> > root@netsphere:~# sockstat |grep exim
> > exim exim-4.1 15156 0 tcp4 192.168.1.1:3851
66.35.250.206:113
> > exim exim-4.1 15156 1 tcp4 192.168.1.1:25
> > 66.35.250.206:53896
> > exim exim-4.1 15156 2 tcp4 192.168.1.1:25
> > 66.35.250.206:53896
> > exim exim-4.1 99992 0 tcp4 192.168.1.1:25 *:*
> > root@netsphere:~# exiwhat
> > 15156 handling incoming connection from [66.35.250.206]
> > 99992 daemon: -q15m, listening for SMTP
> >
> > Nopes :) Once again, three connections, exiwhat only states one
incoming
> > connection...
>
> There is only one exim receiving process, 15156, so there can only be
> one incoming connection. In fact that is what we see in
>
> > exim exim-4.1 15156 1 tcp4 192.168.1.1:25 66.35.250.206:53896
>
> I don't understand the second, almost identical line:
>
> > exim exim-4.1 15156 2 tcp4 192.168.1.1:25 66.35.250.206:53896
>
> ... but note that it has the same IP addresses and ports. Therefore, it
> must represent the same connection as the previous line. What is the
> meaning of the "1" and "2" column that is different between the two
> lines?
USER COMMAND PID FD PROTO LOCAL ADDRESS FOREIGN ADDRESS
exim exim-4.1 22555 1 tcp4 192.168.1.1:25
195.92.249.251:57206
exim exim-4.1 22555 2 tcp4 192.168.1.1:25
195.92.249.251:57206
The 1 and 2 would have been the file descriptors... And funny enough, once
again the same scenario - duplicate connection...
> The first output line,
>
> > exim exim-4.1 15156 0 tcp4 192.168.1.1:3851
66.35.250.206:113
>
> is an *outgoing* connection to port 113 on the remote host. Port 113 is
> the ident port. You caught this process while it is doing the ident
> thing.
Yup, my bad - I just grabed the output from the grep and pasted...
> So I cannot see any inconsistency in that output in terms of
> connections, though the repeated line is odd (but I'm not familiar with
> sockstat).
Erm, it still doesn't count the current connections right in the log files?
John Jetmore posted a reply earlier with basically the same scenario - he's
connection count is just alot more than mine...
--
me
