Our physical Linux servers have an ILO board that may be used to create an non-maskable interrupt. This easily triggers the creation of a kernel dump in case the server is unresponsive. How can this be done in a virtual machine?

In order to debug unresponsible Linux servers, we currently trigger the kernel dump of virtual machines via key combination ALT+SysRq+c. However it does not work via terminal server connection - especially with different keyboard mappings and various client configurations.

To straighten the process, we would like to use the API or to use a native Vmware method that triggers a NMI on hung machines.

Is this possible?

  • I hope you don't have to use this often! – ewwhite Oct 17 '13 at 11:42
  • Not very often but its absolutely needed in order to receive vendor support. The alternative would be to just reboot the VM but that would not solve the root cause. – Reiner Rottmann Oct 18 '13 at 13:48

On the actual ESXi host, logged in as root run this;

/sbin/vmdumper <world_id> nmi

where is that of the VM you want to NMI.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.