[BUG] Incorrect images of Blender in Arm

[manju-rn]

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

In ARM images:

1. The version 5.1.0 actually displays Blender 4.0
2. The version 5.1.1 does not display in the browser at all - it most likely uses some wayland drivers

In AMD images, there is no such problem, 5.1.0 and 5.1.1 correctly displays the Blender version and work beautifully in browser. Hence, some issue on the ARM images

Expected Behavior

The ARM images should also work similar to AMD images - This is tested on Oracle Ampere instances

Steps To Reproduce

1. Get a ARM based machine/ VM (Like Oracle Ampere instance or mac) and run either 5.1.0 and 5.1.1 - the 5.1.0 version will actually show Blender 4.0 when launched in browser. The 5.1.1 does not even get rendered in browser

Environment

- OS:
- How docker service was installed:

CPU architecture

arm64

Docker creation

docker run -p 3001:3001 linuxserver/blender:5.1.1
docker run -p 3001:3001 linuxserver/blender:5.1.0

Container logs

None - no issues shown in logs

[github-actions]

Thanks for opening your first issue here! Be sure to follow the relevant issue templates, or risk having this issue marked as invalid.

[thelamer]

For arm we install from the repos which is whatever is in the ubuntu repos at the time :

https://github.com/linuxserver/docker-blender/blob/master/Dockerfile.aarch64#L25

I am not able to replicate what you are talking about though regarding it not working, it works fine on my identical instance.

2026-04-17.19-06-53.mp4

Try pulling the latest image maybe?

As far as the possiblity of bumping the version of blender on Arm, they don't distro arm https://mirror.clarkson.edu/blender/release/Blender5.1/

So until this gets rebased to ubuntu resolute that is basically where it is stuck.

Just as a side note there is no GPU acceleration support on our arm images (well there is best effort but in general nothing is confirmed) so running this on arm is just kind of pointless impo.

[manju-rn]

As you can see in your vid - the blender version of 4.0 which is pretty old. So from the tag linuxserver/blender:5.1.0 for arm it gives Belnder 4.0. While the same tag in AMD gives the correct 5.1. So maybe double check if the correct original repo is being used for building the image.

I am not able to get the page displayed for tag linuxserver/blender:5.1.1

I understand on the GPU acceleration on ARM images, but since Oracle Ampere isntance is free tier and pretty decent specs, i would lke to use it. If it works like linuxserver/blender:5.1.0 (but with a blendet 5.1 version) it would be good for me.

[thelamer]

Dude are you not reading what I am writing please read my response again let it sync in and then respond again.

[manju-rn]

@thelamer I did not read? it is you who probably did not understand

I am not able to replicate what you are talking about though regarding it not working, it works fine on my identical instance.

You state this above and then showed a vid where it shows exactly what I explained, regarding arm tag 5.1.0 showing Blender 4.0. So you were able to replicate exactly what I indicated.

Try pulling the latest image maybe?

tag 5.1.1 is the latest version. Did you use that and got the same results as your vid?

As far as the possiblity of bumping the version of blender on Arm, they don't distro arm https://mirror.clarkson.edu/blender/release/Blender5.1/
So until this gets rebased to ubuntu resolute that is basically where it is stuck.

So you do not take their (Blender's) original github https://github.com/blender/blender and build from that? I thought you guys did that and not just use a base ubuntunoble image and just install the pacakge - which may be old

[thelamer]

I'm closing this, please do not open another issue.

[manju-rn]

No wonder nobody raises issues here, just to be shutdown without proper discussion

[Roxedus]

Correct, we don't build blender ourselves. I think the right discussion to have here, is the removal of the arm builds, unless someone can point us to non-snap distributions of both arches.