On 05.08.2014 20:08, Noel Jones wrote: > On 8/5/2014 12:53 PM, M. Rodrigo Monteiro wrote: >> 2014-08-05 14:47 GMT-03:00 Noel Jones <njo...@megan.vbhcs.org>: >>> On 8/5/2014 12:23 PM, M. Rodrigo Monteiro wrote: >>>> Hi! >>>> >>>> This postfix act as a Relay. >>>> From all servers, but one, the message size is 20MB. For these one it's >>>> 10MB >>>> >>>> # postconf message_size_limit >>>> message_size_limit = 20971520 >>>> >>>> >>>> Here is the telnet response from all servers >>>> >>>> # telnet relay 25 >>>> Trying XXX... >>>> Connected to XXX. >>>> Escape character is '^]'. >>>> 220 relay ESMTP Postfix >>>> ehlo teste >>>> 250-relay >>>> 250-PIPELINING >>>> 250-SIZE 20971520 >>>> 250-ETRN >>>> 250-STARTTLS >>>> 250-ENHANCEDSTATUSCODES >>>> 250 8BITMIME >>>> quit >>>> >>> >>> OK, looks like normal postfix. >> >> Ok >> >>> >>> >>>> >>>> Here is from just one spefic server >>>> >>>> # telnet relay 25 >>>> Trying XXX... >>>> Connected to relay >>>> Escape character is '^]'. >>>> 220 relay ESMTP Service ready >>>> ehlo server >>>> 250-Requested mail action okay, completed >>>> 250-SIZE 10000000 >>>> 250-ETRN >>>> 250-8BITMIME >>>> 250-STARTTLS >>>> 250 OK >>>> ^] >>>> telnet> quit >>>> Connection closed. >>>> >>> >>> And this doesn't look like postfix. Either there's some sort of >>> proxy interfering with SMTP, or this isn't connecting to the same >>> server. Maybe some security software or firewall on the oddball server? >>> >> >> Yes, I thought that too. But the first 220 replay with my Relay name >> (smtp2.tld.com) indicates otherwise, not? >> But, I'll investigate on that. > > In the samples you provided, the 220 server responses are clearly > different. The first one looks like postfix, the second one with > the different message size does not look like postfix. > > The evidence you provided leads me to believe the second server is > not connecting to postfix. Usually this means there is some sort of > proxy altering the SMTP commands, but there could be other problems, > such as a forgotten /etc/hosts entry pointing to the wrong IP.
I was once bitten by a sendmail instance which was still installed on the host, and which took over the port 25 at boot time, making postfix unable to bind and start. regards, jwi > > At any rate, this isn't a postfix problem. > > > -- Noel Jones >