No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

BSoD Occurs After Hyper-V Is Installed on Two CH222 V3 Compute Nodes (Windows Server 2016)

Publication Date:  2019-04-10 Views:  36 Downloads:  0
Issue Description

A customer installs Hyper-V and then two Microsoft patches on two CH222 V3 compute nodes (Windows Server 2016). Blue screen of death (BSoD) occurs in less than two hours. The patches are KB4093137 and KB4093119


Solution

1. Use WinDbg to open the memory dump files. The memory dump files show that the BSoD problem is related to vmswitch, which is the virtualization driver provided by Windows.

2. Microsoft provides the following memory dump analysis.

Bug check information:

====================

  14: kd> kL

# Child-SP          RetAddr           Call Site

00 ffff9f01`e68955c8 fffff803`2538be66 nt!KeBugCheckEx

01 (Inline Function) --------`-------- nt!PspUnhandledExceptionInSystemThread+0x27

02 ffff9f01`e68955d0 fffff803`25363c3f nt!PspSystemThreadStartup$filt$0+0x44

03 ffff9f01`e6895610 fffff803`2537cf1d nt!__C_specific_handler+0x9f

04 ffff9f01`e6895680 fffff803`252bd011 nt!RtlpExecuteHandlerForException+0xd

05 ffff9f01`e68956b0 fffff803`252bbe34 nt!RtlDispatchException+0x421

06 ffff9f01`e6895db0 fffff803`25386c82 nt!KiDispatchException+0x1e4

07 ffff9f01`e6896460 fffff803`25383274 nt!KiExceptionDispatch+0xc2

08 ffff9f01`e6896640 fffff80e`968e4651 nt!KiPageFault+0x4f4

09 ffff9f01`e68967d0 fffff80e`96906f52 vmswitch!VmsPtAdjustNumQueuePairsInUse+0x19

0a ffff9f01`e6896810 fffff80e`968e4e19 vmswitch!VmsIovDeleteVPort+0x27e

0b ffff9f01`e68968e0 fffff80e`9690a144 vmswitch!VmsVmqPvtDeleteVmq+0x179

0c ffff9f01`e68969b0 fffff80e`968e02d3 vmswitch!VmsVmqPvtDeleteVmqOnNic+0xac

0d ffff9f01`e68969f0 fffff80e`968e5f49 vmswitch!VmsVmqDoVmqOperation+0xcb

0e ffff9f01`e6896a50 fffff80e`968e5a52 vmswitch!VmsPtNicMuxManageResources+0x40d

0f ffff9f01`e6896ad0 fffff803`252b6511 vmswitch!VmsPtNicReallocateResourceWorkerRountine+0xb2

10 ffff9f01`e6896b10 fffff803`25257529 nt!IopProcessWorkItem+0x81

11 ffff9f01`e6896b80 fffff803`2528e0c1 nt!ExpWorkerThread+0xe9

12 ffff9f01`e6896c10 fffff803`2537bcf6 nt!PspSystemThreadStartup+0x41

13 ffff9f01`e6896c60 00000000`00000000 nt!KxStartSystemThread+0x16

 

14: kd> .frame /r 0n0

00 ffff9f01`e68967d0 fffff80e`96906f52 vmswitch!VmsPtAdjustNumQueuePairsInUse+0x19 [d:\rs1\vm\dv\net\nvsp\driver\pt\vmspt.c @ 19024]

rax=ffff9f01e6896808 rbx=ffffbb0a58c59000 rcx=0000000000000000

rdx=00000000ffffffff rsi=ffffbb0a3f858000 rdi=0000000000000000

rip=fffff80e968e4651 rsp=ffff9f01e68967d0 rbp=ffff9f01e6896879

r8=ffffbb0a4eb607c0  r9=0000000000000000 r10=ffff9f01e7195000

r11=ffff9f01e68967e0 r12=ffffbb0a3f849040 r13=0000000000000001

r14=0000000000000000 r15=0000000000000001

iopl=0         nv up ei ng nz na pe nc

cs=0010  ss=0018  ds=002b  es=002b  fs=0053  gs=002b             efl=00010282

vmswitch!VmsPtAdjustNumQueuePairsInUse+0x19:

fffff80e`968e4651 488b7108        mov     rsi,qword ptr [rcx+8] ds:002b:00000000`00000008=????????????????

 Cuase is the member adapter Nic is pointin to Null value. This value checking was triggerd while doing some VMQ queue tunning.  Please share us what is the user behaivor when this server blue screen.

 

14: kd> dt _VMS_PROTOCOL_NIC_MEMBER

vmswitch!_VMS_PROTOCOL_NIC_MEMBER

   +0x000 ProtocolHandle   : Ptr64 Void

   +0x008 Nic              : Ptr64 _VMS_OBJ_NIC

 

MemberAdapter is NULL

 

14: kd> dt vmswitch!_VMS_PROTOCOL_EXT 0xffffbb0a3f849040 MemberAdapterWrapper

   +0xb50 MemberAdapterWrapper : (null)

 

14: kd> dt vmswitch!_VMS_OBJ_NIC 0xffffbb0a3f858000 NicExt.ExtContext

   +0xc38 NicExt            :

      +0x000 ExtContext        : 0xffffbb0a`3f849040 Void  --------------> ptNicExt

 

14: kd> dt _VMS_PROTOCOL_EXT 0xffffbb0a`3f849040 NicIndexes

vmswitch!_VMS_PROTOCOL_EXT

   +0x5c0 NicIndexes : [64] (null)

 

14: kd> dx -r1 (*((vmswitch!_VMS_OBJ_NIC_HEADER * (*)[64])0xffffbb0a3f849600))

(*((vmswitch!_VMS_OBJ_NIC_HEADER * (*)[64])0xffffbb0a3f849600))                 [Type: _VMS_OBJ_NIC_HEADER * [64]]

    [0]              : 0x0 [Type: _VMS_OBJ_NIC_HEADER *]

    [1]              : 0xffffbb0a3f85e970 [Type:  _VMS_OBJ_NIC_HEADER *] --------------------------> nicHeader

 

From current dump log, I didn’t find which  application or service free the member adapter pointer. We need some vmswitch ETL logs before dump generated. 


When the Virtual Machine Queue feature (Hyper-V) is deleted, the OS accesses the NIC public attribute resources. If MemberAdapter is an empty pointer, BSoD occurs.

To resolve this problem, the customer needs to deploy serial port redirection and kdump on the OS, which requires assistance from the OS vendor.


END