Weird EVC mode - EVC (sandy) enabled yet VMs will not migrate across cluster (both sandy)

[jkensy]

Hi all - I recently reinstalled/rebuilt a lab VCSA to latest 6.7 U2 and current build (13843469)

I have 2 cluster "Management" and "Workload". Management has one host which is a E5-2650v2 (Ivy?) host w/ Sandy Bridge EVC enabled. Workload has two hosts one is E5-2670v0 (Sandy) and the other is E5-2650v2 (Ivy?). I have EVC mode for Sandy Bridge enabled for both clusters, as I have done for many deployments over the years! However, when I spin up a VM and try to move it from "Management" to "Workload" I get compatibility issues regarding EVC/CPU type. Makes no sense. However, my VCSA, thankfully, can bounce between both clusters. I've tried maintenance mode on all hosts, remove from cluster, re-add, turn EVC off, turn on. All VMs powered off, powered on. Made sure to power VMs on in the EVC cluster(s) so that they aren't running higher instructions than EVC mode... nada.

Anyone else having this issue? Thanks!

PS- no, it's not Per-VM EVC. That's only available to HW14+ VMs and many of my issues are on appliances like Operations Manager, etc. which are < HW10/11.

Thanks again

[rcporto]

Can you please let us know the exact build number of each ESXi host and also the exact message that you get regarding the CPU compatibility?

[jkensy]

Surely - I am running ESXi 6.7 13644319  on all hosts in question.  The message I get is:  The target host does not support the virtual machine's current hardware requirements. Use a cluster with Enhanced vMotion Compatibility (EVC) enabled to create a uniform set of CPU features across the cluster, or use per-VM EVC for a consistent set of CPU features for a virtual machine and allow the virtual machine to be moved to a host capable of supporting that set of CPU features. See KB article 1003212 for cluster EVC information. com.vmware.vim.vmfeature.cpuid.mdclear

But of course I do have EVC enabled on each cluster configured for Sandy Bridge.  The super weird part is that two of the hosts have E5-2650 v2 CPUs and even though they're the same, with the same ESXi build and same EVC mode, it reports the message above.

[rcporto]

I have seen this before, but problem was related to Spectre/Meltdown patch not present in some hosts which was running different build numbers, but in your case build numbers are the same. In that case, I would check the BIOS settings and firmware version if are the same across all three hosts.