I keep getting the following message to the shell.

Any ideas on the cause?

 

Message from syslogd@phab1 at Dec 26 19:16:16 ...
kernel:Uhhuh. NMI received for unknown reason 30 on CPU 0.

Message from syslogd@phab1 at Dec 26 19:16:16 ...
kernel:Do you have a strange power saving mode enabled?

Message from syslogd@phab1 at Dec 26 19:16:16 ...
kernel:Dazed and confused, but trying to continue

sir1-diagnostics-20161226-1523.zip

@d.bech

im getting the same exact error. Any luck in finding a solution? The VM does seem to work fine though.

 

What are you using your CentOS vm for?

I haven't found a solution yet.

I am using CentOS 7.

 

For sure doesn't seem to affect the running of the VM.

 

It seems to come and go.

 

What hardware are you using?

 

I'm using a HP G6 DL360

I have seen the same on a Fedora 25 VM

Hi, i encountered this issue when passing through a usb pcie controller to the VM's

https://www.amazon.co.uk/gp/product/B00JEVLEFQ

 

the VM's start showing this message, for various ID's (10, 20, 30, 21, 31), 5 minutes after VM started.

 

Removed the passthrough of this card from the vm definition, and message disappeared.

 

i will update if removing the usb devices from the pcie usb controller card is helping...

 

-d

 

 

 

 

 

hi,

 

i bought a pcie riser x1 and moved the usb pcie card from the x16 slot (the one receiving lanes from CPU) to a x1 pcie slot (the small slot under my primary 2-slot wide gpu, closest to the CPU).

 

Now after this change, the following happens:

1) no more such errors as reported in this post on the guest VM's that have the pcie usb card passed though. 

2) i can passthrough this usb pcie card to my Windows 10 VM (in the past it was not possible, the windows vm was not initiating, but it was working only in Linux VM's)

3) my windows 10 VM can start with more than 4 cpu's allocated (in the past allocating more than 4 would fail to load the windows 10 VM) - i tried now with 3:2 and 4:2 and it's working fine

 

my understanding is that some pcie slots are using the PCIe lanes from CPU while other slots (x4, x1) from the motherboard chipset.

due to my case and motherboards space constraints, i initially placed this usb pcie card in the x16 slot that usually is for dual sli.

And although it was having different iommu group from the primary GPU, somehow it was causing these issues.

 

Now i placed my 2 GPU's in the apropriate x16 slots and took away this usb pcie card with the help of the x1 riser.

Passthrough of the 2 GPUs is fine and no issues encountered so far.

-d

Fedora 25 issue is pretty annoying. Wish there was a newer libvirtd/qemu/kvm binaries that may fix this.