你好,
您这个是一个Hyper V cluster 的环境吗?然后在群集中频繁出现1146的报错,并且Hyper V manager无法启动吗?
综合这两个现象来看,我们一般需要收集kernel dump来看。
On problematic nodes
Confirm the cluster node had configured Kernel memory dump and paging file\dump location
disk has enough space
----------------------------
To set up a kernel memory dump, steps can be referred to:
Enable the system to generate a kernel memory dump by changing the following
registry key:
HKEY_LOCAL_MACHINE \System\CurrentControlSet\Control\CrashControl
Value Name: CrashDumpEnabled
Data Type: REG_DWORD
Value: 2
Please explicitly specify the paging files on the system drive.
Locate to the registry HKEY_LOCAL_MACHINE\System\CurrentControlSet\Control\Session Manager\Memory Management
Double click 'PagingFiles' and change the paging file on C drive as "C:\pagefile.sys
8300 8300" which changes the initial size and maximum size of paging file.
Please specify the following registry key to change the dump file location:
Locate to the registry HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\CrashControl
Double click 'DumpFile' to change the location of the dump file. Please choose another disk drive with more free space.
Reboot
the server for settings take effect.
Add a DWORD registry value named "DebugBreakOnDeadlock"
value 3 atHKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\ClusSvc\Parameters\DebugBreakOnDeadlock
Restart the cluster service.
DownloadRHSMon.zip(D:\Alogs\收藏\RHS)
from attachment, runrun.cmd.
Upload system event log and cluster log from both nodes.
When the RHS deadlock issue occurs, rhsmon.exe will detect it. It will launch dumpncrash.cmd to create user dump for rhs.exe/clussvc.exe and then crash the box to get the memory
dump. Please remove the DebugBreakOnDeadlock key from this node once the issue been reproduced and dump generated.
由于论坛不能收集dump分析,如果您想进一步排除,建议您向微软开case做深入分析:
Best Regards,
Anne
Please remember to mark the replies as answers if they help.
If you have feedback for TechNet Subscriber Support, contact tnmff@.