To ESMTP or not?

Hi,

I have a couple of external domains that our mail server can not connect with and the gwia log generates "420 TCP Read error"
Yet, these same external mail servers are able to successfully send emails to my users.
Besides this issue all other emails are going out OK.
When I perform a tracert I get a successful completion (it took just over a minute and my GWIA SMTP Timeout settings are at the default).

When I perform a telnet to these mail servers it connects and shows ESMTP in the response.
Can their server be setup NOT to accept SMTP, just ESMTP?

Are there any disadvantages to enabling ESMTP or should I be looking somewhere else?

As always, thanks for your responses!

Stan

Tags:

  • In article <Demaximis.6vt20n@no-mx.forums.microfocus.com>, Demaximis
    wrote:
    > I have a couple of external domains that our mail server can not connect
    > with and the gwia log generates "420 TCP Read error"
    > Yet, these same external mail servers are able to successfully send
    > emails to my users.
    > Besides this issue all other emails are going out OK.


    I would say there is some sort of networking problem between your GWIA and
    those external domains. I would check
    - your MTU settings as there may be a negotiation problem there
    - make sure there aren't errors showing on the server's NIC
    - check your router
  • Hi

    420 TCP Read error


    This is a communications issue. 99% of the time it's a misconfigured router/switch and it could be, as Andy pointed out, that MTU negotiation is the problem.

    The best method would be to determine which device on the way between you and the recipient is incorrectly configured, and fix that.

    Let us know how it goes.

    Cheers,
  • Hi Stan,

    In addition to what has been said above, you may find this TID of interest: https://www.novell.com/support/kb/doc.php?id=7007770

    It could be a firewall configuration at the destination domain site.

    Cheers,
  • Hi Stan,

    I've just had a chat to my "Team Cisco" and they recommend the following tests, assuming a Linux box!

    First, determine the lowest MTU size between your GWIA and the recipient's system: tracepath -n x.x.x.x (destination IP address).
    This will output the path and the last line will have something like this: Resume: pmtu 1500 - in this example it is telling me that the PMTU size is 1500.

    Next, do a ping with the Do Not Fragment flag set: ping -M do -s 1462 x.x.x.x (destination IP address)

    If the packet is too big you should get an ICMP response stating that the packet needs to be fragmented but the DF switch has been specified. If you don't get an ICMP response stating this then the recipient firewall is not configured to give the correct ICMP responses - that equals a problem.

    Try this and let us know the outcome please.

    Looking forward to hearing back from you :)

    Cheers,
  • Thanks for the suggestions.

    I have wireshark on my mail server and captured data while sending to one of the troublesome domains.
    Wireshark shows red lines on the communications between our mail server and one of these domains.
    Our mail server is on an up-to-date-patched SuSE 11 SP3 server running GW 2014 (14.0.1 -118418), which is a VM, on a Citrix XenServer.

    The bottom line is that we switched to cloud-based gwava and while incoming emails went through gwava, outgoing was not.
    These domains would not accept the switch between the ip addresses.
    We found this out after we called gwava support and had our outgoing email go through gwava.
    Test emails to persons in these domains were successfully transferred.

    Thanks so much for your helpful suggestions and this forum!

    Stan
  • Hi Stan,

    Glad that you managed to solve your issue and thanks very much for reporting back to us.

    Cheers,