[Support] ich777 - AMD Vendor Reset, CoralTPU, hpsahba,...


Recommended Posts

1 hour ago, ich777 said:

I would recommend that you open up a issue over here: Click with the Diagnostics attached and the post from above linked.

 

I think the TS-470+ has another chip in it but I could be very wrong about that.

 

Happy New Year! 🎇🍾🧨🧯

Sorry I have the TVS-471 not TS-470. Opening up a ticket on github.

Edited by gregzee
  • Like 1
Link to comment

Just wanted to say a massive thank you for everything you do.

I wanted to get sys temps monitored but couldn't because I have a gigabyte board with the IT8686E chip

I tried searching everywhere and couldn't find anywhere to make it work with linux. I should learn to look at the plugins first! 

Works perfectly, thanks for all the great work!

  • Like 1
Link to comment
  • 2 weeks later...
  • 4 weeks later...

Thanks for providing the "RTL8152/3/4/6 USB Drivers" plugin. This is awesome!

 

My USB RTL8156b is used as eth0 in a failover bond0 with an integrated 1Gbit LAN as eth1.

The bond is set to active-backup... which leads to eth1 only usage.

 

After investigating with iperf3, I found out that I've "retries" when using eth0 only. So if I disconnect eth1, eth0 works somehow fine (with reties) giving me abaout 1.7 Gbits/sec. As soon as I connect my eth1 again, I get 0.95 Gbits/sec provided by eth1.

 

So it seams that I get drops on eth0 with RTL8156b. Is this a  known issue? And how should I force the bon0 to use eth0 rather then eth1?

 

Thx.

Link to comment

Hi all, just noticed my OpenRGB patch stopped working, after some troubleshooting I realised it was due to the recent setting up of my 1st ever VM (steep learning curve that 1....lol)

 

Basically once I bound my onboard Intel sound card to VFIO it stopped the patch from working, which makes total sense as SMbus & Serial Bus controller are in the same IOMMU group that is being passed through.

 

I tried a few thing to free up resources for the Patch via modifying the syslinux config, but to no avail, those being:

 

"modprobe.blacklist=i2c_i801,i2c_smbus" and unbinding the IOMMU group and binding only the sound device via "vfio-pci.ids=8086:43c8"

 

So my question is can I utilise both at the same time?

Edited by wacko37
Link to comment
4 hours ago, wacko37 said:

So my question is can I utilise both at the same time?

1st i would move to the proper forum part, this is more a virtualisation point then this plugin.

 

as starter, play with these settings and see if you can get them seperated, be aware in terms of passthrough's, the addressing will change ... so current passthrough's will most likely break and have to be reassigned.

 

also read the help text what it does ... can be useful, but not always ending up good ;)

image.thumb.png.e75fc9732743d45e38008b2249a86098.png

 

and then rather ask in the virtualisation part from the forum.

  • Thanks 2
Link to comment
15 hours ago, Jazer said:

So it seams that I get drops on eth0 with RTL8156b. Is this a  known issue? And how should I force the bon0 to use eth0 rather then eth1?

It wont be likely that you've got drops on eth0, maybe eth1 is responding quicker since it's your built in NIC and not a USB NIC (overhead) and so it uses eth1.

AFAIK a fail over bond will always use the quickest responding NIC and not the fastest one.

 

15 hours ago, Jazer said:

After investigating with iperf3, I found out that I've "retries" when using eth0 only.

This could be also the case because you have your NICs configured as a bond and it always has to try, have you yet tried to use only eth0 for testing purposes without any bond configured?

Link to comment
3 hours ago, alturismo said:

1st i would move to the proper forum part, this is more a virtualisation point then this plugin.

 

as starter, play with these settings and see if you can get them seperated, be aware in terms of passthrough's, the addressing will change ... so current passthrough's will most likely break and have to be reassigned.

 

also read the help text what it does ... can be useful, but not always ending up good ;)

image.thumb.png.e75fc9732743d45e38008b2249a86098.png

 

and then rather ask in the virtualisation part from the forum.

Thanks for the reply @alturismo and for pointing me in the right direction. 

  • Like 1
Link to comment
  • 2 weeks later...

I just upgraded to v2023.02.15 of the Intel GPU TOP plugin on unraid 6.9.2 and noticed gpu statistics wasn't working on the dashboard.

Running intel_gpu_top from the command line returned:

intel_gpu_top: /lib64/libc.so.6: version `GLIBC_2.33' not found (required by intel_gpu_top)

libc is 2.30

 

EDIT: Happened to have a backup of v2022.09.27. Working again.


 

Edited by ncohafmuta
Link to comment
26 minutes ago, ich777 said:

Please upgrade to 6.11.5 your installed version 6.9.2 is more than outdated...

 

Yeah, it's in the plans for this year, just running uptime stability testing for the last 4 months with another 2 to go. I'll just not update the plugin until then. Maybe just update the "Min OS" on the apps page?

Thanks

  • Haha 1
Link to comment
  • 2 weeks later...
15 minutes ago, positronicP said:

Does this support 13th gen Intel yet? I know there should be little change but I'm getting no stats - and it says Gen12.

Yes, because your iGPU is 12th gen and not 13th gen (intel_gpu_top always displays the generation from the iGPU and not from the CPU and yes, the can be different :D ).

 

For example my i5-10600 (10th gen CPU) has a 9th gen iGPU built in:

intel-gpu-top: Intel Cometlake (Gen9) @ /dev/dri/card0

 

Is something transcoding on your iGPU or using your iGPU? If not nothing would be displayed because the GPU is basically in render standby (RC6) as you can see on top.

 

IMC bus utilization is not supported by all CPUs, basically the same as if you have a Nvidia GPU, not all metrics are supported by all cards or better speaking by nvidia-smi.

  • Thanks 1
Link to comment

I installed amd vendor reset plugin in order to solve VM with amd gpu passthrough crashing the VM tab when I turn off or hibernate the VM. The hard power rests were the only solution to fix this issue.

 

I had this plugin installed for few days, but still unraid gui VM tab goes un-responsive once the VM in question turns off.

 

Attaching my diagnostic file. Hopefully someone can help this poor guy who is crashing his head over this.

kumonvm-diagnostics-20230307-1238.zip

Link to comment
29 minutes ago, sonisame said:

Attaching my diagnostic file. Hopefully someone can help this poor guy who is crashing his head over this.

The main issue is that the AMD Vendor Reset is a workaround and not a real solution to the cause of the issue.

I had many people where the server crashed occasionally after n amount of reboots and for some cards the AMD Vendor Reset didn't work properly...

Maybe your cards are affected and the Vendor Reset is not working...

 

The only real solution would be to sell these cards and buy something that isn't affected by the Reset Bug.

  • Like 1
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.