On 2007-09-05 16:28:50 +0200, Johan Almqvist wrote: > I just had a look at the stats graph I created over a year ago. > http://www.almqvist.net/hamspam-week.png > > The most noteworthy change is that the number of earlytalkers has > decreased by an order of magnitude. The time for this coincides > with my migration from tcpserver to forkserver. Could there be a relation?
Do you accept more connections in parallel than before the migration?
The earlytalker plugin also reports clients which closed the socket
before the timeout. If there is a significant time between the
completion of the tcp handshake (which is handled entirely by the
kernel) and the time qpsmtpd (or tcpserver) accepts the connection, it
is possible that the client already gave up. In this case an earlytalker
would be reported even though the client is an early non-talker :-).
hp
--
_ | Peter J. Holzer | I know I'd be respectful of a pirate
|_|_) | Sysadmin WSR | with an emu on his shoulder.
| | | [EMAIL PROTECTED] |
__/ | http://www.hjp.at/ | -- Sam in "Freefall"
signature.asc
Description: Digital signature
