I'm experiencing a triple-fault while booting an OS of my own creation. I have "Gather debugging information" set to "full". VMware is failing to generate the suspend and memory files needed to generate a debug core dump file via vmss2core.
According to dariusd's comment in VMWare workstation monitor.suspend_on_triplefau... |VMware Communities , two files of the form debug.guest should have been created.
I've attached my vmware.log file.
Hi
I am no expert on this type of problems but I noticed that you use
monitor_control.log_vmsample = "TRUE"
while you do not use
monitor.suspend_on_triplefault="true"
as suggested by DariusD in VMWare workstation monitor.suspend_on_triplefau... |VMware Communities
Hi
I am no expert on this type of problems but I noticed that you use
monitor_control.log_vmsample = "TRUE"
while you do not use
monitor.suspend_on_triplefault="true"
as suggested by DariusD in VMWare workstation monitor.suspend_on_triplefau... |VMware Communities
Ah, I had assumed "full debug" also enabled the suspend_on_triplefault feature. Please forgive my naïveté. Thus far I've only used VMware with "supported" OSes. So, I was an "uninformed user". As a result I haven't needed to tinker with my .vmx files directly.
I am able to get a core file now.
Thanks for the suggestion.
Anyone know why I'm getting access denied while trying to access the referenced article (VMWare workstation monitor.suspend_on_triplefau... |VMware Communities)?
Try VMWare workstation monitor.suspend_on_triplefault fails
Thank you -- that link worked. But I'm using Fusion, so how do I set "Gather Debugging Information" to "Full" ? I don't see any UI to do that, so I assume it's a .vmx setting?
Add
monitor_control.log_vmsample = "TRUE"
monitor.suspend_on_triplefault="true"to your vmx-file while the VM is powered off.
Perfect, thank you! Any chance I can get a register dump at the point of the triple fault via another vmx setting? I noticed that the VMware.log monitor ring buffer entries that indicated the chain of faults didn't change the RIP or the RSP for each fault (in my case, they were definitely different -- not sure why it didn't change).