gmail.com problems and this list

David Ford david at blue-labs.org
Sun Jul 8 18:42:44 CEST 2007


Wolfgang S. Rupprecht wrote:
> Harald Welte <laforge at openmoko.org> writes:
>   
>> yes, I can confirm this.  This is from our exim4 installation on the
>> list server:
>> 2007-07-07 09:17:19 1I74UH-0005cE-SD <= robin.paulson at gmail.com H=py-out-1112.google.com [64.233.166.176] P=esmtp S=3940 id=2f3aa2770707070010k7407439axfaa269de33395ea2-JsoAwUIsXosN+BqQ9rBEUg at public.gmane.org
>> 2007-07-07 09:27:18 1I74dw-000682-6l <= robin.paulson at gmail.com H=py-out-1112.google.com [64.233.166.181] P=esmtp S=3940 id=2f3aa2770707070010k7407439axfaa269de33395ea2-JsoAwUIsXosN+BqQ9rBEUg at public.gmane.org
>> ...
>>     
>
> This isn't enough to tell which side is causing the misbehavior.  I
> find that slapping a tcpdump(1) on the interface and logging to a file
> is the simplest way to see what is really happening (eg. tcpdump -w
> /xxx/logfile -s 1500 -i eth0).  Running wireshark() (formerly
> ethereal()) and using the "follow tcp stream" shows the back and forth
> conversation.  In this case I bet you'll see that gmail is either
> timing out due to slow replies or packets are getting dropped and
> gmail is asking for lots of retransmissions and eventually throws in
> the towel.
>
> My gut feel is that the receiver is most likely the villain with
> either and overloaded CPU and/or lots of slow antispam filters that
> need to be run.
>
> (If you want help setting up the tcpdump or looking at the logs, feel
> free to send mail to my gmail address.  I've tracked down enough mail
> problems like this that this is very old hat.)
>
> -wolfgang
>   


The log that I posted showed the initial retransmit happening within
google's gmail servers.

-david




More information about the community mailing list