A hotfix has been released for Windows Server 2008 R2 Hyper-V for when computers that use AMD central processing units hang or restart during the Live Migration process.
This issue is caused by caused by AMD erratum 383. For detailed information, see the “Symptoms,” “Cause,” and “Resolution” sections of this article.
The article explains that:
When you run Windows Server 2008 R2 Hyper-V on a computer that uses AMD Family 10h central processing units (CPUs), the host computer may restart or stop responding unexpectedly. When this issue occurs, you receive no error messages.
This issue may occur when the Live Migration feature is used to move a virtual machine (VM) from one Windows Server 2008 R2 Hyper-V-based node to another node. This problem usually affects the source node that is the original location of the VM and usually does not affect the target node that is the new location where the VM will be moved.
Note This issue may occur in other scenarios and does not occur only in Live Migration scenarios.
This issue occurs because of erratum 383 in AMD Family 10h processors. This erratum generates a machine check exception. This machine check exception causes the hardware to restart or to stop responding.
A hotfix has been released to fix this issue.