• We had to restore from a backup today after a failed software update. Backup was from 0000 EDT and restored it at 0800 EDT so we lost about 8hrs. Today is 07/20/2024. More info here.

[SOLVED] Problems (re)installing Ubuntu 20.04 on Laptop with UEFI SSD - get "ubi partman failed with exit code 10"

Kibi

New Member
Joined
Jul 20, 2022
Messages
12
Reaction score
6
Credits
122
Hi there

I saw this question here and Condoman suggested I open a new topic - sorry if I've done it wrong.

I'm trying to install Ubuntu 20.04 on my laptop and getting this error: "ubi partman failed with exit code 10".

To be honest it's a reinstall, because I had everything working and then messed it up by deleting a partition in the Disks app - my fault.

Anyway, I don't care too much about saving data I just want a system which will boot. The laptop is a little unusual in that it has a 256GB SSD and a 1TB HDD - I want to boot off the SSD to make the boot a bit faster.

The SSD is UEFI (I have to admit I don't really understand that) - but I know I had the system booting and working fine (Ubuntu only) for 2 years until I muffed it a couple of days ago.

I followed the instructions in that question I references to use GParted on the disk and made a small EFI partition (it's actually 512Mb), bootable
I noticed that the rest of the disk (ext4) is also bootable
But rerunning the installation I still get the same error - I looked at the syslog in var/log and if I understand correctly, the relevant lines are:

Code:
Jul 20 07:55:24 ubuntu ubiquity[23896]: switched to page prepare
Jul 20 07:55:29 ubuntu ubiquity: No drivers found for installation.
Jul 20 07:55:29 ubuntu ubiquity[23896]: debconffilter_done: ubi-prepare (current: ubi-prepare)
Jul 20 07:55:29 ubuntu ubiquity[23896]: Step_before = stepPrepare
Jul 20 07:55:29 ubuntu activate-dmraid: No Serial ATA RAID disks detected
Jul 20 07:55:30 ubuntu ubiquity: umount: /target: target is busy.
Jul 20 07:55:30 ubuntu ubiquity[23896]: debconffilter_done: ubi-partman (current: ubi-partman)
Jul 20 07:55:30 ubuntu ubiquity[23896]: dbfilter_handle_status: ('ubi-partman', 10)
Jul 20 07:55:36 ubuntu ubiquity[23896]: dbfilter_handle_status: response -7
Jul 20 07:55:36 ubuntu ubiquity[23896]: log-output -t ubiquity debconf-copydb configdb targetdb -p ^oem-config/ --config=Name:targetdb --config=Driver:File --config=Mode:0644 --config=Filename:/target/var/cache/debconf/config.dat
Jul 20 07:55:36 ubuntu ubiquity[23896]: log-output -t ubiquity debconf-copydb configdb targetdb -p ^keyboard-configuration/ --config=Name:targetdb --config=Driver:File --config=Mode:0644 --config=Filename:/target/var/cache/debconf/config.dat
Jul 20 07:55:37 ubuntu ubiquity[23896]: log-output -t ubiquity debconf-copydb configdb targetdb -p ^console-setup/ --config=Name:targetdb --config=Driver:File --config=Mode:0644 --config=Filename:/target/var/cache/debconf/config.dat
Jul 20 07:55:37 ubuntu ubiquity:                      USER        PID ACCESS COMMAND
Jul 20 07:55:37 ubuntu ubiquity: /target:
Jul 20 07:55:37 ubuntu ubiquity:         
Jul 20 07:55:37 ubuntu ubiquity:  root 
Jul 20 07:55:37 ubuntu ubiquity:  swap  /target/swapfile
Jul 20 07:55:37 ubuntu ubiquity:   
Jul 20 07:55:37 ubuntu ubiquity:     
Jul 20 07:55:37 ubuntu ubiquity: 
Jul 20 07:55:37 ubuntu ubiquity:  root      mount
Jul 20 07:55:37 ubuntu ubiquity: /target
Jul 20 07:55:37 ubuntu ubiquity: kernelkernel
Jul 20 07:55:37 ubuntu ubiquity[23896]: log-output -t ubiquity fuser -cv /target
Jul 20 07:55:37 ubuntu ubiquity: umount: /target: target is busy.
Jul 20 07:55:37 ubuntu ubiquity[23896]: log-output -t ubiquity umount /target
Jul 20 07:55:37 ubuntu systemd[2020]: vte-spawn-783f789f-64f5-46f5-b66d-a0514d925e37.scope: Succeeded.
Jul 20 07:55:41 ubuntu dbus-daemon[2041]: [session uid=999 pid=2041] Activating service name='org.gnome.Nautilus' requested by ':1.100' (uid=999 pid=15207 comm="/usr/bin/gnome-shell " label="unconfined")
Jul 20 07:55:41 ubuntu dbus-daemon[2041]: [session uid=999 pid=2041] Successfully activated service 'org.gnome.Nautilus'
Jul 20 07:55:41 ubuntu nautilus[26096]: Couldn't register with accessibility bus: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

Any idea what I'm doing wrong?

Thanks so much for any help
 
Last edited:


Delete all partitions on drive you want to install with Gparted from the live USB and create only one partition for whole drive format it to ext4 and reinstall Ubuntu and let the Ubuntu installer format and make the required partitions
 
Thank you @Lord Boltar - I did that. When I reboot sfter the reinstall I now get a message:

No Boot Device Found. Press any key to reboot the machine

(Pressing any key doesn't do anything)

To be clear - having formatted to a single partition of ext4, the Ubuntu install has now repartitioned to a small partition of 512Mib in fat32 which is marked boot, esp and the rest is in an ext4 partition

Any clue what to do now?

Should probably add for completeness, this is a Dell Precision 5510
Is it possible I should change any BIOS settings?
 
Last edited:
On reinstall, why not use the so-called "Experimental" option and use ZFS?
It worked great for me with a xubuntu 20.04.2 called Enchantment for about a year.
I'd recommend it on top of what @Lord Boltar suggests.
 
In the Bios make sure secure boot and fast boot are turned off
Have a look here - https://itsfoss.com/no-bootable-device-found-ubuntu/
Secure boot is off. It now boots into Grub 2.04

I tried 'ls' and looked at the various drives - none of them show me much info. just the sizes of the partitions I have. I didn't see any Fast Boot setting in the BIOS.

I followed the instructions in the link you sent - those suggest turn Secure boot on and finding the file and set up a new boot with the file /ubuntu/shimx64.efi (which I found) - however, it works no better (goes to grub)) so I have gone back to turning Secure boot off.

I guess I'm going to try @forester 's idea of Using ZFS now - I'm pretty stuck.

Any more ideas?
 
On reinstall, why not use the so-called "Experimental" option and use ZFS?
It worked great for me with a xubuntu 20.04.2 called Enchantment for about a year.
I'd recommend it on top of what @Lord Boltar suggests.
OK, tried that - unfortunately the same issue - just boots to grub now
 
The next option is to change the BIOS from UEFI to Legacy and then reinstall Ubuntu - in the BIOS look under the Boot Tab then click on boot mode - I have an Acer Laptop with an InsydeH2O BIOS that is what I had to do to get Linux Installed change it from UEFI to Legacy and it worked just fiine
 
Last edited by a moderator:
The next option is to change the BIOS from UEFI to Legacy and then reinstall Ubuntu - in the BIOS look under the Boot Tab then click on boot mode - I have an Acer Laptop with an InsydeH2O BIOS that is what I had to do to get Linux Installed change it from UEFI to Legacy and it worked just fiine
OK. so I changed to Legacy boot - didn't instantly fix anything. After changing to Legacy boot, I did yet another reinstall with the "erase disk" option - didn't choose fancy ZFS this time.

"No Boot Device Found. Press any key to reboot the machine"

Pressing any key doesn't do anything. Need to hold the power key in for 5 seconds to restart

Trying to install again - it thinks there is a version of 20.04 to install next to - but since I can't boot that I'm trying again with the "erase" option - this time using ZFS because let's exhaust all possibilities:

"No Boot Device Found. Press any key to reboot the machine"

So, I'm afraid I'm still stuck
 
Set your BIOS back to UEFI boot.
Boot from the liveCD.
Install efibootmgr if it's not installed (everything you install in the LiveCD session is in RAMFS, it won't touch your machine)
Run the following as root/superuser (you are already, this is for reference for anyone not doing this live)
Code:
efibootmgr
// Now note all the boot options. They are named "BootNNNN BloodyDevice" where NNNN is a hex number and BloodyDevice is a boot option.
// Example: Boot0000 FD
// Note the numbers, NNNN, of each item that isn't a BIOS standard, eg: BootNNNN UEFI Bloody-Thing
// Delete them all like so
efibootmgr --delete-bootnum -b 0xNNNN
// Where NNNN is what you noted earlier
// Finally run
efibootmgr --delete-bootorder
Do a fresh install.
Boot live from CD afterwards.
Mount your boot partition at /mnt/
Do this as root/superuser:
Code:
// I do not remember Ubuntu specifically, so you may need to tweak the paths where shimx64 is (not where you're gonna put it)
mkdir -p /mnt/EFI/BOOT/
cp /mnt/EFI/ubuntu/shimx64.efi /mnt/EFI/BOOT/BOOTX64.efi
// Not serious if the below isn't found:
cp /mnt/EFI/ubuntu/BOOTX64.CSV /mnt/EFI/BOOT/
Reboot and hope for the best.
 
Set your BIOS back to UEFI boot.
Boot from the liveCD.
Install efibootmgr if it's not installed (everything you install in the LiveCD session is in RAMFS, it won't touch your machine)
Run the following as root/superuser (you are already, this is for reference for anyone not doing this live)
Code:
efibootmgr
// Now note all the boot options. They are named "BootNNNN BloodyDevice" where NNNN is a hex number and BloodyDevice is a boot option.
// Example: Boot0000 FD
// Note the numbers, NNNN, of each item that isn't a BIOS standard, eg: BootNNNN UEFI Bloody-Thing
// Delete them all like so
efibootmgr --delete-bootnum -b 0xNNNN
// Where NNNN is what you noted earlier
// Finally run
efibootmgr --delete-bootorder
Do a fresh install.
Boot live from CD afterwards.
Mount your boot partition at /mnt/
Do this as root/superuser:
Code:
// I do not remember Ubuntu specifically, so you may need to tweak the paths where shimx64 is (not where you're gonna put it)
mkdir -p /mnt/EFI/BOOT/
cp /mnt/EFI/ubuntu/shimx64.efi /mnt/EFI/BOOT/BOOTX64.efi
// Not serious if the below isn't found:
cp /mnt/EFI/ubuntu/BOOTX64.CSV /mnt/EFI/BOOT/
Reboot and hope for the best.
Hi @Fanboi

After all that it boots into Grub 2.04 with "minimal BASH-like editing.... :-(

Just because you didn't mention it - I want to check - I still have Secure Boot OFF
Also, when you said
"Note the numbers, NNNN, of each item that isn't a BIOS standard, eg: BootNNNN UEFI Bloody-Thing" (and delete them) I just removed all the things which I had previously added in, plus I removed the "ubuntu" and "Linux something" which ubuntu had added over various installs - I hope that was right.

Also, you said I should do a:
mkdir -p /mnt/EFI/BOOT/

but that directory already existed and I had to overwrite BOOTX64.efi which was already there with the shimx64.efi forcefully (with sudo cp)

In any case, still not working I'm afraid.
 
Ok, from grub type
Code:
ls (hd0,1)/
you should see a file named vmlinuz or linux, and initrd.img
Type
Code:
linux (hd0,1)/vmlinuz root=/dev/sda1
or
Code:
linux (hd0,1)/linux root=/dev/sda1
depending on what you found with ls (hd0,1)/, then:
Code:
initrd (hd0,1)/initrd.img
Code:
boot

If you get initramfs rescue mode enter your password, then
Code:
startx
. Hopefully you get a desktop.
Use gparted to check your file system, if it reports an error, then you need to boot from a LiveCD or other media to fix it .... DO NOT attempt to repair a mounted partition.
The following three commands fix many grub boot problems. They run quick so just do all three instead of trying to find which one you need.
Code:
sudo grub-install /dev/sda
Code:
sudo update-grub
Code:
sudo update-initramfs -u

Reboot and see what you get.
 
Last edited by a moderator:
Ok, from grub type
Code:
ls (hd0,1)/
you should see a file named vmlinuz or linux, and initrd.img
Type
Code:
linux (hd0,1)/vmlinuz root=/dev/sda1
or
Code:
linux (hd0,1)/linux root=/dev/sda1
depending on what you found with ls (hd0,1)/, then:
Code:
initrd (hd0,1)/initrd.img
Code:
boot

If you get initramfs rescue mode enter your password, then
Code:
startx
. Hopefully you get a desktop.
Use gparted to check your file system, if it reports an error, then you need to boot from a LiveCD or other media to fix it .... DO NOT attempt to repair a mounted partition.
The following three commands fix many grub boot problems. They run quick so just do all three instead of trying to find which one you need.
Code:
sudo grub-install /dev/sda
Code:
sudo update-grub
Code:
sudo update-initramfs -u

Reboot and see what you get.
Hi @Lord Boltar thanks for your reply

So I'm having a problem quite early on in these particular instructions:
If I type
Code:
ls (hd0,1)/
I get a reply of just `efi/`

If I try
Code:
ls (hd1,1)/
I see `efi/ grub/`

Just to make that clear I checked on the disks (using the liveCD) and this means hd0 is associated with the HDD and hd1 is associated with the SSD. If you recall I was hoping to boot from the SSD but to be honest, now I'll take what I can get.

But either way, neither shows me files named vmlinuz or linux, and initrd.img

Still with the live CD I can see on the larger partition of the SSD a bunch of directories, including one called "boot" which has files named vmlinuz and initrd.img - should I try copying those files into the root of the boot partition? Thease would there be on (hd1,2)/boot/

Also, you wrote some stuff about (if I understood correctly) setting the root to /dev/sda1 . On my setup I think that is the HDD which also doesn't seem to have any of the linux or initrd.img on the first small partition. would it instead make sense for me to do something like
Code:
linux (hd1,1)/vmlinuz root=/dev/nvme0n1p1
Or were you specifically trying to go for the HDD over the SSD?

Please forgive my ignorance, I'm learning just as fast as I can
 
Last edited:
Hi @Lord Boltar thanks for your reply

So I'm having a problem quite early on in these particular instructions:
If I type
Code:
ls (hd0,1)/
I get a reply of just `efi/`

If I try
Code:
ls (hd1,1)/
I see `efi/ grub/`

Just to make that clear I checked on the disks (using the liveCD) and this means hd0 is associated with the HDD and hd1 is associated with the SSD. If you recall I was hoping to boot from the SSD but to be honest, now I'll take what I can get.

But either way, neither shows me files named vmlinuz or linux, and initrd.img

Still with the live CD I can see on the larger partition of the SSD a bunch of directories, including one called "boot" which has files named vmlinuz and initrd.img - should I try copying those files into the root of the boot partition? Thease would there be on (hd1,2)/boot/

Also, you wrote some stuff about (if I understood correctly) setting the root to /dev/sda1 . On my setup I think that is the HDD which also doesn't seem to have any of the linux or initrd.img on the first small partition. would it instead make sense for me to do something like
Code:
linux (hd1,1)/vmlinuz root=/dev/nvme0n1p1
Or were you specifically trying to go for the HDD over the SSD?

Please forgive my ignorance, I'm learning just as fast as I can
Yes, that's exactly what @Lord Boltar means: to manually point GRUB to where the installation should have. The issue may be that you have two devices wiith efi partitions, but only one of them has a valid OS (which explains a lot). Running the commands he suggested will fix that, but I suggest you remove the unwanted efi partition first (you can delete it from the hdd in gParted). If your board does not support booting from nvme -- and some older ones don't -- then you may have to use the hdd to boot, but I doubt that and you can always cross that bridge if you come to it.
 
Last edited:
Yes, that's exactly what @Lord Boltar means: to manually point GRUB to where the installation should have. The issue may be that you have two devices wiith efi partitions, but only one of them has a valid OS (which explains a lot). Running the commands he suggested will fix that, but I suggest you remove the unwanted efi partition first (you can delete it from the hdd in gParted). If your board does not support booting from nvme -- and some older ones don't -- then you may have to use the hdd to boot, but I doubt that and you can always cross that bridge if you come to it.
OK @Fanboi thanks

So, I'm documenting what I'm doing now for posterity, let's see how it goes.
  1. Rebooted into grub
  2. Code:
    ls (hd1,2)/boot/
    gives me an initrd.img and vmlinuz (among other files)
  3. Code:
    linux (hd1,2)/boot/vmlinuz root=/dev/nvme0n1p1
  4. Code:
    initrd (hd1,2)boot//initrd.img
  5. Code:
    boot
    - so I got lots of logs on screen and ended with an initramfs prompt
  6. Code:
    startx
    gives the error "sh: startx: not found
:-( So I can't get to all the logging from `boot`, but the part which I can see seems to go wrong near the end like this:


Code:
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: mountin /run on /root/run failed: No such file or directory
etc
etc

Maybe that gives you a clue. Restarting computer puts me back in grub
trying again:

But then I remembered what you said about the two EFI partitions - so I booted from the liveCD and deleted the EFI partition on the HDD. Rebooted and... IT BOOTS!!!!

Yes!!!! Thank you so much @Fanboi and @Lord Boltar !!! You have been very patient with a slow learner.
I'm just running the updates and will try exiting and rebooting to make sure all is working... yup seems to be fine!

Thank you so much again!
 
Glad it works! You can mark the thread [Solved].
 
Glad it works! You can mark the thread [Solved].
Once again forgive the ignorance @Fanboi but how do I do that? Do I just edit the thread title to put in the word SOLVED? [at the beginning or the end?] or is there a button somewhere to push/link to click?
 
Once again forgive the ignorance @Fanboi but how do I do that? Do I just edit the thread title to put in the word SOLVED? [at the beginning or the end?] or is there a button somewhere to push/link to click?
On your first entry you need to edit the title to say - [SOLVED] Problems (re)installing Ubuntu 20.04 on Laptop with UEFI SSD - get "ubi partman failed with exit code 10" - so just add [SOLVED] to the title
 

Members online


Latest posts

Top