As with core dumps, there are a few ways to create crash dumps. When the system does this, it is of course disruptive. Usually a crash dump is created as part of the system’s panic routine. Normally, when the kernel encounters an error that could pose a threat to data, it has three objectives to pursue, in priority order:
Protect the data
Inform the user
Keep the system running
When the cause of the problem is a process or ...