vSphere Upgrade Saga: It is All About EVC

I was a little confused about EVC recently, and hopefully this will clear it up for many others. My vSphere 5 Upgrade has been halted due to work, family, and my travel schedule. I wanted to put out one more staging tip for vSphere 5 with respect to EVC. This will come in handy later as I add in a new blade with a pair of Intel Westmere chips so I can play with AES&I as well as the new Intel TXT functionality. 

My issue with EVC stemmed from trying to enable it on an existing set of nodes so that I can easily transfer VMs between them. Exactly how EVC was planned to be used. However, in my environment somewhere along the line EVC was disabled. To enable it, I had to find the setting that would enable. I tried to first enable it for the chipset I had, that did not work. Then I enabled it for the previous chipset, of course that did not work. The solution was to enable it for the chipset I did not already have. In other words, I am running the chipset before Westmere, so in order for EVC to enable on my cluster, I had to set EVC mode to be Westmere chips. In other words, I am masking off all Westmere special functions.

This is what confused me, we all know we use EVC to mask off what we do not want, but if you have an existing cluster, you need to max off the chipset ABOVE the one you are on. That is, unless you desire to shutdown all your existing VMs, set EVC to the mode you desire and then power them al back on.

That was not something I was willing to do.

So, now I add to my pre-stage steps, setting EVC to mask off chipset functionality above what I already have. This will become a fairly major requirement when I add a Westmere based blade into my cluster.

Edward Haletky

Edward L. Haletky, aka Texiwill, is an author, analyst, developer, technologist, and business owner. Edward owns AstroArch Consulting, Inc., providing virtualization, security, network consulting and development and TVP Strategy where he is also an Analyst. Edward is the Moderator and Host of the Virtualization Security Podcast as well as a guru and moderator for the VMware Communities Forums, providing answers to security and configuration questions. Edward is working on new books on Virtualization.

7 thoughts on “vSphere Upgrade Saga: It is All About EVC”

  1. I’m not currently able to try this but my understanding is if that if you set a cluster to EVC mode you prefer you can add hosts to that cluster from the Datacenter level. I believe you are able to perform these steps with live VMs but have not tested as of yet.

    1. Add/Move host to Datacenter level out of a cluster.
    2. Move host to EVC enabled cluster.

    I am wondering if this step would allow you to set the EVC the way you want it, then migrate hosts in with live VMs.

    1. Hi Josh,

      It was even easier than that, I was able to successfully set the EVC mode to Westmere without doing anything on the running cluster. Since all three blades are identical, this works just fine. THe problem will be to ensure its set to Westmere on the new blade. So it is possible to set it up on a Cluster with running VMs but you need to mask off anything ABOVE your current chipset. Since I do not currently have Westmere, I masked off Westmere and viola it worked.

      Best regards,
      Edward

    2. Josh,

      You won’t always be able to add hosts into an empty cluster with any EVC mode enabled with powered on VMs. This is because the vCPU presented to your VMs already have any/all masks presented to them. In order to change those maskings (depending on the desired EVC mode) you would need to power off all VMs on that host.

      Using a different EVC mode as Edward states seems to work for that particular chipset/EVC mode, but it definitely won’t work for all. I have ran into this multiple times and no EVC modes would enable without powering off of VMs.

      1. Hello Josh,

        If all your hosts have the same chipset when you enable EVC it is an issue of finding the proper EVC mode for those hosts. In my case the EVC mode I needed was to mask off Westmere specific functionality…. This should work for all ‘like-to-like’ hosts. I.e. all mine are the same brand, chipset, CPUs, etc. If something does not match, then enabling EVC may be more difficult.

        I also did not add hosts into a new cluster. I enabled EVC on the existing cluster. Granted I am running vCenter 5, but vSphere 4, perhaps there is something new there with respect to EVC. Or perhaps since all the CPU and chipsets match, then all VMs have the same elements presented to them therefore enabling EVC was easy to do.

        My goal is to just add a new host into my EVC enabled cluster. But prior to vCenter 5, my experience was to reboot all the VMs as well. Unfortunately, I never tried to perform this action before upgrading vCenter!

        Best regards,
        Edward L. Haletky

  2. Edward,
    thank you very much, worked very well in our environment. When you are in a 24/7 environment and the VMs can not be power down, post like yours saves the day:) Richard.

  3. I’m running into the same issue as first outlined. We have a cluster consisting of Xeon E5530(Nehalem) procs and now we just purchased a new blade with X5650 (Westmere) procs. Our cluster does not currently have EVC enabled. Looking to enable it the only mode where validation succeeded is Intel Xeon Core i7 so it will admit Nehalem and Westmere processors but run them all in Nehalem mode. I’m hoping once I enable that on the existing cluster, do a few test vmotions and then add in the new host to the cluster all will be well. We have over 150 VMs running so i’m hoping from what i’m reading that we won’t have to do any shutdowns or forced reboots. I guess running the EVC in Nehalem mode means there won’t be any changes to the HALs of the guest OS? They already think they are running E5530 so no changes? Then when we get a couple more Westmere powered hosts I guess we can look at putting them in their own cluster. Sorry for the novel, any thoughts, comments are appreciated.

    G.

    1. Hello Gord,

      You will need to set the EVC mode to Westmere, which will mask OFF all higher order functionality of the Westmere (or newer chipsets) to leave behind only the Nahalem functionality. No reboot of any VM is required when you do this and therefore the HALs do not need to change as well. But do it before you add in the new Westmere server to the cluster.

      The method I outlined will allow EVC to be enabled without the need to reboot or change any VM.

      Best regards,
      Edward L. Haletky

Leave a Reply

Your email address will not be published. Required fields are marked *

eighteen − twelve =