[Plugin] Nvidia-Driver

[Milk.Daddy]

1 hour ago, ich777 said:

Please post your Diagnostics.

 

tower-diagnostics-20211230-0223.zip

[ich777]

10 minutes ago, Milk.Daddy said:

tower-diagnostics-20211230-0223.zip 82.21 kB · 0 downloads

This is really strange, I see the card recognized in you system overview but the driver doesn't even recognizes the card.

Have you tried to connect a display to the card and see if you get a output from the card?

Does the board deliver enough power to the PCIe slot where the card is installed in? Can you maybe try another PCIe slot?

 

Please double check if you are enabled Above 4G Support and Resizable BAR support in your BIOS.

Also make sure that you are on the latest BIOS version and update it if a newer version is available.

[thymon]

Hello All,

 

I'have 2 GPU

- RTX 3060 12Go to Emby Transcode with this plugin

- RTX 2070 Super to VM Gaming.

PSU :850W

 

Very regularly the RTX 3060 disappear... I don't know why. Need to reboot Unraid. 

 

 

 

At the moment, I have to restart the server almost every day.

What to do ? I can't find anything on Google ...

Edited December 30, 2021 by thymon

[ich777]

1 hour ago, thymon said:

At the moment, I have to restart the server almost every day.

Can you post your Diagnostics please?

 

Are you running a Ryzen based server?

[thymon]

@ich777 thank you for your interest. 

 

Yes, it is a Ryzen based server.

AMD Ryzen 9 3900X 

 

So far I had a GPU for transcoding and had never had a problem.

It's since the add of the RTX3060 for transcoding instead of RTX2070.

 

Diag in attachment.

unraid-diagnostics-20211230-2152.zip

Edited December 30, 2021 by thymon

[ich777]

34 minutes ago, thymon said:

@ich777 thank you for your interest. 

 

Yes, it is a Ryzen based server.

AMD Ryzen 9 3900X 

 

So far I had a GPU for transcoding and had never had a problem.

It's since the add of the RTX3060 for transcoding instead of RTX2070.

 

Diag in attachment.

Please see this post:

 

@HellraiserOSU reported that turning on IOMMU in the BIOS solved the issue but other users reported turning off IOMMU in the BIOS solves the issue.

Please also consider disabling C-States in the BIOS, also enable 4G Above Decoding and Reiszable BAR support in your BIOS.

Make also sure that you are on the latest BIOS version.

[thymon]

@ich777 If I turn-off IOMMU, I will cannot use my other GPU in passtrough, right ?

[ich777]

4 hours ago, thymon said:

@ich777 If I turn-off IOMMU, I will cannot use my other GPU in passtrough, right ?

Should be possible.

I would try it since you can always revert back.

[alturismo]

4 hours ago, thymon said:

@ich777 If I turn-off IOMMU, I will cannot use my other GPU in passtrough, right ?

 

exactly, so prolly no option in your case ... its sad enough but these issues are more really AMD Systems related, its really a trial & error forwarding with different BIOS settings like @ich777 mentioned which are the base settings, really turn all energy savers off and so on ... hopefully you find a working combination.

[thymon]

@ich777

No more issue for the moment.  
I changed two things. One of twice is probably the solution.
 

-1: Removed the "GPU stat" plug-in

-2: Added an HDMi dongle.

 

wait & see…

thank a lot !

Edited January 1, 2022 by thymon

[Commuting-Reliably6-Hardsh]

i have the error:

 

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.:

 

what can i do to try to make it work ?

 

[alturismo]

53 minutes ago, Commuting-Reliably6-Hardsh said:

what can i do to try to make it work ?

 

may post a screen from your vfio's from tools, system devices ...

 

sample

 

and i assume your card is currently NOT used by any VM.

[Commuting-Reliably6-Hardsh]

 

43 minutes ago, alturismo said:

 

may post a screen from your vfio's from tools, system devices ...

 

sample

 

and i assume your card is currently NOT used by any VM.

no it is not used by anything i plugged it and tried every driver that's about it.

[alturismo]

12 minutes ago, Commuting-Reliably6-Hardsh said:

no it is not used by anything i plugged it and tried every driver that's about it.

 

ok, then may take a look at page 1 under troubleshooting and follow the "supported" cards ... and corresponding drivers.

[Commuting-Reliably6-Hardsh]

46 minutes ago, alturismo said:

 

ok, then may take a look at page 1 under troubleshooting and follow the "supported" cards ... and corresponding drivers.

i did and it is supported by all 470 drivers

[alturismo]

13 minutes ago, Commuting-Reliably6-Hardsh said:

i did and it is supported by all 470 drivers

 

ok, then ... when i see your iommu, do use downstream, multi function as ACS Override ?

 

if so, may try turn this off

 

next would be, try another pcie slot on your board

 

also, take a look at the bios setting if you boot in uefi, may try change this (or vice vers)

 

not much more to say ... as it looks all good

[Commuting-Reliably6-Hardsh]

11 minutes ago, alturismo said:

ok, then ... when i see your iommu, do use downstream, multi function as ACS Override ?

how may one do this ?

[alturismo]

4 minutes ago, Commuting-Reliably6-Hardsh said:

how may one do this ?

 

if its disabled, keep it disabled, just looks almost like its enabled

 

 

best bet, try another slot and/or change boot type

[Commuting-Reliably6-Hardsh]

1 hour ago, alturismo said:

 

if its disabled, keep it disabled, just looks almost like its enabled

 

 

best bet, try another slot and/or change boot type

yes it was disabled. i removed the extra 2 network adapter leaving only the 2 raid controller and placed the gpu on another one
plugged a screen to check that i have a display output from the gpu and it works.

 

still got the error...

[ich777]

49 minutes ago, Commuting-Reliably6-Hardsh said:

still got the error...

Please post your Diagnostics.

[Commuting-Reliably6-Hardsh]

2 minutes ago, ich777 said:

Please post your Diagnostics.

here it is

tower-diagnostics-20220103-1931.zip

[ich777]

1 hour ago, Commuting-Reliably6-Hardsh said:

here it is

I'm very sorry but the driver reports this:

Jan  3 18:37:33 Tower kernel: NVRM: The NVIDIA GeForce GT 730 GPU installed in this system is
Jan  3 18:37:33 Tower kernel: NVRM:  supported through the NVIDIA 390.xx Legacy drivers. Please
Jan  3 18:37:33 Tower kernel: NVRM:  visit http://www.nvidia.com/object/unix.html for more
Jan  3 18:37:33 Tower kernel: NVRM:  information.  The 470.94 NVIDIA driver will ignore
Jan  3 18:37:33 Tower kernel: NVRM:  this GPU.  Continuing probe...
Jan  3 18:37:33 Tower kernel: NVRM: No NVIDIA GPU found.

 

From what I've found online there are a few cards out there who are not compatible, is it this card: Click

If yes, you are may be out of luck.

There are "old" and "new" GT 730's out there and only the "new" ones are supported.

[Commuting-Reliably6-Hardsh]

2 minutes ago, ich777 said:

I'm very sorry but the driver reports this:

Jan  3 18:37:33 Tower kernel: NVRM: The NVIDIA GeForce GT 730 GPU installed in this system is
Jan  3 18:37:33 Tower kernel: NVRM:  supported through the NVIDIA 390.xx Legacy drivers. Please
Jan  3 18:37:33 Tower kernel: NVRM:  visit http://www.nvidia.com/object/unix.html for more
Jan  3 18:37:33 Tower kernel: NVRM:  information.  The 470.94 NVIDIA driver will ignore
Jan  3 18:37:33 Tower kernel: NVRM:  this GPU.  Continuing probe...
Jan  3 18:37:33 Tower kernel: NVRM: No NVIDIA GPU found.

 

From what I've found online there are a few cards out there who are not compatible, is it this card: Click

If yes, you are may be out of luck.

There are "old" and "new" GT 730's out there and only the "new" ones are supported.

oh this is so sad

Would it be possible to install the 390.xx as recommended or should i just bite it and find another gpu ?

[ich777]

1 minute ago, Commuting-Reliably6-Hardsh said:

Would it be possible to install the 390.xx as recommended or should i just bite it and find another gpu ?

For what would you use it with the legacy drivers? This card isn't even capable of transcoding most of the things out there today.

Also the legacy drivers won't work properly with Docker container in general.

 

I can recommend you to get a Nvidia T400 you can get them brand new for about $130,- (2GB Turing card).

[milfer322]

I installed that plugin with Unraid 6.9.2 and restart the server.

After restart dont start the docker service.

What can I do?
I tried disable and enable the service via gui, and restart the server and  uninstall plugin and restart server.

 

Thanks so much!!

 

Edited January 3, 2022 by milfer322