Single Dell Host Refusing to Boot VMs with EVC Compatibility

[behbhelpdesk]

I have a simple 3 Host vSphere Essentials Plus Cluster. I recently enabled "Sandy Bridge" for compatibility between these three hosts:

VMHOST1 -  Xeon E5-2470

VMHOST2 -  Xeon E5-2470

VMHOST3 -  Xeon E5-2680 v4

When I set EVC in the cluster settings it validated and set them. All individual VMHOSTs show "Sandy-Bridge" for VMware EVC Mode (see attached). When I look at individual booted VMs on VMHOST1 and VMHOST2, they show either Westmere or Sandy Bridge in their EVC Mode. When I look at booted VMs on VMHOST3, they show "N/A" for EVC Mode.

I have tried adding and removing this host with VMs off to another cluster with the same setting, etc. and am getting the same result.

I looked at the Dell BIOS setting and VT is enabled, but I do not see this option: "No-Execute Memory Protection" in the BIOS anywhere (see attached) as recommended here. I'm at a loss here. We are at ESXi 5.1 U3 build number 3872664.

I'm at a loss at this point.

I can live migrate between VMHOST1 and VMHOST2. I can live migrate from VMHOST1/2 to VMHOST3. I cannot live migrate VMHOST3 to either of the other two hosts.

Any ideas?

[MattiasN81]

Do you see "vmx| CPUID differences from hostCPUID" in the VMs vmware.log file?

[behbhelpdesk]

MattiasN81​,

Clarifying question, the vmware.log of the individual VMs that I am running on VMHOST3?

- Elan

[RajeevVCP4]

These are jpg file not vmware.log

but you can try this KB..

EVC mode mismatch causes virtual machine migration issues (2014835) | VMware KB

[behbhelpdesk]

Thanks Rajeev for your help. Unfortunately, that didn't resolve the issue.

This is happening for every VM on VMHOST3.

[behbhelpdesk]

MattiasN81​,

Yes I do.

28T19:40:25.415Z| vmx| I120: CPUID differences from hostCPUID.

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[1] level 00000001, 0: 0x000406f1 0x01200800 0x77fefbff 0xbfebfbff

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[1] level 0000000b, 0: 0x00000001 0x00000002 0x00000100 0x00000001

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[1] level 0000000b, 1: 0x00000005 0x0000001c 0x00000201 0x00000001

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[1] level 0000000b, 2: 0x00000000 0x00000000 0x00000002 0x00000001

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[1] level 0000000d, 0: 0x00000007 0x00000240 0x00000340 0x00000000

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[2] level 00000001, 0: 0x000406f1 0x02200800 0x77fefbff 0xbfebfbff

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[2] level 0000000b, 0: 0x00000001 0x00000002 0x00000100 0x00000002

2017-03-28T19:40:25.415Z| vmx| I120: CPUID[2] level 0000000b, 1: 0x00000005 0x0000001c 0x00000201 0x00000002

What does that mean?

[behbhelpdesk · Accepted solution]

We figured out the issue. Our processor was so new that ESX 5.1U3 didn't support. We had to upgrade the host to 5.5U3 in order to get EVC support for this generation.