VMware Cloud Community
bmekler
Contributor
Contributor

Cannot VMotion some VMs despite EVC being enabled

I have a cluster of three hosts of varying ages, running ESXi 5.1 update 2 with EVC set to Nehalem, to which I added a brand new host with ESXi 5.1 update 3 and Xeon E5-v4 CPUs. I VMotioned some VMs to the new host, and now most of them won't move back to any other host in the cluster with the following error message:

The virtual machine requires hardware features that are unsupported or disabled on the target host:

* Carryless multiply (PCLMULQDQ) (see KB 1034926)

* AES instructions (AES-NI) (see KB 1034926)

If possible, use a cluster with Enhanced vMotion Compatibility (EVC) enabled; see KB article 1003212.

CPUID details: incompatibility at level 0x1 register 'ecx'.

Host bits: 0000:0000:1001:1000:0010:0010:0000:0001

Required: x1xx:xx1x:11x1:1xx0:xxxx:xx1x:xxxx:xx11

The thing is, EVC is enabled on the cluster, and the message is not even consistent - some VMs are allowed to move back to the older hosts, while others aren't. I drained one of the older hosts and upgraded it to 5.1 update 3, same build (2323236) as the new host, but it doesn't help - some VMs can move back and forth, while others can't. Amusingly enough, if I right click a VM on the new host, click migrate, and pick the new host as the validation target, it shows me the same error - for the host that the VM is currently running on.

Is there a known bug around this or something? Is there a workaround?

Edit: never mind, I figured it out - this is the proper KB article: Migrating a virtual machine configured with Virtual Hardware version 7/8 to an ESXi host with newer ...

The few VMs I had configured with hardware version 9 migrate back and forth without issues; the ones with hardware version 7 or 8 can migrate to the new host but not back.

Tags (1)
0 Kudos
1 Reply
drew_gensler
Contributor
Contributor

I'm actually having this same issue on vSphere 6.5u1e

VMware ESXi, 6.5.0, 7526125

Source Host: Dell M820 Intel(R) Xeon(R) CPU E5-4650 v2

Source Cluster:cluster3

Destination Host: Dell M830 Intel(R) Xeon(R) CPU E5-4650 v4

Destination Cluster: cluster2

EVC Mode for both clusters: Ivy Bridge

This worked up until the last set of reboots for some syslog/scratch location changes last week.

Anyone have any ideas?

0 Kudos