Jake Vickers wrote:
> Eric "Shubes" wrote:
>> This problem has me stumped at the moment.
>>
>> A (relatively) high volume toaster (10.0.1.69) on CentOS4. It's been
>> running
>> fine for the most part. However, when processing messages from a local
>> web
>> server (10.0.1.66) relayed from postfix, the message doesn't go through.
>> FWIW (doesn't make sense to me) the sending postfix queue shows that the
>> connection timed out or no route to host.
>>
>> I turned on recordio to see what's happening, and here the smtp log for a
>> session:
>>
>> 2007-01-04 18:23:53.798287500 tcpserver: status: 0/100
>> 2007-01-04 18:23:54.875954500 tcpserver: status: 1/100
>> 2007-01-04 18:23:54.875962500 tcpserver: pid 5787 from 10.0.1.66
>> 2007-01-04 18:23:54.875964500 tcpserver: ok 5787 spin4:10.0.1.69:25
>> :10.0.1.66::43332
>> 2007-01-04 18:23:54.875967500 5787 < [EOF]
>> 2007-01-04 18:23:54.970660500 5787 > 220 mydomain.com - Welcome to Qmail
>> Toaster Ver. 1.3 SMTP Server ESMTP^M
>> 2007-01-04 18:23:54.970671500 5787 > [EOF]
>> 2007-01-04 18:23:54.970673500 tcpserver: end 5787 status 256
>> 2007-01-04 18:23:54.970675500 tcpserver: status: 0/100
>>
>> The smtp session rapidly terminates, out of the blue.
>>
>> I can telnet from the web server to the toaster just fine.
>>
>> tcp.smtp:
>> 127.:allow,RELAYCLIENT=""
>> 10.0.1.:allow,RELAYCLIENT=""
>> :allow,BADMIMETYPE="",BADLOADERTYPE="M",CHKUSER_RCPTLIMIT="50",CHKUSER_WRONGRCPTLIMIT="10",QMAILQUEUE="/var/qmail/bin/simscan",DKSIGN="/var/qmail/control/domainkeys/%/private"
>>
>>
>> qmail-dk has been taken out (qmail-queue -> qmail-queue.orig)
>>
> Is the Postfix machine mad because of the server's HELO?
I don't think so. In another case in the log, the session terminates before
the helo is even sent:
2007-01-04 18:23:03.310745500 tcpserver: status: 0/100
2007-01-04 18:23:03.562492500 tcpserver: status: 1/100
2007-01-04 18:23:03.562500500 tcpserver: pid 5542 from 10.0.1.66
2007-01-04 18:23:03.562503500 tcpserver: ok 5542 spin4:10.0.1.69:25
:10.0.1.66::42927
2007-01-04 18:23:03.563811500 5542 < [EOF]
2007-01-04 18:23:03.626124500 tcpserver: end 5542 status 256
2007-01-04 18:23:03.626134500 tcpserver: status: 0/100
2007-01-04 18:23:03.626302500 5542 > 220 SPiNInternetHosting.com - Welcome
to Qmail Toaster Ver. 1.3 SMTP Server ESMTP^M
2007-01-04 18:23:03.626442500 5542 > [EOF]
2007-01-04 18:23:04.284530500 tcpserver: status: 1/100
Do you suppose that the initial [EOF] coming from postfix is causing
tcpserver to end? (It doesn't appear to in the first example, but the delay
in logging might make things look a little out of whack, as in this example)
> Are the Postfix logs showing anything?
Yes:
Jan 4 20:24:31 spin1 postfix/smtp[27795]: connect to
mail.mydomain.com[10.0.1.70]: No route to host (port 25)
Now this part I don't quite understand. 10.0.1.69 is the toaster's eth0,
while 10.0.1.70 is the toaster's eth0:5
Could that be hosing things up? (I'm not familiar with eth0:5 notation)
--
-Eric 'shubes'
---------------------------------------------------------------------
QmailToaster hosted by: VR Hosted <http://www.vr.org>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
