ESXi host fails with a PSOD due to an Intel Virtualization Technology!
I have read a KB on VMware Knowledge Base and it says that an Intel Virtualization Technology can be cause of PSOD.
This is little funny because ESXi will be affected by wide range of Intel Xeon processor family:
- Intel® Xeon® Processor 55xx Series
- Intel® Xeon® Processor 56xx Series
- Intel® Xeon® Processor 65xx Series
- Intel® Xeon® Processor 75xx Series
- Intel® Xeon® Processor E5-1400 v2 Product Family
- Intel® Xeon® Processor E5-1600 v2 Product Family
- Intel® Xeon® Processor E5-1600 v3 Product Family
- Intel® Xeon® Processor E5-2400 Product Family
- Intel® Xeon® Processor E5-2400 v2 Product Family
- Intel® Xeon® Processor E5-2600 Product Family
- Intel® Xeon® Processor E5-2600 v2 Product Family
- Intel® Xeon® Processor E5-2600 v3 Product Family
- Intel® Xeon® Processor E5-2600 v4 Product Family
- Intel® Xeon® Processor E5-4600 Product Family
- Intel® Xeon® Processor E5-4600 v2 Product Family
- Intel® Xeon® Processor E5-4600 v3 Product Family
- Intel® Xeon® Processor E5-4600 v4 Product Family
- Intel® Xeon® Processor E7-2800 Product Family
- Intel® Xeon® Processor E7-4800 Product Family
- Intel® Xeon® Processor E7-8800 Product Family
- Intel® Xeon® Processor E7-8800/4800/2800 v2 Product Families
- Intel® Xeon® Processor E7-8800/4800 v3 Product Families
- Intel® Xeon® Processor E7-8800/4800 v4 Product Families
There is a workaround for preventing the problem and PSOD on server but VMware recommends to contact with your vendor hardware.
Here is some lines from the article:
Most Intel® Xeon® processor based platforms incorporate a work around for the erratum. However, some do not and may require additional action by the customer as described:
- For Intel® Xeon® processor based platforms with 255 or less logical processors (local APIC in xAPIC mode) VMware recommends to add the iovDisableIR=TRUE boot option to the ESXi host that disables the use of Intel® VT-d interrupt remapperNote: VMware does not expect this boot option to have any negative functional or performance effects.
- For Intel® Xeon® processor based platforms with more than 255 logical processors (local APIC in x2APIC mode) VMware recommends that customers experiencing this issue should contact the server vendor to ensure the appropriate erratum work around.
