindows wrote:

Your ssh session should be started outside of the chroot session (and may require copying the configuration from the original into the boot OS to get it to work)

ok, i'll try to do this and I'll let you know

nomorewindows wrote:

Your /dev/sde1 I believe is your /dev/sdb1 according to your /etc/fstab, not your 4GB key.
Your crashed OS should have a pacman database, which seems to me to be your boot OS.

yes the sdb1 is now the sde1 (it's a second 4gb usb key that I use just for the pacman downloaded packages)

nomorewindows wrote:

Your boot mounting for your crashed OS should be under the same place as where you are mounting your other partitions, so that /dev/sdd6 /mnt/arch (or wherever) is your first, and /dev/sdd1 /mnt/arch/boot is where the boot is mounted. And then your second drive should mount /dev/sde1 /mnt/arch/var/cache/pacman/pkg. Then add the other three lines that are correct (proc, sys, dev), which enable you to rebuild your kernel. Then you can chroot /mnt/arch (or wherever). From there you might be able to mount the line with /dev/data/databin, it should just go on it's own if everything is correct just by typing 'mount /data'

nomorewindows · 2012-03-09 15:38:51

These three lines aren't right:
mount -o bind /boot boot/
mount -o bind /var var/
mount -o bind /usr usr/

Just don't mount these when you reboot or go back and unmount them. This probably is the reason you are getting errors with your /var and /usr directories when using pacman, but it also is affecting other directories not spawned by this, so this is still somewhat unexplained.

I don't know about storing pacman packages on a memory stick, it will fill up quickly and I just store them on a bigger disk somewhere else when they get older. This is of course personal preference. Of course, you realize you can reuse the packages on other installations, given you keep the architectures straight.

Last edited by nomorewindows (2012-03-09 15:45:10)

miky76 · 2012-03-10 13:11:31

Let start again

I've rebooted the system with the booting USB in

loadkeys it
mkdir /mnt/oldarch

mount /dev/sdd6 /mnt/oldarch

cd /mnt/oldarch
mount -t proc proc proc/
mount -t sysfs sys sys/
mount -o bind /dev dev/

mount /dev/sdd1 /mnt/oldarch/boot

cp -L /etc/resolv.conf /mnt/oldarch/etc/resolv.conf

cp -L /etc/ssh/sshd_config /mnt/oldarch/etc/ssh/sshd_config

/etc/rc.d/sshd restart    [b]DONE[/b]

mount /dev/sdc1 /mnt/oldarch/var/cache/pacman/pkg   [b]#it used to be sde1 now is sdc1[/b]

cd /mnt/oldarch

chroot . /bin/bash

vi etc/fstab

/dev/data/databin    /data ext4  defaults 0 1
/dev/sda1      /boot    ext2  defaults 0 1
/dev/sda6     /      ext4   defaults 0 1
/dev/sda5    swap  swap     defaults 0 0
/dev/sdb1   /var/cache/pacman/pkg   ext4     defaults 0 0
tmpfs   /tmp   tmpfs  nodev,nosuid  0  0

/etc/rc.d/network restart

ping 173.194.35.177  

the ping is working so the network is on

trying from another machine to connect through SSH, I get "connection refused", last time I could insert login and password

mkinitcpio -p linux

It's working!!!

it's generating Initramfs-linux.img and initramfs.linux-fallback.img

it finished successfully, and now?

Last edited by miky76 (2012-03-10 13:12:16)

nomorewindows · 2012-03-10 14:39:29

You copied the ssh config from your boot OS to your working configuration! If you were using ssh to get your machine working again, it was supposed to be copied the other way around (leave your working ssh config alone and just copy it temporarily to the boot OS), but other than that looks good!

So if that was the only problem, you can just back out of chroot and 'init 6' or control-alt-delete and let it unmount everything for you, and we'll see if that was all it needed.

miky76 · 2012-03-10 15:25:54

I've removed all the USB keys and left only the one with the fixed OS

the boot now fail because the USB key with the OS is not sda anymore but is sdc

- how can I make the OS start from sdc now? just changing the fstab?
- how can I do that everytime I reboot it will not change the name of the usb? sda, sdb, sdc, etc...? with UUID or by-label?

thanks

nomorewindows · 2012-03-10 18:01:59

miky76 wrote:

I've removed all the USB keys and left only the one with the fixed OS
the boot now fail because the USB key with the OS is not sda anymore but is sdc
- how can I make the OS start from sdc now? just changing the fstab?
- how can I do that everytime I reboot it will not change the name of the usb? sda, sdb, sdc, etc...? with UUID or by-label?
thanks

Let me understand, all the USB keys are removed except one? It should be booting by itself from the hard drive, and not by USB. USB is a temporary arrangement, unless your /var/cache/pacman/pkg is messing up the order in which the BIOS thinks it is occurring. Plugging and unplugging your USB key device is a point of possible failure in the future. I think that there is also a command to change the order that grub sees the devices, to be able to flip flop (hd0) with (hd1); the OS with the USB key.

Everything will need to be consistent.

*Probably all you need to do is remove the USB key when grub comes up, and then after it starts booting, reinsert the USB key.*

If you are using grub boot loader and you need to make a change you can do it from grub, by interactively editing the kernel line to reflect /dev/sdc or adding the flip flop (this will only affect this booting session). However, if the kernel doesn't recognize the root filesystem as being on /dev/sda when it hands off to init, then it will kernel panic or if it does mount it will mount read-only. In which case, you would have to temporarily go back in and change /etc/fstab to make the lines /dev/sdc, and reboot, then when it boots sucessfully, go back and change it back, and make sure your /boot/grub/menu.lst has /dev/sda, and mkinitcpio -p linux along with grub-install /dev/sda.

Last edited by nomorewindows (2012-03-10 18:03:39)

miky76 · 2012-03-12 20:26:08

sorry for my late answer, bu tI couldn't work much on it in this days as I was away

My "normal" booting partition is a USB, because I do not have an HD for the OS, I will buy it in few days (tomorrow or the day after), and migrate the OS on it. This is why I'm using so many USB keys...

With your help I made it works!!

rebooting, the system was recognizing the fixed OS partition as sdc, so I've changed, right now, the menu.lst and the fstab, it rebooted and it's working now.

few things are still not working, but I can work on it, for example kde, sshd, samba and few other things are active but not really working

So i'm trying to upgrade them again using pacman (I have to upgrade also pacman) finger crossed!!

nomorewindows thanks for your help and your patience

for sure I will disturb you again for new mess I will do with ArchLinux

Arch Linux

#26 2012-03-03 14:52:39

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#27 2012-03-03 16:01:04

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#28 2012-03-03 16:54:37

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#29 2012-03-04 23:11:37

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#30 2012-03-04 23:56:15

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#31 2012-03-06 20:57:24

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#32 2012-03-08 13:30:07

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#33 2012-03-08 13:39:38

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#34 2012-03-08 14:12:50

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#35 2012-03-08 21:02:45

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#36 2012-03-09 01:04:47

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#37 2012-03-09 14:39:38

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#38 2012-03-09 15:38:51

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#39 2012-03-10 13:11:31

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#40 2012-03-10 14:39:29

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#41 2012-03-10 15:25:54

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#42 2012-03-10 18:01:59

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

#43 2012-03-12 20:26:08

Re: [SOLVED] unable determine major/minor number of root device 'dev/sda6'

Board footer