5 hours ago, NullZeroNobody said:

 

Thanks Saarg, you are correct. How do I fix this? And if I make any changes, will it affect the VM I have it passed through to? I would shut off the VM when using the GPU for Handbrake Docker encoding, and vice versa. 

I'm not sure if the card is released from vfio-pci automatically or you have to force it somehow. You might have to remove it from the VM template and reboot.

I haven't playes around with this, but maybe someone else have and can chime in.

7 hours ago, saarg said:

I'm not sure if the card is released from vfio-pci automatically or you have to force it somehow. You might have to remove it from the VM template and reboot.

I haven't playes around with this, but maybe someone else have and can chime in.

Went into Settings -> VFIO-PCI Config, unchecked the bind for the 2060, and was able to assign it to the handbrake docker and run successfully. If i stop the docker, then the VM with the 2060 attached booted without issue, which was what I was looking for. 

 

Thanks for your help!

Since 6.8.3 I haven't been able to pass my GPU through to a VM without crashing the server, requiring a power cycle.

Tho only other thing that changed is I changed my GUI mode GPU from an old AMD card to a GT710 and added a 10GbE NIC

 

I haven't been able to suss out the reason, hoping something sticks out in my config to someone here.

note: pass thru works fine on vanilla 6.8.3

fortytwo-diagnostics-20200906-0802.zip

Edited September 6, 20206 yr by tjb_altf4

10 hours ago, tjb_altf4 said:

Since 6.8.3 I haven't been able to pass my GPU through to a VM without crashing the server, requiring a power cycle.

Tho only other thing that changed is I changed my GUI mode GPU from an old AMD card to a GT710 and added a 10GbE NIC

 

I haven't been able to suss out the reason, hoping something sticks out in my config to someone here.

note: pass thru works fine on vanilla 6.8.3

fortytwo-diagnostics-20200906-0802.zip 308.78 kB · 1 download

If the GPU is being used by something, you will get a crash. If it's not in use, something probably changed in either the kernel or Nvidia driver.

You have to remember that swapping it around like that isn't a normal thing to do. You should first unload the Nvidia modules and then bind it to vfio.

1 hour ago, saarg said:

If the GPU is being used by something, you will get a crash. If it's not in use, something probably changed in either the kernel or Nvidia driver.

You have to remember that swapping it around like that isn't a normal thing to do. You should first unload the Nvidia modules and then bind it to vfio.

Definitely nothing using the GTX1060 (card being passed thru), the GT710 is for host (GUI) and no VMs or dockers running this time.

No point binding the card, as I won't be able to use it for docker and vanilla unraid works just fine for that use case.

32 minutes ago, tjb_altf4 said:

Definitely nothing using the GTX1060 (card being passed thru), the GT710 is for host (GUI) and no VMs or dockers running this time.

No point binding the card, as I won't be able to use it for docker and vanilla unraid works just fine for that use case.

 

Vanilla unraid doesn't load the drivers for the GPU at all, so it's no wonder there isn't any issues passing it through to the VM. When you want to.

 

You have to unbind the card from the Nvidia modules before you use it in your VM. There is no way around it if unraid crashes when you start the VM.

10 minutes ago, saarg said:

You have to unbind the card from the Nvidia modules before you use it in your VM. There is no way around it if unraid crashes when you start the VM.

Can that be done on the fly, or only at boot time?

6 hours ago, tjb_altf4 said:

Can that be done on the fly, or only at boot time?

If you look at the instructions about dumping the bios here on the forum, you should find a command to unbind the card from modules. Hopefully that is enough.

I don't remember which thread it was in, so you will need to search for it. Probably best doing it on Google.

Hi all!  I'm getting this set of errors when going to the Nvida plugin page.  Not sure what has happened, but something is up, as I've never received this set of errors before.

 

<a href="https://ibb.co/KznBzBb"><img src="https://i.ibb.co/sJ7cJcb/Screen-Shot-2020-09-11-at-7-30-20-AM.png" alt="Screen-Shot-2020-09-11-at-7-30-20-AM" border="0"></a>

4 minutes ago, de la trevie said:

Hi all!  I'm getting this set of errors when going to the Nvida plugin page.  Not sure what has happened, but something is up, as I've never received this set of errors before.

 

<a href="https://ibb.co/KznBzBb"><img src="https://i.ibb.co/sJ7cJcb/Screen-Shot-2020-09-11-at-7-30-20-AM.png" alt="Screen-Shot-2020-09-11-at-7-30-20-AM" border="0"></a>

Your link didn't work. We prefer attachments instead of links to external sites anyway. Now that you have been approved you should be able to attach the image to your NEXT post.

Sorry about that.  Here's the screen cap of the errors.

On 9/11/2020 at 7:44 AM, de la trevie said:

Sorry about that.  Here's the screen cap of the errors.

Is this something someone else has encountered before, or is this something that is going to have to take a full reset of unRAID to fix?

Hey there, i've got issues with following card:
palit geforce gtx 1650 kalmX (https://www.palit.com/palit/vgapro.php?id=3494&lang=en)

I tried several Unraid Nvidia Builds that are avaible from the WebGUI, nothing seems to work.

result of running "lspci"

04:00.0 VGA compatible controller: NVIDIA Corporation TU117 [GeForce GTX 1650] (rev a1)

a old Geforce GT 730 works just fine. Somebody an idea how to get this card working?

13 hours ago, blackbunt said:

Hey there, i've got issues with following card:
palit geforce gtx 1650 kalmX (https://www.palit.com/palit/vgapro.php?id=3494&lang=en)

I tried several Unraid Nvidia Builds that are avaible from the WebGUI, nothing seems to work.

result of running "lspci"

04:00.0 VGA compatible controller: NVIDIA Corporation TU117 [GeForce GTX 1650] (rev a1)

a old Geforce GT 730 works just fine. Somebody an idea how to get this card working?

Not sure what your problem is, but I have one of those cards installed in one of my servers and it is working fine with the current Nvidia builds.   I specifically went for that card because of the fact that it is fanless and therefore silent.

1 hour ago, itimpi said:

Not sure what your problem is, but I have one of those cards installed in one of my servers and it is working fine with the current Nvidia builds.   I specifically went for that card because of the fact that it is fanless and therefore silent.

i am running the exact same card in my main server without any issues.

i popped the working gtx 1650 SUPER from my main server and have the same issues with that one in my backup server

Weirdly it shows up under SysDEvs as following:

 

IOMMU group 13:[10de:2187] 04:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)

[10de:1aeb] 04:00.1 Audio device: NVIDIA Corporation TU116 High Definition Audio Controller (rev a1)

[10de:1aec] 04:00.2 USB controller: NVIDIA Corporation Device 1aec (rev a1)

[10de:1aed] 04:00.3 Serial bus controller [0c80]: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)

is there a way to force the driver to use the gpu?

Edited September 13, 20206 yr by blackbunt

10 hours ago, blackbunt said:

i am running the exact same card in my main server without any issues.

i popped the working gtx 1650 SUPER from my main server and have the same issues with that one in my backup server

Weirdly it shows up under SysDEvs as following:

 

IOMMU group 13:[10de:2187] 04:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)

[10de:1aeb] 04:00.1 Audio device: NVIDIA Corporation TU116 High Definition Audio Controller (rev a1)

[10de:1aec] 04:00.2 USB controller: NVIDIA Corporation Device 1aec (rev a1)

[10de:1aed] 04:00.3 Serial bus controller [0c80]: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)

is there a way to force the driver to use the gpu?

The GPU is probably stubbed. Check which module is loaded with lspci -k

20 hours ago, saarg said:

The GPU is probably stubbed. Check which module is loaded with lspci -k

04:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
        Kernel driver in use: nvidia
        Kernel modules: nvidia_drm, nvidia
04:00.1 Audio device: NVIDIA Corporation TU116 High Definition Audio Controller (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
04:00.2 USB controller: NVIDIA Corporation Device 1aec (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
        Kernel driver in use: xhci_hcd
04:00.3 Serial bus controller [0c80]: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850

2 hours ago, blackbunt said:

04:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
        Kernel driver in use: nvidia
        Kernel modules: nvidia_drm, nvidia
04:00.1 Audio device: NVIDIA Corporation TU116 High Definition Audio Controller (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
04:00.2 USB controller: NVIDIA Corporation Device 1aec (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850
        Kernel driver in use: xhci_hcd
04:00.3 Serial bus controller [0c80]: NVIDIA Corporation TU116 [GeForce GTX 1650 SUPER] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device 3850

What is the actual problem? You only said it doesn't work.

Card isn’t recognized.

Hi,

 

I am trying to install the drivers on my machine and it just halts after trying to start downloading bzroot-gui. I have tried a few times now and it stalls at the same place each time. When I go back to stock build it download and copies without errors.

 

*****

Now installing Nvidia version 6.8.3

Base URL: https://lsio.ams3.digitaloceanspaces.com/unraid-nvidia/6-8-3/nvidia

TO AVOID CORRUPTION
DO NOT CLOSE THIS WINDOW UNTIL YOU SEE THE DONE PROMPT

Downloading: /tmp/mediabuild/bzimage ... done
Downloading: /tmp/mediabuild/bzroot ... done
Downloading: /tmp/mediabuild/bzroot-gui ...

****

 

Any idea what the problem might be?

 

Br,

 

EDIT: This seemed to be caused by the DNS server. I use a local installation of Pi-Hole as DNS. After changing to 1.1.1.1 it worked.

Edited September 15, 20206 yr by RecycledBits
Found the solution

13 hours ago, blackbunt said:

Card isn’t recognized.

What happens if you runthe below command:

watch nvidia-smi

 

Does that work?

If it does, then there is a command somewhere in this thread, posted by chbmb, that will show the UUID.

49 minutes ago, saarg said:

What happens if you runthe below command:

watch nvidia-smi

 

Does that work?

If it does, then there is a command somewhere in this thread, posted by chbmb, that will show the UUID.

Output: No devices were found

 

17 hours ago, blackbunt said:

Output: No devices were found

 

Then I don't know what could be the problem.

You could try to switch slots or try connecting a monitor and see if that helps.

22 hours ago, saarg said:

Then I don't know what could be the problem.

You could try to switch slots or try connecting a monitor and see if that helps.

Switching slots is sadly not possible, got just one. I'll try the monitor option  thx!

edit:

tried different Monitors dp and hdmi-didn‘t work. I also tried a 1660 super - also no luck. It seems somethings wrong with my mobo.

Edited September 17, 20206 yr by blackbunt

Will the next nVidia unRAID release use later drivers than 440.100?

 

I'm looking to resolve a Plex Transcoder power C-State issue that nVidia fixed in > 450.36

 

Thanks,