ldornak
Contributor
Contributor

Can't Vmotion to EVC Cluster

I'm receiving the following error when I try to migrate VMs from a non-EVC cluster to a EVC enabled cluster:

Unable to migrate from host1 to host2: Host CPU is incompatible with the virtual machine's requirements at CPUID level 0x80000001 register 'edx'. host bits: 1110:0011:1101:0011:1111:1011:1111:1111 required: 11x0:1x11:1101:0011:1111:x0x1:1111:1111 Mismatch detected for these features: *RDTSCP; refer to KB article 1993 for a possible solution.

Current setup:

  • VC 2.5u2

  • ESX 3.5u2

  • 3 - DL585 G2s (2 servers with 8218 CPUs, 1 server with 8220 CPU)

  • No-Execute Page-Protection - Enabled

  • AMD V - Enabled

  • Vmotion currently works between all hosts in non-EVC cluster

  • No masking is setup on any of the VMs

Here's my method:

  • Create new AMD EVC enabled cluster

  • Put host1 in mainenance mode

  • Move host1 from non-EVC cluster to EVC enabled cluster

  • Exit maintenance mode on host1

  • Attempt to migrate several VMs to the EVC enabled cluster and receive above error on each attempt.

Am I missing something? Is there anyway to migrate to an EVC enabled cluster without shutting down all VMs?

0 Kudos
5 Replies
stvkpln
Virtuoso
Virtuoso

What's the compatibility configuration set to in EVC?

-Steve
0 Kudos
ldornak
Contributor
Contributor

Here's the details of the EVC setting:

Applies baseline feature set of AMD Second Generation Opteron processors to all hosts in this EVC cluster. "AMD Second Generation Opteron" refers to AMD Opteron processors, such as Rev. E and Rev. F, that belong to the K8 family

0 Kudos
jandie
Enthusiast
Enthusiast

Are you still having issues with this? If you're not, how did you resolve it.

I'm having similar issues but only to the Windows 2008 guests.

0 Kudos
ldornak
Contributor
Contributor

Yes, I still have the issue. I have not received any resolution, but I do have a window around x-mas where I can shutdown all my VMs and enable EVC on the cluster.

0 Kudos
Rhodan
Contributor
Contributor

We had to shutdown the VM's and cold migrate them.

0 Kudos