Problems with shutting down server on Red Hat reboot
- From: "John Smith" <SMasiutin@xxxxxxxxx>
- Date: 26 Oct 2006 07:56:15 -0700
I have TCP server that accepts client and then receives data from them.
Server has SIGTERM handler, on which all sockets are closed and server
exits.
If server has several clients connected and I manually send it kill
-TERM <server_pid> everything goes correct: client application's
sockets are disconnected and server shuts down the right way. But if I
reboot machine some strange behavior occur: client application's
sockets still think they are connected, and after server receives
SIGTERM all semaphores that were in blocked state become unblocked.
Can someone explain me what's going on?
.
- Prev by Date: Re: select() indicates writability before a non-blocking TCP socket has actually connected
- Next by Date: close /proc/self/cmdline
- Previous by thread: select() indicates writability before a non-blocking TCP socket has actually connected
- Next by thread: close /proc/self/cmdline
- Index(es):
Relevant Pages
|
|
