what process killers does linux have?

oom is currently the only thing that kills automatically.

dmesg

and /var/log/messages should show oom kills.

If the process can handle that signal, it could log at least the kill.

Normally memory hogs get killed. Perhaps more swap space can help you, if the memory is only getting allocated but is not really needed.

Else: Get more RAM.


Coming to mind:

  • getting a SIGSEGV or SIGBUS for invalid memory access
  • SIGILL for invalid instructions
  • SIGFPE when doing an invalid arithmetic operation like division by 0
  • SIGXCPU or SIGXFSZ if the process exceeds some limits
  • SIGHUP for disconnected terminals and some other situations
  • SIGPIPE when trying to write to a closed pipe or socket