rump.halt —
halt a rump kernel
The rump.halt utility exits a rump kernel. The file
  system cache, if present, is flushed. Since a rump kernel does not control its
  clients, they are not directly affected by rump.halt.
  However, they will be unable to request further services from the halted rump
  kernel.
The options are as follows:
  - -d
- Create a core dump. The core file is saved according to standard userland
      program coredump rules, and can be later examined with a debugger.
- -h
- By default the process hosting the rump kernel exits. Using this option
      shuts down rump kernel activity, but does not cause the hosting process to
      exit.
- -n
- Do not flush the file system cache. This option should be used with
      extreme caution. It can be used if a virtual disk or a virtual processor
      is virtually on fire.
Therump.halt command appeared in
  NetBSD 6.0.
While using -h makes it impossible to issue further
  system calls, it does not necessarily stop all activity in a rump kernel. It
  is recommended this option is used only for debugging purposes.