Accidentally deleted EFI partition

stekte

root@solus /home/live # mount /dev/sdc1 /target/boot
root@solus /home/live # mount --types proc /proc /target/proc
mount --rbind /dev /target/dev
mount --rbind /sys /target/sys
mount --make-rslave /target/dev
mount --make-rslave /target/sys
root@solus /home/live # chroot /target

In /target you may mount the system partition not the /boot. The /boot partition will be
mounted when in chroot environment.
So from live you mount /dev/sdc3 on /target and when you chrooted in /target you
mount /dev/sdc1 /boot/efi and then you start the recovery procedure

JanAlbert

stekte Mounted boot once chrooted, didn't work.

Device Size Type /dev/sda1 1G EFI System /dev/sda3 219.4G Linux filesystem root@solus /home/live # mount --types proc /proc /target/proc mount --rbind /dev /target/dev mount --rbind /sys /target/sys mount --make-rslave /target/dev mount --make-rslave /target/sys root@solus /home/live # chroot /target root@solus / # mount /dev/sda1 /boot/efi mount: /boot/efi: mount point does not exist. dmesg(1) may have more information after failed mount system call.

minh

JanAlbert sudo efibootmgr -c -d /dev/sdb -p 4 -l "\EFI\systemd\systemd-bootx64.efi"

Mine is different.
Boot0001* Solus Linux Bootloader HD(1,GPT,5574757f-9f73-11ee-9cc2-7085c28ac21f,0x800,0x100000)/\EFI\COM.SOLUS-PROJECT\BOOTLOADERX64.EFI
Also from your previous post, the /target should mount to your sdb disk (existing solus), not sdc (live cd)?

JanAlbert

minh That only changed the name of the option in the F11 boot menu. Still not booting.

BuzzPCSOS

Not trying to tread on toes here, just suggesting another path that might be successful.
@JanAlbert There has been some excellent advice coming your way but I must admit that I prefer using a GUI to CLI. For me there is too much chance of getting things wrong and making it all worse. My go-to tool in a pickle like yours would be Rescatux which can be booted into a live environment to perform an in-situ repair.
Make a boot media with Rescatux on it. Maybe switch off Secure Boot in the BIOS. With Secure Boot enabled I can no longer boot from any Solus 4.5 iso but your mileage may vary depending on BIOS updates.
Start Rescatux. Select the Menu button and then in the Boot section try any of the following.
UEFI Partition Status
Check UEFI Boot
Create UEFI Boot Entry
Alternatively look for some online tutorials on it's use or use the Rescatux chat function on the Live system. It has been around a good while now and has often helped me to bring a computer out of limbo with the minimum of fuss.
Good Luck with your repair.

JanAlbert

BuzzPCSOS I appreciate the suggestion. I'll keep it in the back of my mind if no success otherwise.

JanAlbert

Whenever I choose the first option, my existing OS, it hangs on the Pro Series logo.

The third option is the one I recently installed to have at least a workable OS.

HomerSp

Quite new here, but I have done the same thing and was able to recover.
While in the chroot, after running clr-boot-manager update, run this, and post the output here:
blkid efibootmgr ls -lR /boot

JanAlbert

HomerSp

`root@solus / # sudo clr-boot-manager update
[ERROR] cbm (../src/bootman/update.c:L189): Cannot determine the currently running kernel
root@solus / # blkid
/dev/loop1: UUID="bf0f7656-8889-44c6-8d03-2954fbfd9e2d" BLOCK_SIZE="4096" TYPE="ext4"
/dev/sdd1: LABEL="backup" UUID="0c047187-220b-402e-941b-bf0439531fce" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="36fe5fb2-5a88-47b9-90d0-02f4514330b7"
/dev/sdb: LABEL="HD" UUID="999149fb-b994-4d58-8872-b60ea5321df7" BLOCK_SIZE="4096" TYPE="ext4"
/dev/loop0: BLOCK_SIZE="1048576" TYPE="squashfs"
/dev/sde2: SEC_TYPE="msdos" LABEL_FATBOOT="SOLUSESP" LABEL="SOLUSESP" UUID="CF7D-884C" BLOCK_SIZE="512" TYPE="vfat" PARTUUID="7941b268-02"
/dev/sde1: BLOCK_SIZE="2048" UUID="2024-01-08-18-20-40-00" LABEL="SolusLiveBudgie" TYPE="iso9660" PTUUID="7941b268" PTTYPE="dos" PARTUUID="7941b268-01"
/dev/sdc2: UUID="35b2da2b-862a-4bd9-8adf-d4a157deec96" BLOCK_SIZE="4096" TYPE="ext4" PARTLABEL="root" PARTUUID="ab380813-e354-499f-bcf4-f9194ec6e148"
/dev/sdc1: UUID="503B-FA60" BLOCK_SIZE="512" TYPE="vfat" PARTLABEL="EFI" PARTUUID="f358023e-366e-45fc-854b-3ab7e4a89a23"
/dev/sda3: UUID="c8d899b5-15c8-4183-b733-4bbf612bef4e" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="62a30208-d945-4310-be3f-fec55cdd86e8"
/dev/sda1: UUID="F8A8-AC68" BLOCK_SIZE="512" TYPE="vfat" PARTLABEL="EFI" PARTUUID="cbc1929a-3a1c-4208-b29e-b66722bed858"
/dev/zram0: LABEL="zram0" UUID="f7f09200-0624-4e41-b6ce-ea66d07f646b" TYPE="swap"
root@solus / # efibootmgr
bash: efibootmgr: command not found
root@solus / # ls -lR /boot
/boot:
total 12
drwxr-xr-x 4 root root 4096 Jul 11 17:00 EFI
drwxr-xr-x 3 root root 4096 Jul 11 17:00 loader
-rwxr-xr-x 1 root root 1321 Jul 20 09:03 solus-enroll-me.cer

/boot/EFI:
total 8
drwxr-xr-x 2 root root 4096 Jul 20 09:03 BOOT
drwxr-xr-x 2 root root 4096 Jul 20 09:03 com.solus-project

/boot/EFI/BOOT:
total 1016
-rwxr-xr-x 1 root root 949424 Jul 20 09:03 BOOTX64.EFI
-rwxr-xr-x 1 root root 87816 Jul 20 09:03 fbx64.efi

/boot/EFI/com.solus-project:
total 126064
-rwxr-xr-x 1 root root 949424 Jul 20 09:03 bootloaderx64.efi
-rwxr-xr-x 1 root root 152 Jul 20 09:03 BOOTX64.CSV
-rwxr-xr-x 1 root root 75056 Jul 20 09:03 grubx64.efi
-rwxr-xr-x 1 root root 73072988 Jul 11 17:00 initrd-com.solus-project.current.6.8.12-293
-rwxr-xr-x 1 root root 41535952 Jul 11 17:00 initrd-com.solus-project.current.6.8.12-293.nvidia
-rwxr-xr-x 1 root root 12590384 Jul 11 17:00 kernel-com.solus-project.current.6.8.12-293
-rwxr-xr-x 1 root root 848080 Jul 20 09:03 mmx64.efi

/boot/loader:
total 8
drwxr-xr-x 2 root root 4096 Jul 11 17:00 entries
-rwxr-xr-x 1 root root 38 Jul 11 17:00 loader.conf

/boot/loader/entries:
total 4
-rwxr-xr-x 1 root root 486 Jul 11 17:00 Solus-current-6.8.12-293.conf
root@solus / #

HomerSp

JanAlbert Thanks, but it looks like you don't have efibootmgr installed in the system. The live iso should have it though, so try running it outside of the chroot. If it's still running, just exit out of the chroot and run efibootmgr again and post the output here. I need it to verify the boot list is correct (including the PARTUUID from the blkid command)

JanAlbert

HomerSp I appreciate you taking the time to look at this. Here it is:

jad2@jad2-solus ~ $ efibootmgr BootCurrent: 0000 Timeout: 1 seconds BootOrder: 0001,0002,000A,0000,0007,000B Boot0000* Solus Linux Bootloader HD(1,GPT,f358023e-366e-45fc-854b-3ab7e4a89a23,0x1000,0x200000)/\EFI\COM.SOLUS-PROJECT\BOOTLOADERX64.EFI Boot0001* Solus Linux Bootloader HD(1,GPT,f358023e-366e-45fc-854b-3ab7e4a89a23,0x1000,0x200000)/\EFI\com.solus-project\bootloaderx64.efi Boot0002* Solus Linux Bootloader HD(1,GPT,cbc1929a-3a1c-4208-b29e-b66722bed858,0x800,0x200000)/\EFI\COM.SOLUS-PROJECT\BOOTLOADERX64.EFI Boot0007* UEFI OS HD(1,GPT,cbc1929a-3a1c-4208-b29e-b66722bed858,0x800,0x1f4000)/\EFI\BOOT\BOOTX64.EFI0000424f Boot000A* UEFI OS HD(1,GPT,cbc1929a-3a1c-4208-b29e-b66722bed858,0x800,0x200000)/\EFI\BOOT\BOOTX64.EFI0000424f Boot000B UEFI OS HD(1,GPT,f358023e-366e-45fc-854b-3ab7e4a89a23,0x1000,0x200000)/\EFI\BOOT\BOOTX64.EFI0000424f jad2@jad2-solus ~ $

By the way, I'm in NZ, so probably a bit of time difference with most people here.

HomerSp

JanAlbert Very strange, it does all look correct. Try running this in the live session (not chroot):
sudo efibootmgr -c --loader \\EFI\\com.solus-project\\bootloaderx64.efi --label SOLUS --disk /dev/sda --part 1
then reboot and check if SOLUS shows up in the UEFI boot list, and if it does, try booting it.
If it still doesn't boot, try pressing the Spacebar key just after you press the boot option. You may have to mash the key or press and hold. This should bring up the systemd boot menu, in which case we know the bootloader is working at least. Let me know how it goes.
Alternatively, you can, without having to press spacebar, modify the loader config to always show the menu. From the live session, mount the boot partition, and add timeout to loader.conf:
sudo mkdir /target-boot sudo mount PARTUUID=cbc1929a-3a1c-4208-b29e-b66722bed858 /target-boot echo "timeout 5" | sudo tee -a /target-boot/loader/loader.conf

JanAlbert

HomerSp
Yes, SOLUS shows up and then still hanging on logo after booting.
And Spacebar brings up systemd boot menu. After booting from there hanging on blank screen.

plutuplutu

JanAlbert still hanging on logo after booting

Apologies if I'm out of topic, but did you try "failsafe" boot?
Select Solus, before return hit e to add the single boot option.
Might (or might not) give you additional hindsight.

Good luck

HomerSp

JanAlbert That's good news, because it means there's nothing wrong with the bootloader. Don't know why it still doesn't boot though, but let's try and figure it out.
While in the systemd boot menu you can try pressing 'e' to edit the boot options and remove the 'quiet splash' part, this should give us a better understanding of what's going on as you should see the boot logs.
You can also try reinstalling the kernel from the live session:
sudo mkdir /target sudo mount PARTUUID=62a30208-d945-4310-be3f-fec55cdd86e8 /target sudo mount PARTUUID=cbc1929a-3a1c-4208-b29e-b66722bed858 /target/boot sudo mount --type proc /proc /target/proc sudo mount --rbind /dev /target/dev sudo mount --rbind /sys /target/sys sudo mount --rbind /run /target/run sudo chroot /target eopkg install --reinstall linux-current

The sudo mount --rbind /run /target/run is required to have internet access while inside the chroot.
It looks like you're running on a slightly older kernel though, so it might get updated while you try to reinstall (which is fine).

JanAlbert

plutuplutu
Thanks for the suggestion. Adding single and hitting enter resulted in the same hanging on a black screen.

JanAlbert

HomerSp
We might be getting somewhere. After removing 'quiet splash' this came up:

Stuck on a line with 'uuid-5c6...'.

And this is the line I took 'quiet splash' out (in 4 parts):

At the end that same UUID starting with 5c6.

I haven't seen that UUID before. Could this be a clue?

JanAlbert

Could it be an old swap partition that's no longer there? If so, how to resolve?

HomerSp

JanAlbert Yeah, that kernel option is for hibernation - you are supposed to create a partition that's as big as your RAM that's used for swap. If you're not using hibernation you can simply remove the whole resume= part from the boot edit menu and see if it boots. If it does, you may also need to remove it from /etc/kernel/cmdline, and run clr-boot-manager update again.

JanAlbert

HomerSp
Yay, SUCCESS! Such a relief to have my familiar DE back!
Yes, I removed the resume part in the boot edit menu as well as the 10_resume.conf file in /etc/kernel/cmdline.d.
So grateful for your support by guiding me towards the solution one step at a time.

« Previous Page