unRAID OS Version:

6.2 rc4, rc5

 

Description:

Windows VMs hang on boot screen or do not start at all under 6.2 rc4 and rc5.  Receive errors related to hyperthreading not being supported on CPU.

 

How to reproduce:

Seems that this occurs with the 6.2 update with AND FX series CPUs.  Other users experiencing this bug in a thread found here:

http://lime-technology.com/forum/index.php?topic=48988.0

 

Expected results:

Would expect VM to run.

 

Actual results:

VM does not install, does not start.

 

 

Other information:

Also seems to be an issue with Plex slowing filling the memory after the 6.2 upgrade so disregard the errors related to memory in the diagnostics.

tower-diagnostics-20160914-2043.zip

Any luck if you Edit XML on one of your VMs and look for the following <cpu> block:

  <cpu mode='host-passthrough'>

 

and replace it with:

  <cpu mode='custom' match='exact'><model fallback='allow'>Opteron_G3</model>

 

I would eshultz but I had to revert back to 6.1.9 as there are too many bugs.  The upgrade was causing Plex to fill my RAM completely amd it was throwing the hyperthreading errors at my Ubuntu VM, which I use for network functions and can't afford to have down.  Check out the thread I posted of others having the same problems.  If need be I can upgrade again and do some more testing when someone will be available to work one on one, but I can't afford to have the VMs down for any length of time.  I'm disappointed as all went well as far as the upgrade was concerned, just the dockers and VMs still have a lot of issues to be corrected before this cam be a stable release.

 

Send me a PM some evening if you or any other devs want more info, but this appears to be a widespread problem.

 

Sent from my SM-N910V using Tapatalk

 

 

...just the dockers and VMs still have a lot of issues to be corrected before this cam be a stable release.

What issues with dockers? (Saves me from having to search out your posts  )

Plex filling up the memory and it broke my openvpn docker.  Those are the only two I use.

 

Sent from my SM-N910V using Tapatalk

 

 

I updated to 6.2 and started fresh.  I got windows 10 to install with the cpu in host passthrough mode and it ran fine for a day and then back to the usual errors:

 

internal error: process exited while connecting to monitor: warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 0]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 1]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 2]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 3]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 4]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 5]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 6]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 7]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 8]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 9]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 12]

warning: host doesn't support requested feature: CPUID.80000001H:EDX [bit 13]

 

Windows 7 will not even install, it hangs on the splash screen.  Has anyone done any testing to resolve this issue?  It is becoming extremely frustrating.  The <cpu mode='custom' match='exact'><model fallback='allow'>Opteron_G3</model> edit does get Windows 10 working but doesn't help the Windows 7 vm.  I'd be glad to do some testing but I have no idea on where to start to find the problem.  What has changed from 6.1.9 to cause this?

Hi

 

any sign of this being looked into, I have rebuilt with different versions of 6.2 and not having any luck.

 

I have tried the solution given by eschultz and the error messages about hyper-threading disappear in the logs and all looks to be fine but no pass through monitors are working on either GPU. Had a glimmer of it on one of the GPU but then went through the windows 10 installer, got the drivers for the storage device and despite saying have 100GB of disk it showed 0b available in windows 10.

 

Almost to the point of giving up and splitting back into two separate machines, had high hopes of using this as both a NAS and 2 Windows 10 machines in one box and playing with dockers.

Hi

 

any sign of this being looked into, I have rebuilt with different versions of 6.2 and not having any luck.

 

I have tried the solution given by eschultz and the error messages about hyper-threading disappear in the logs and all looks to be fine but no pass through monitors are working on either GPU. Had a glimmer of it on one of the GPU but then went through the windows 10 installer, got the drivers for the storage device and despite saying have 100GB of disk it showed 0b available in windows 10.

 

Almost to the point of giving up and splitting back into two separate machines, had high hopes of using this as both a NAS and 2 Windows 10 machines in one box and playing with dockers.

 

We're about ready to release a new 'mainline' which has large number of package updates, including qemu.  Let's see how things once we're on the latest code before diving into this issue further.

Hi

 

any sign of this being looked into, I have rebuilt with different versions of 6.2 and not having any luck.

 

I have tried the solution given by eschultz and the error messages about hyper-threading disappear in the logs and all looks to be fine but no pass through monitors are working on either GPU. Had a glimmer of it on one of the GPU but then went through the windows 10 installer, got the drivers for the storage device and despite saying have 100GB of disk it showed 0b available in windows 10.

 

Almost to the point of giving up and splitting back into two separate machines, had high hopes of using this as both a NAS and 2 Windows 10 machines in one box and playing with dockers.

 

We're about ready to release a new 'mainline' which has large number of package updates, including qemu.  Let's see how things once we're on the latest code before diving into this issue further.

 

Any idea when?  I just reverted back to 6.1.9 and magically all the VMs work without issue.  I'd be glad to test out a beta and report back any findings, considering there are no AMD test servers.

Hi

 

any sign of this being looked into, I have rebuilt with different versions of 6.2 and not having any luck.

 

I have tried the solution given by eschultz and the error messages about hyper-threading disappear in the logs and all looks to be fine but no pass through monitors are working on either GPU. Had a glimmer of it on one of the GPU but then went through the windows 10 installer, got the drivers for the storage device and despite saying have 100GB of disk it showed 0b available in windows 10.

 

Almost to the point of giving up and splitting back into two separate machines, had high hopes of using this as both a NAS and 2 Windows 10 machines in one box and playing with dockers.

 

Can you post some system specs as well as what you are trying to do?

An update to this issue.  In the latest RC, code has been put in to detect the case of an AMD CPU and to set the CPU topology to not use multiple threads in this instance (only multiple cores).  However, you will need to click "edit" and "update" on any of your VMs for this change to take affect.

An update to this issue.  In the latest RC, code has been put in to detect the case of an AMD CPU and to set the CPU topology to not use multiple threads in this instance (only multiple cores).  However, you will need to click "edit" and "update" on any of your VMs for this change to take affect.

So the AMD FX CPUs issue is fixed in 6.2.4?

 

I tried back when 6.2.0 first came out but had to immediately downgrade back to 6.1.9.  One weird thing in the downgrade back to 6.1.9 is still seeing the GUI line in the boot menu.

 

 

Sent from my XT1563 using Tapatalk

 

 

Anyone else having issues with AMD and having a plex docker and Win 10 VM auto-starting and fighting for resources? Docker isn't isolated to cores but the VM is set for the last 6 out of 8 on my 8350. Literally my webui hangs and I have to manually kill plex for my VM to properly start, then the webui becomes responsive and all is well and I can restart my plex docker.

unraid-box-diagnostics-20161116-2029.zip

Anyone else having issues with AMD and having a plex docker and Win 10 VM auto-starting and fighting for resources? Docker isn't isolated to cores but the VM is set for the last 6 out of 8 on my 8350. Literally my webui hangs and I have to manually kill plex for my VM to properly start, then the webui becomes responsive and all is well and I can restart my plex docker.

Not the "real" solution, but this plugin will let you still have Plex and your VM autostart, but will let you delay the start of Plex for however long you choose (ie: the average number of seconds it takes for your VMs to get up and running)

 

http://lime-technology.com/forum/index.php?topic=53731.0

I have successfully installed Win7 and WinXP VMs on an AMD FX-4130.

 

In order to get passed the Windows boot logo, you need to select the i440fx-2.5 machine, SeaBIOS and VNC cirrus or vmvga video driver (cirrus won't allow more tha 1024x768 while vmvga allows larger resolutions). If you select the default one (QXL) or a Q35 machine it will not boot.

 

I am adding my settings in a screenshot if that can help!