stable-20250126.1 is unbootable on Nvidia systems #2157
Comments
|
Same here, it seems to be failing on the sb key. |
|
This turns out to be a nvidia kernel module issue for me. I think the NV drivers are not bumped somewhere? |
|
I'm also experiencing this issue on builds newer than 1/24 of Bluefin. |
|
@francoism90 @dylanmtaylor please edit your posts and toss in an rpm-ostree status if you don't mind. |
castrojo
changed the title
|
p5 has pushed the last known good -nvidia images to the tags, so if you do an update you should be good. Can ya'll confirm if that is indeed the case? Then when the team fixes it we'll just republish the new images. |
|
The latest bluefin/aurora-nvidia (normal and dx) builds seem to date from 12h ago Running Did you mean something else by
? |
|
@castrojo Everything after 26 Jan doesn't work. I don't fully understand, do you mean you have pushed an update? And do you mean updating using ujust? |
That said, the 570 driver is in both repos now so there should be no more mismatch |
|
@dylanmtaylor @bsherman (hopefully tagging the correct persons) - it works again. :) $ rpm-ostree status
State: idle
AutomaticUpdates: stage; rpm-ostreed-automatic.timer: no runs since boot
Deployments:
● ostree-image-signed:docker://ghcr.io/ublue-os/bluefin-dx-nvidia:latest
Digest: sha256:4838d6eefbe781581a27f216815e02de0bc900821caf7ff3a4baa9ca06df8644
Version: latest-41.20250129.2 (2025-01-29T05:16:06Z)Thanks for the fix! Let's hope the new (opensource) nvidia driver will prevent this in the future. I had this issues in the past on Arch a couple of times. |
|
The new images work! However, as of now, the currently tagged "stable" for *-nvidia images are still not pointing to the fixed images (meaning that stable users updating their OS will still run into this issue) Would there be a way to hotfix this @castrojo? |
|
This is fixed for me on the new image versions |
@Malix-Labs I've just checked nothing 570 is in those stable images... I believe @p5 mitigated this yesterday around 12:00 EST by re-tagging stable to known good images. New builds are looking good. I think we can close this issue now. |
|
@bsherman, I just ran I am back at my previous deployment again https://paste.centos.org/view/3242a436 Note that the deployment was using the cache for sure
but when I upgrade again, the deployment is way too fast to not be using cache |
|
Ok let's just push a new stable, try again when this turns green, thanks! https://github.com/ublue-os/bluefin/actions/runs/13039922089 |
This is odd. Here's a snippet of me rebasing to downgrade to this image. This is the same digest as what you report, but it has me moving to a consistent Nvidia 565 build. Not sure what the problem is on your system Regardless, with Jorge building new releases of stable, I expct that will resolve things. |
|
|
|
It...
...
GG for everyone involved except my slow ass internet <3 |
|
Could there be any added unit tests to prevent as much causes of unbootable system as possible? |
|
I tried many things but the new image seem to not work either It is probable than yesterday night I didn't booted to the new deployment by mistake, sorry for the false resolved status Back on 41.20250119.1 again |
Describe the bug
stable-20250126.1 seems unbootable (tested with Bluefin-DX-Nvidia on a Lenovo Legion 5)
My booting screen (personally, with a Lenovo and a Bluefin logo plus a spinning loading icon) shows up, then seems to freeze after a few seconds, then the screen turns full black forever
What did you expect to happen?
A working stable release
Output of
rpm-ostree statusOutput of
bootc statusOutput of
sudo bootc statusOutput of
fastfetchExtra information or context
Need to rollback to previous deploy (last week's stable - stable-20250119.1)
The text was updated successfully, but these errors were encountered: