8 Posts
0
12668
The Target Host Does Not Support The Virtual Machine's Current Hardware Requirements
Hello everyone,
We currently have a 4 host VxRail Cluster. Unfortunately one of the hosts failed, and we weren't able to get a duplicate server so we do have a server with slightly different hardware:
Host 1: VxRail S570 Intel(R) Xeon(R) Silver 4210 CPU @ 2.20GHz
Hosts 2-4: VxRail S470 Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz
After working with Dell, everything had seemed fine and the host is now fully incorporated into the cluster. However, I just ran VxVerify to see if we can upgrade from v4.7.530 to v4.7.541. There are a few virtual machines that are reporting the following error message:
The target host does not support the virtual machine's current hardware requirements.
Majority of these VMs were migrated from a completely different host (not the VxRail) onto the VxRail relatively recently. There are a few that I had built on the VxRail and I'm not sure why it is reporting this error message. I did attempt to look into VMWare EVC, and attempted to look into setting it into Sandy Bridge (since that's what the three other hosts are?) but it reports that I can't because there are VMs that are online that don't support that. There is only one setting that only has host 1 with that error, which is Broadwell.
What are my options here in order to reduce the downtime of machines as much as possible, so that I can successfully migrate these machines without issue to the best supported setting?
DELL-Naoyuki K
4 Operator
4 Operator
•
1.8K Posts
0
March 16th, 2022 07:00
https://docs.vmware.com/en/VMware-vSphere/7.0/com.vmware.vsphere.vm_admin.doc/GUID-EE6F4E5A-3BEA-43DD-9990-DBEB0A280F3A.html
Maybe it's because of the virtual hardware version of each VM, if you can configure it for some VMs but not for others in the same cluster.
The virtual hardware version must be 14 or upper as requirement of per-vm EVC.
So first thing to do is to check the hardware version of each VM and upgrade it if needed.
Note that upgrading virtual hardware version is one way operation, i highly recommend you to take a shapshot of each VM before upgrading.
DELL-Naoyuki K
4 Operator
4 Operator
•
1.8K Posts
0
March 15th, 2022 08:00
Let me know following.
- Current EVC mode (CPU generation) of the cluster.
- you can migrate those affected VMs by vMotion from current host to another or not.
In my experience, this error message can be caused by CPU spec mismatch.
And S570 does have higher CPU generation than S470.
If there is no EVC on the cluster, and those affected VMs are all on the host 1 (S570). I would use per-VM EVC for those VMs.
If those affected VMs are all on S470 nodes, and can migrate among S470 nodes by vMotion, you can ignore the error message because VxRail Upgrade put it in maintenance mode only one ESXi at a time.
Punkrulz24
8 Posts
0
March 15th, 2022 09:00
Hey Naoyuki,
- Current EVC Mode: Presently disabled.
- Unable to migrate those VMs due to the following error message:
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.misc.mds_no
com.vmware.vim.vmfeature.misc.rsba_no
com.vmware.vim.vmfeature.misc.ibrs_all
com.vmware.vim.vmfeature.misc.rdcl_no
- I was investigating the Per-VM EVC. Majority of our VMs in our infrastructure all say that they are Intel Broadwell Generation except for of course the ones that are missing this. Another strange thing: Many VMs, including those that say they are "Intel Broadwell Generation" are missing the VMWare EVC configuration option... and I can't figure out why. It doesn't appear to be power status, nor what version of OS.
- All devices appear to be on the S570 node.
DELL-Naoyuki K
4 Operator
4 Operator
•
1.8K Posts
0
March 15th, 2022 19:00
Looking into the info you provide, it looks like caused by CPU feature difference between source host and target host.
all the feature specified by vmotion check are related with IA32_ARCH_CAPABILITIES of the CPU.
So i'm not completely sure but it looks like can be resolved by Per-VM EVC (needs shutdown VM), unless you did tweak CPU masking manually for the VMs or ESXis.
https://www.intel.com/content/www/us/en/developer/articles/technical/software-security-guidance/technical-documentation/cpuid-enumeration-and-architectural-msrs.html
Punkrulz24
8 Posts
0
March 16th, 2022 04:00
Hey Naoyuki,
Since I'm unfamiliar with any sort of CPU Masking, I can pretty much guarantee that we did not tweak that manually for the VMs or ESXis. The biggest issue we have here is that there is no Per-VM EVC Setting for many of the PCs; it is just completely missing, and I don't know why. Even with machines powered off, the setting isn't there for a lot of them. I don't know the rhyme or reason as to why it would appear on some machines and disappear for most.
Punkrulz24
8 Posts
1
March 20th, 2022 10:00
Naoyuki,
Thank you for the wonderful information. We were able to successfully upgrade the VM hardware of the machines in question, then enable per-VM EVC. Once that was enabled on all of them, I was able to easily change the EVC mode of the entire cluster to Broadwell and everything is working well. Thank you!
reneeF
1 Message
0
October 27th, 2022 15:00
did you have to reboot each VM after upgrading hardware?