On Mon, Apr 03, 2000 at 05:02:48PM +0300, Mitja Sarp wrote:
> I'd like to do some mail benchmarking, and netscape's mailstone seemed to be
> a good choice. Running the smtp-test gives the following result though:
>
> nurja[9374] t=954768787: mailclient (4.1 built Jun 21 1999 12:03:47)
> nurja[9374] t=954768787: procs=1 threads=1 seconds=10 ramptime=10...
> nurja[9375] t=0: Child starting
> nurja[9375] t=0: readResponse(6) errno=0: Error 0
> nurja[9375] t=0: Error reading [QUIT] response: errno=0: Error 0
> nurja[9375] t=0: ERROR process=0 thread=0 sendCommand(6, 'QUIT') errno=32:
>Broken pipe
> nurja[9375] t=0: Error sending [QUIT] command to server: errno=0: Error 0
> nurja[9375] t=0: readResponse(6) errno=0: Error 0
> .. etc
>
> When running from tcpserver, running from inetd gives a similar result, but without
> 'Broken pipe' at the end.
Hmm. This is the code fragment in qmail-smtpd:
void smtp_quit()
{
smtp_greet("221 "); out("\r\n"); flush(); _exit(0);
}
So the mailstone program should be seeing this response to the QUIT. Are you in a
position to tcpdump or recordio the traffic?
Regards.
