On 8/9/06, David Boyes <[EMAIL PROTECTED]> wrote:
Unless something has changed in 5.2, the TCPIP stack applications have a way to do clean shutdowns, but the stack itself does not. This has been a lingering remnant of the old IP stack code since the beginning – it's probably not critical, but it is annoying that the only way to shut down the stack is CP FORCE or sending it an external interrupt, essentially crashing it.
#cp ext does not crash the stack, but is picked up as a signal to stop the stack. My understanding is that from that point the stack does not accept new connections but continues to run the open ones and give them time to close. I suppose this might make sense to complete a current ftp transfer or web transaction. So it might make sense to have TCPIP understand a signal shutdown like that. If the sessions don't close all before the signal expired, CP will take the virtual machine down. Would work. Rob
