Thanks for the rapid release

 

Updating to 6.9 beta30 now.

On 9/29/2020 at 5:47 PM, thor2002ro said:

...

update zfs to 2.0 rc2

fix zfs startup script
...

Thx for the work, I'm on 5.9.0rc8-thor-Unraid+.NV.6.9b30.zip and zfs-2.0.0-rc2-x86_64-thor.tgz

Jfyi the ZFS zpool devices in /dev/<zpool>/<poolname> are still missing.

Does running in terminal 

 

./etc/rc.d/rc.zfs start

 

 

Work? 

 

 

 

updated zfs package and reuploaded it

I have a AMD 2200G with built in graphics.  The machine boots up fine but the display stops working at this point

 

Tower kernel: amdgpu 0000:09:00.0: Direct firmware load for amdgpu/raven_gpu_info.bin failed with error -2

Tower kernel: amdgpu 0000:09:00.0: amdgpu: Failed to load gpu_info firmware "amdgpu/raven_gpu_info.bin"

Tower kernel: amdgpu 0000:09:00.0: amdgpu: Fatal error during GPU init

Tower kernel: [drm] amdgpu: finishing device.

Tower kernel: amdgpu: probe of 0000:09:00.0 failed with error -2

 

what do I need to add to stop the amdgpu fail?

 

46 minutes ago, rilles said:

I have a AMD 2200G with built in graphics.  The machine boots up fine but the display stops working at this point

 

Tower kernel: amdgpu 0000:09:00.0: Direct firmware load for amdgpu/raven_gpu_info.bin failed with error -2

Tower kernel: amdgpu 0000:09:00.0: amdgpu: Failed to load gpu_info firmware "amdgpu/raven_gpu_info.bin"

Tower kernel: amdgpu 0000:09:00.0: amdgpu: Fatal error during GPU init

Tower kernel: [drm] amdgpu: finishing device.

Tower kernel: amdgpu: probe of 0000:09:00.0 failed with error -2

 

what do I need to add to stop the amdgpu fail?

 

After lots of googling and some guessing - this worked

 

modprobe.blacklist=amdgpu

 

you forgot to update the bzfirmware 

look into the "5.8.1-20200814" release theres a bzfirmware update the on included with unraid is really basic....

5.9.1-20201026 release 

 

update to 5.9.1
updated paragon ntfs3 v10
improvements to pcie aer
nvidia driver update 450.80.02
updated bzfirmware

 

update/add drivers out of tree:
corefreq kernel module(added utils for corefreq module run corefreq-cli-run) d058a1e
zfs drivers 3928ec5
tbsecp3 drivers 3cdeaee
asus-wmi-sensors driver 3
r8125 driver 9.003.05
r8152 driver 2.13.0
ryzen_smu driver 44a0f687
tn40xx driver 0.3.6.17.3
zenpower driver 0.1.12

bzfirmware was the missing bit.  all good - hope to do some load testing now to see if my AMD still crashes with this newer then 6.8.3 kernel.

 

thanks for doing this.

thank you for this amazing kernel. My performance in general and in VM's feels more snappier. But i do have one issue when using the nvidia driver utils. I have installed everything as shown in the first post, but when using GPU transcoding in Jellyfin docker container, the playback stalls complaining about a CUDA init error. The same transcoding works when using the unraid nvidia builds. any idea what could be the issue?

as the last comment the VM's feels better but i cant get the Nvidia driver to work on dockers.

 

when i run nvidia-smi i get

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.80.02    Driver Version: 450.80.02    CUDA Version: 11.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Quadro P2000        On   | 00000000:0D:00.0 Off |                  N/A |
| 47%   37C    P8     4W /  75W |      0MiB /  5059MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

but in plex or emby its not working

 

 

if i try - docker run --runtime=nvidia --rm nvidia/cuda nvidia-smi

 

i get this error

 

docker: Error response from daemon: OCI runtime create failed: container_linux.go:349: starting container process caused "process_linux.go:449: container init caused \"process_linux.go:432: running prestart hook 1 caused \\\"error running hook: exit status 1, stdout: , stderr: nvidia-container-cli: requirement error: unsatisfied condition: cuda>=11.1, please update your driver to a newer version, or use an earlier cuda container\\\\n\\\"\"": unknown.

 

 

Thanks for your work.

 

yes I can reproduce it... I think the last version kernel has an issue with docker nvidia..... will look into it in the next version....

I found the problem it needs an updated driver .... the nvidia docker is using cuda 11.1 and the 450.80.02 driver that is included in the last kernel has cuda 11.0 lol its a mismatch ...

5.9.4-20201105

 

update to 5.9.4
alternative unraid driver disk spindown using hdparm
nvidia driver update 455.38
updated bzfirmware(get last from 5.9.1-20201026 release)

 

update/add drivers out of tree:
corefreq kernel module(added utils for corefreq module run corefreq-cli-run) 5816cf3
zfs drivers 52e585a
tbsecp3 drivers 3cdeaee
asus-wmi-sensors driver 3
r8125 driver 9.003.05
r8152 driver 2.13.0
ryzen_smu driver 44a0f687
tn40xx driver 0.3.6.17.3
zenpower driver 0.1.12

dont know what im doing wrong but when i try your kernel all works but the nvidia drivers with docker stop working, the plex give me a unkown error.

Zfs is working, Zen is working only the nvidia drivers with docker. the performance in vms is just better and snappy.

On 11/11/2020 at 12:50 PM, loomitz said:

dont know what im doing wrong but when i try your kernel all works but the nvidia drivers with docker stop working, the plex give me a unkown error.

Zfs is working, Zen is working only the nvidia drivers with docker. the performance in vms is just better and snappy.

are you on the latest 5.9.4?

 

I'm using a test version of 5.10 but it uses the same nvidia driver should work

 

 

Hello  Great work, thank you  
Is there any reason why it wouldn't work with plex (linuxserver/plex)

 

The card shows up in the nvidia-smi from the host and the container, but plex (despite enabling hardware acceleration) doesn't use it. 

 

 

 

I have installed your kernel + nvidia driver utils from the latest release on unraid 6.8.3. Do you have any advice as to what might be the reason?

 

Thanks in advance  

good news I figured out the nvidia issue and nvidia and docker will work in the next version

release 5.10rc4-20201118

 

update to 5.10rc4
nvidia driver update 455.45.01
fixed nvidia docker
corsair power supply kernel driver
Add RCEC handling to PCI/AER v11
add support for virtio-mem: Big Block Mode (BBM)
update ntfs3 driver v12
updated bzfirmware(get last from 5.9.1-20201026 release)
removed all old AMD GPU reset quirks
add support for new AMD GPU reset module vendor_reset(auto resets all AMD GPUS when it needs to please dont run any other reset methods, to use add in /config/go file "modprobe vendor_reset")
supports:
AMD | Polaris 10 |  
AMD | Polaris 11 |  
AMD | Polaris 12 |  
AMD | Vega 10 | Vega 56/64
AMD | Vega 20 | Radeon VII
AMD | Navi 10 | 5600XT, 5700, 5700XT
AMD | Navi 12 | Pro 5600M
AMD | Navi 14 | Pro 5300, RX 5300, 5500XT

 

update/add drivers out of tree:
corefreq kernel module(added utils for corefreq module run corefreq-cli-run) e4c271f
zfs drivers 0ca45cb
tbsecp3 drivers removed not building
asus-wmi-sensors driver 3
r8125 driver 9.003.05
r8152 driver 2.13.0
ryzen_smu driver 44a0f687
tn40xx driver 0.3.6.17.3
zenpower driver 0.1.12
vendor_reset module 6140e2f

RR272x_1x does not build so removed

Edited November 18, 20205 yr by thor2002ro

I updated to the newest version. It does seemthat the nvidia installation has gone further. Unfortunately somehow it failed: 

 

logs have following errors:

/bin/bash: line 14: nvidia-smi: command not found

 

installation...

 

failed to initialize NVML: Driver/library version mismatch 

 

Here is a photo of this:

 

 

For some reason, it caused webpanel to not start. 

 

p.s. hypervisor/libvirt errors are not related, it was there before (something with vm backup plugin)

 

 

you need to update the hole thing not just the packages....

the packages are tied to the kernel.... bzimage and bzmodules

Edited November 18, 20205 yr by thor2002ro

I did replace all of the provided files using new release, for both nvidia drivers and kernel package (did it once again just to be sure), but the problem persist. 

 

In the first post you put this: 

 

dont try to install the utils packages on any modified rootfs .... they need to be installed on stock unraid rootfs

 

Does that mean that i should replace those files using original unraid file from before i started? 

no..... you dont need the original files .....

but the kernel you are loading seams to be 5.9.4 not 5.10rc4..... so the bzimage is not the right one for sure....

I am probably doing something very stupid here, but it seems to me that those are exactly the same files:

 

Edited November 18, 20205 yr by dunioo

dono what to say.... maybe it was a flash write error .... try to overwrite the files again and be sure to safely remove the flash stick and try again....