Boot failing: No init found

[cstad]

Hi, my stretch install is failing to boot and I'm not sure how to interpret the logs:

Code: Select all

JBD2: Unrecognised features on journal
EXT4-fs (sda6): error loading journal
mount: mounting /dev/sda6 on /root failed: Invalid argument
done.
Begin: Running /scripts/local-bottom ... done
Begin: Running /scripts/init-bottom ... mount: mounting /dev on /root/dev failed: No such file or directory
mount: mounting /dev on /root/dev failed: No such file or directory
done.
mount: mounting /run on /root/run failed: No such file or directory
run-init: current directory on the same filesystem as the root: error 0
Target filesystem doesn't have requested /sbin/init.
run-init: current directory on the same filesystem as the root: error 0
run-init: current directory on the same filesystem as the root: error 0
run-init: current directory on the same filesystem as the root: error 0
run-init: current directory on the same filesystem as the root: error 0
run-init: current directory on the same filesystem as the root: error 0
No init found. Try passing init= bootarg.

And then it gives me an "(initramfs)" prompt. I tried running "fsck /dev/sda6" from there but it says it is clean.

I've looked at a bunch of posts related to "No init found" but nothing exactly like this, and I'm not sure if that is just a symptom of the journal issues. Before rebooting I was trying to install virtualbox, so maybe that triggered this somehow?

I don't have a good understanding of the boot process, and I'm scared of losing data, so any help or suggestions would be very deeply appreciated. Thank you.

[p.H]

I've looked at a bunch of posts related to "No init found"

You're looking to the wrong direction. No init found is just a consequence of the failure to mount the root filesystem (which contains init). The important information is :

Code: Select all

JBD2: Unrecognised features on journal
EXT4-fs (sda6): error loading journal

Did you try fsck with -f ?

[cstad]

Thanks, very useful to know that is the root problem. I ran "fsck -f /dev/sda6" but it didn't find anything to fix. I found this article which had the same journal issue, but fsck fixed it for them.

[bw123]

Code: Select all

mount: mounting /dev/sda6 on /root failed: Invalid argument

What's up with invalid argument? why would you want to mount a volume on /root anyway? what's in fstab?

[cstad]

The /etc/fstab on initramfs is empty. How do I read the "real" fstab?

why would you want to mount a volume on /root anyway?

I don't know, where would it normally be mounted? There's an env var "rootmnt=/root", is that related?

Thanks!

[bw123]

You might need to explain a little better how your system got into this. It's possible nobody here will know for sure what you did based on the error message alone, or how to fix it.

Did you change anything in the grub config? Did you alter the settings for initramfs somehow?

Since you said you're scared to lose any data, I'm sure you have a backup made of anything important? If not, that would be the first thing to do. From a live system probably would be the best way.

[cstad]

That makes sense, and I really appreciate the help. I did not intentionally change anything before this happened. The only thing out of the ordinary was that I was installing virtualbox, so I will try to remember exactly what I ran while doing that. Yes, I need to backup some things first though.

[p.H]

why would you want to mount a volume on /root anyway? what's in fstab?

The error happens during execution of the initramfs, before the final root filesystem is mounted. The main purpose of the initramfs is to mount the final root filesystem (which seems to be in /dev/sda6). It tries to mount it on /root, then switches to it as the new /. There is no fstab yet as it is in the root filesystem. The initramfs gets the root filesystem specification from the root= parameter in the kernel command line passed by the boot loader.

I once encountered "unrecognized features" errors when trying to mount an ext4 filesystem created with Stretch's mke2fs default options on a Wheezy system with a 3.2 kernel. But IIRC they were not related to the journal but to some other new features such as metadata checksumming.

[cstad]

Thanks p.H, any suggestions for how to figure out what the "unrecognised features" are?

I was able to use a live USB to backup my important data and pull some information which will hopefully be useful.

It sounds like this is probably not the issue, but here is /etc/fstab:

Code: Select all

# / was on /dev/sda6 during installation
UUID=3eda7621-8a70-4c32-96e2-049fb3d8102f / ext4 errors=remount-ro 0 1
# /boot/efi was on /dev/sda1 during installation
UUID=E02A-F4E4 /boot/efi vfat umask=0077 0 1
# swap was on /dev/sda7 during installation
UUID=29762f86-b82a-472c-ae0a-66f0521b29ec none swap sw 0 0
/dev/sr0 /media/cdrom0 udf,iso9660 user,noauto 0 0

And here is what I ran before rebooting:

Code: Select all

wget -q https://www.virtualbox.org/download/oracle_vbox_2016.asc -O- | sudo apt-key add -
sudo apt update
sudo apt install virtualbox-5.2
sudo apt install ./vagrant_2.1.5_x86_64.deb 
vagrant box add [...]
vagrant init [...]
vagrant up
VBoxManage --version
./etc/init.d/vboxdrv setup
sudo /etc/init.d/vboxdrv setup
sudo apt-get install dkms build-essential linux-headers-`uname -r`
sudo /sbin/vboxconfig
sudo rcvboxdrv setup
dpkg-reconfigure virtualbox-dkms
modprobe vboxdrv

[p.H]

any suggestions for how to figure out what the "unrecognised features" are?

You can try

Code: Select all

tune2fs -l /dev/sda6

However there should not be any "normal" unrecognized features if the filesystem was created with the same version of Debian which is installed on it, and has been running fine until now. So I suspect some metadata corruption.

Now you have backed up the data, you could try to remove and re-create the journal with

Code: Select all

tune2fs -O ^has_journal /dev/sda6
tune2fs -O has_journal /dev/sda6

[cstad]

I tried re-creating the journal and that did get me farther then before, but then there were a ton of messages like this:

Code: Select all

EXT4-fs warning (device sda6): dx_probe:472: Corrupt dir inode 7602194, running e2fsck is recommended.
EXT4-fs warning (device sda6): dx_probe:398: dx entry: limit != root limit

And a bunch of things that say "[FAILED]" but scroll by too fast to read.

I then got a terminal from which I can log in, but no GUI, and running anything triggered more of the above error messages. fsck still says it's clean though.

Unless there are any more suggestions I'm planning on re-installing. Thanks again!

[p.H]

Did you try again running fsck with -f while the filesystem was not mounted ?

[debiman]

And here is what I ran before rebooting:

Code: Select all

wget -q https://www.virtualbox.org/download/oracle_vbox_2016.asc -O- | sudo apt-key add -
sudo apt update
sudo apt install virtualbox-5.2
sudo apt install ./vagrant_2.1.5_x86_64.deb 
vagrant box add [...]
vagrant init [...]
vagrant up
VBoxManage --version
./etc/init.d/vboxdrv setup
sudo /etc/init.d/vboxdrv setup
sudo apt-get install dkms build-essential linux-headers-`uname -r`
sudo /sbin/vboxconfig
sudo rcvboxdrv setup
dpkg-reconfigure virtualbox-dkms
modprobe vboxdrv

crap, is this the proper way to install & set up virtualbox on debian?
to me it looks either outdated or way too low-level.
where did you get this from?
the virtualbox wiki does not have anything below apt install virtualbox-5.2.

[cstad]

Did you try again running fsck with -f while the filesystem was not mounted ?

Yes, from Debian on a live USB I ran "sudo fsck -f /dev/sda6" and it didn't report any issues.

crap, is this the proper way to install & set up virtualbox on debian?

Basically after the apt install it wasn't working so I then started googling and running things from stackoverflow. Maybe not the best strategy. The virtualbox manual does say you may need to install linux kernel headers and then run "rcvboxdrv setup" (https://www.virtualbox.org/manual/ch02.html#idm936), but I'm not sure exactly why I ran the other things.