SCVMM Hypervisor host rebooting/crashing due to excessive CPU utilization RRS feed

  • Question

  • First off, I come from a VMWare ESXi background and have little exposure to HyperV.

    There is a small 2 node cluster I inherited on SCVMM 2012 R2.  On occasion one of the hypervisor nodes will flatline at 100% CPU utilization which eventually causes it to crash.  All the VM's will then migrate to node that's left standing.  With the new load in addition to th existing load the second node will then proceed to flatline at 100% CPU and crash.  On the next cycle, the VMs will all live on one hypervisor, and the ones with resources will be online, the ones without resources will be in a failed state.

    In VMWare, if you over provisioned CPU's the VM performance would get sluggish but there would never be the case where VM usage would directly crash ESXi.  I realize HyperV is a different model in the sense that the Hypervisor is still running on top of Windows OS; but is there some control I can put into place that prevents above scenario?

    Monday, October 7, 2019 12:16 PM


All replies