httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From (Scott Michel)
Subject Re: restart disaster with 1.2-dev
Date Mon, 19 Aug 1996 04:31:39 GMT
Randy Terbush says:
> Actually, I'm finding that -HUP will fail with "address already in use"
> on FreeBSD 2.1.5 on the *first* -HUP attempt. I have stopped my nightly
> HUP until I get a chance to look at this more.

Sounds like you've been caught by the 2 RTT CLOSE_WAIT. You've killed
the daemon, but TCP is keeping the socket open for 2 RTT to check for
stray packets received after close.

Basically, if memory serves me, 2 RTT = 4 minutes. So you can either
sleep or turn this feature off. Which socket option this is/was is not
on the top of my cranium at the moment.


View raw message