Link to home
Start Free TrialLog in
Avatar of maxihost
maxihost

asked on

vMotion CPU Modifying CPUID masks

Hello,

I have a cluster running four hosts using Westmere Intel processors (5630). I have added a server using processors 5530 in the same cluster.

When testing the vMotion I received an error regarding CPU compatibility.

I have found two knowledbase topics:
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1993
and
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1991

I have changed the "ecx" for a guest to:
---- --0- ---- ---- ---- ---- ---- --0-
as described at the topic:

Group D
Models include:
         
Intel Xeon CPUs based on Intel Nehalem microarchitecture. For example, Intel Xeon 75xx Series, Intel Xeon 65xx Series, Intel Xeon 55xx Series ("Nehalem-EP"), and Intel Xeon 34xx Series ("Lynnfield").

Group E
Models include:
         
Intel Xeon CPUs based on Intel Westmere microarchitecture. For example, Intel Xeon 56xx Series ("Westmere-EP").

For D<->E VMotion, apply AES mask. (Not supported prior to ESX 3.5. Experimentally supported for ESX 3.5 and later only.)

AES      a      ecx      ---- --0- ---- ---- ---- ---- ---- --0-

After changing those values, the error is the same.

Any ideas ?
ASKER CERTIFIED SOLUTION
Avatar of bgoering
bgoering
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of maxihost
maxihost

ASKER

Hello,

Its giving me the following error:


The cluster cannot be configured with the selected Enhanced vMotion Compatibility mode; CPU features disabled by that mode may currently be in use by powered-on or suspended virtual machines in the cluster.

SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Oh my god. I have 75 VMs running. Ok. Thanks for the solution.
LOL - I feel your pain. It has to be that way because vms that may (or may not) be using the features of the newer processor can become unstable when migrated to a cpu that doesn't support those features. You may be able to lessen the pain by moving all of your westmere hosts out of the cluster then setting the evc level.

Next you can power off and migrate vms one at a time back to your host in the cluster until one of your westmere hosts is empty, then add it back to the cluster where it will respect the new evc mode.

Repeat for your other three westmere hosts... You still have to power off each vm, but doing it this way they don't have to all be down at the same time.

Good Luck