JoshStrobl Thank you for helping. Should i be updating to the latest packages before running this command?
Can't boot after latest update
I found out the clock and date was set wrong
Question, I have Installed the Nvidia drivers via DoFlicky, is it still the same actions as mentioned in the dev tracker?
tom73287428764 Running the command should be fine. Please note that I am not entirely certainly what the behavior will be when installing, if clr-boot-manager will have three entries instead of two (with .10 and both .12), so might be worth having a Solus USB ready to run through the boot rescue doc and revert to .10 (by reverting the eopkg transaction).
RunningAroundIC Run eopkg li | grep 'nvidia'
If it isn't glx
but something like 390
, these instructions will not work for you. If you are using glx
then use the CLI commands from the task.
The individual on the development tracker reported that the kernels referenced on the task resolved the issue, so assuming positive responses from you folks, these will get pushed out to stable.
JoshStrobl It seems to have worked! Thanks!! clr-boot-manager list-kernels has two entries "5.14.12-201" and "5.14.10-199"
5.14.12-201 has a * next to it so i'm guessing that's the one that's selected? is there anything else i should be doing or should i wait until the packages are updated again a week from now and updata via software centre then
- Edited
5.14.12-201 has a * next to it so i'm guessing that's the one that's selected
That is correct.
is there anything else i should be doing or should i wait until the packages are updated again a week from now and updata via software centre then
You'll be on the latest kernel that will get pushed from unstable to stable today. No additional changes needed on your part, just keep on rollin. You will see a modaliases package update related to nvidia-glx-driver once these land in stable, this is just for hardware detection and not something you have to worry about immediately updating and rebooting for.
Apologies for the disruption. This worked on my laptop, which is a LUKS based installation. Obviously wouldn't have pushed .12 to begin with if it hadn't.
Hey, it installed the kernel and Nvidia drivers fine, but I still am stuck at bios startup screen, nothing happening. Any ideas what it can be?
- Edited
RunningAroundIC I would double check that you are using the -201 version. If you are unsure and haven't seen a prompt / list when booting, spam the spacebar while booting up and you should see a menu.
- Edited
The -201 is not listed
But in my eopkg history I see it installed, do I need to configure more?
When sudo usysconf run -f
I see that it is not updating the clr-boot-manager
- Edited
RunningAroundIC Sounds like clr-boot-manager
did not successfully run, possibly a full boot partition.
If you still have access to 5.14.10
and it's available in the boot menu, I would boot into that. You have two options after this, frankly I would suggest the first option.
The Easy Thing
- Revert the eopkg transaction(s) responsible for the installation of
-201
by following https://getsol.us/articles/package-management/history-and-rollback/en/ -- For example if it was transaction number 1000, you would dosudo eopkg history -t 999
to basically uninstall it. - Wait until
201
is pushed to the stable repo.
The Hard(er) Thing
sudo clr-boot-manager update
sudo clr-boot-manager list-kernels
Return the result from the last command. If 5.14.12-201
isn't listed there, it may be indicative of a full EFI System Partition (ESP). Do the following:
- Run
lsblk
to determine the partition device names. You may seesda
or it might be even benvme0n1
if it is an NVMe drive. - Run
sudo fdisk -o Device,Size,Type -l /dev/###
, replacing###
with either sda, nvme0n1, etc. You should see something like the following...
[~] sudo fdisk -o Device,Size,Type -l /dev/nvme0n1
Disk /dev/nvme0n1: 931.5 GiB, 1000204886016 bytes, 1953525168 sectors
Disk model: Samsung SSD 970 EVO Plus 1TB
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 1CDC09CE-51FC-456F-8E52-A5B5EDE6E12C
Device Size Type
/dev/nvme0n1p1 100M EFI System
/dev/nvme0n1p2 16M Microsoft reserved
/dev/nvme0n1p3 214.2G Microsoft basic data
/dev/nvme0n1p4 505M Windows recovery environment
/dev/nvme0n1p5 500M EFI System
/dev/nvme0n1p6 700.6G Linux filesystem
/dev/nvme0n1p7 15.6G Linux swap
Look for the EFI System
. If you are dual-booting with Windows and had correctly followed our UEFI instructions, you should have a 500M EFI System partition. If you didn't, you are probably sharing it with Windows (big no no) and it's full with both Solus and Windows stuff.
In my example, my EFI System Partition is /dev/nvme0n1p5
and so I will run:
sudo mount /dev/nvme0n1p5 /boot
After this, I run sudo tree /boot
, which returns
├── EFI
│ ├── Boot
│ │ └── BOOTX64.EFI
│ ├── com.solus-project
│ │ ├── initrd-com.solus-project.current.5.14.10-199
│ │ ├── initrd-com.solus-project.current.5.14.12-200
│ │ ├── initrd-com.solus-project.next.5.10.0-17
│ │ ├── kernel-com.solus-project.current.5.14.10-199
│ │ ├── kernel-com.solus-project.current.5.14.12-200
│ │ └── kernel-com.solus-project.next.5.10.0-17
│ └── systemd
│ └── systemd-bootx64.efi
└── loader
├── entries
│ ├── Solus-current-5.14.10-199.conf
│ ├── Solus-current-5.14.12-200.conf
│ └── Solus-next-5.10.0-17.conf
└── loader.conf
6 directories, 12 files
This is my desktop, not my laptop, and thus I have some extra stuff and not actually 201 (my laptop is what I tested 201 on).
In this scenario (as in, in your case, as you should be on .10), I would do the following:
- Remove 5.14.12* by doing:
sudo rm /boot/EFI/com.solus-project/*.solus-project.current.5.14.12-*
- Remove the 5.14.12* config by doing:
sudo rm /boot/loader/entries/Solus-current-5.14.12-*.conf
This would clean up space, and thus I would re-install the -201 kernels (following either the command from the task, or whatever means you did it). This is the easiest way to ensure clr-boot-manager is triggered and has the necessary files.
After this, re-run sudo clr-boot-manager list-kernels
and sudo tree /boot
(assuming it is still mounted) and double check that 201 is now listed. If it failed and /boot
is still mounted, try doing sudo umount /boot
then try the clr-boot-manager update again.
- Edited
Okay, so i did "The Easy Thing" and rolled back to before the 5.14.12 kernel.
Also i did "The hard(er) thing" by following your instructions and removing any trace of the 5.14.12 kernel
Im in my system again
Thank you so much for your time and help! its greatly appreciated!
I have learned a lot, and now theirs some guides i can follow is it ever happen again
Now deployed to stable repo. Marking as solved.
Would appreciate those that were impacted by this change test and reply on https://dev.getsol.us/T10006 for Linux 5.15.1 testing. Thanks in advance. @RunningAroundIC @tom73287428764
- Edited
JoshStrobl Ran the following commands:
sudo eopkg install https://joshuastrobl.com/5.15-testing/linux-current-5.15.1-206-1-x86_64.eopkg
and
sudo eopkg install https://joshuastrobl.com/5.15-testing/nvidia-glx-driver-32bit-470.82.00-414-1-x86_64.eopkg https://joshuastrobl.com/5.15-testing/nvidia-glx-driver-common-470.82.00-414-1-x86_64.eopkg https://joshuastrobl.com/5.15-testing/nvidia-glx-driver-current-470.82.00-414-1-x86_64.eopkg https://joshuastrobl.com/5.15-testing/nvidia-glx-driver-modaliases-470.82.00-414-1-x86_64.eopkg
Rebooted and everything seems to be working so far
EDIT: I have noticed i get a green + purple line across the screen right after the LUKS prompt. I'm guessing this isn't an issue but posting just in case