An update from Vmware. My SR#1101983251. Wondering if anyone in Vmware engineering can pull some strings and get me an updated kernel RPM w/ the micro-code update? I can easily verify wether the issue is fixed in my lab environment.
"Hi, Greg,
Oh, it is not a good news. I looked through your newest log , it shows the same error as previous 7 PSOD instances.
ESC[31;1m21:05:47:35.224 cpu2:1101)ALERT: MCE: 169: Machine Check Exception:
General Status 0000000000000005ESC[0m
ESC[31;1m21:05:47:35.224 cpu3:1406)ALERT: MCE: 169: Machine Check Exception:
General Status 0000000000000004ESC[0m
ESC[31;1m21:05:47:35.224 cpu2:1101)ALERT: MCE: 193: Machine Check Exception:
Bank 0, Status b200004000000800ESC[0m
ESC[31;1m21:05:47:35.224 cpu3:1406)ALERT: MCE: 193: Machine Check Exception:
Bank 0, Status b200004000000800ESC[0m
Panic from another CPU (world 1406): ip=0x62583c Machine Check Exception: Unable to continue
The MCE code : b200004000000800 indicates that no micro-op retired in a given set of time.
A request from the CPU that goes to bus that does not complete within a few seconds can cause the timeout failure.
The kinds of requests that usually fit this pattern are reads of I/O ports or MMIO locations. This error is also something that software can cause: by reading wrong memory locations or trying to read from a non-existent I/O port.
Progress from engineering : Intel has revised the micro-code for Harpertown (54xx Xeon) for us to test last night (Jan 22nd), the test is in progress now.
What I can think of , at current stage , is to change the CPU to Clovertown (53xx Xeon) inside your SuperMicro chassis since it is the one on the HCL list and the engineering can't simulate the above issue with this family of Xeon. If this alternative is not feasible , we will have no choice but to wait for the engineering. I will advise you NOT to put production VMs on the ESX that is supermicro/with 5425 CPU. On my side, I will keep on eye on the engineering progress and keep you updated."
