[Mailman-Users] delayed distribution of messages

Lluis Montoliu montoliu at cnb.csic.es
Sat Sep 17 19:17:56 CEST 2011


Thanks Mark for a detailed and useful answer.

Here are the headings of one of these delayed messages, sent yesterday  
(at 18:47), that was eventually distributed 45 minutes later (at 19:32):

Message was sent from my address (my computer at work) to the list  
called "istt_list at transtechsociety.org", hosted at d344.dinaserver.com

_________________________________________

Received:

from [161.111.10.139] (helo=ironport1.csic.es) by buzon1.csic.es with  
esmtp (Exim 4.72) (envelope-from  
<istt_list-bounces at transtechsociety.org>) id 1R4cHE-0002Hp-GH for  
montoliu at cnb.csic.es; Fri, 16 Sep 2011 19:32:36 +0200

from d344.dinaserver.com ([82.98.151.168]) by ironport1.csic.es with  
ESMTP; 16 Sep 2011 19:32:36 +0200

from d344.dinaserver.com (localhost [127.0.0.1]) by  
d344.dinaserver.com (Postfix) with ESMTP id 5D6E724B62; Fri, 16 Sep  
2011 19:32:29 +0200 (CEST)

from ironport.csic.es (ironport.csic.es [161.111.10.143]) by  
d344.dinaserver.com (Postfix) with ESMTP id 4329A23BBE for  
<istt_list at transtechsociety.org>; Fri, 16 Sep 2011 18:47:42 +0200 (CEST)

from estafeta.csic.es ([161.111.10.248]) by ironport4.csic.es with  
ESMTP/TLS/ADH-AES256-SHA; 16 Sep 2011 18:47:41 +0200

from [150.244.81.39] (helsinki.cnb.csic.es [150.244.81.39]) by  
estafeta.csic.es (Postfix) with ESMTP id BFE741A0C29 for  
<istt_list at transtechsociety.org>; Fri, 16 Sep 2011 18:47:46 +0200 (CEST)

_____________________________________________

At the CSIC (the network from where the message was sent) we are now  
using SPF, as it is indicated also in the header:

Received-SPF:  Pass identity=mailfrom; client-ip=82.98.151.168;  
receiver=ironport1.csic.es;  
envelope-from="istt_list-bounces at transtechsociety.org";  
x-sender="istt_list-bounces at transtechsociety.org";  
x-conformance=spf_only; x-record-type="v=spf1"
None identity=helo; client-ip=82.98.151.168;  
receiver=ironport1.csic.es;  
envelope-from="istt_list-bounces at transtechsociety.org";  
x-sender="postmaster at d344.dinaserver.com"; x-conformance=spf_only

Thus, according to the timings above, it looks as the delay indeed  
occurs at our d344.dinaserver.com server, not sure whether we can say  
this is related to mailman or to server performance with these data,  
could we?

Thanks for comments,

regards

Lluis




Quoting Mark Sapiro <mark at msapiro.net>:

> Lluis Montoliu wrote:
>>
>> I am responsible for a list of ~500 members and other much smaller
>> (<25 members). All these lists, specially the largest, are
>> experiencing unexpected delays in distributing messages of about 45-60
>> min from sending a message to the list and the message being
>> distributed. This is very problematic and makes the use of these lists
>> a pain, specially for maintain a dynamic discussion on any subject.
>>
>> The mail system itself seems to be OK. If I do send a message to the
>> same 500 members from a PHP script the message gets distributed
>> immediately.
>>
>> I don't have root privileges although we are on our own (administered)
>> server. All I can do is complain before my hosting services and they
>> say they can't do much and refer to internal mailman problems to
>> explain this poor performance (that I haven't seen before in other
>> servers).
>
>
> The first thing you need to do is examine the Received: headers in one
> of these delayed messages to determine where the delay occurs. It may
> or may not even involve Mailman. Note that Received: headers are added
> at the top of the headers by each MTA as it receives the message,
> thus, the chain reads chronologically from bottom to top.
>
> Yo need two find the two headers which indicate delivery to Mailman and
> delivery from Mailman. For example, in the message to which I'm
> replying as received by me from the list, the two relevant headers are:
>
>
> Received: from albatross.python.org (localhost [127.0.0.1])
> 	by mail.python.org (Postfix) with ESMTP id 3S1Gcn4GdGzP1M
> 	for <mark at msapiro.net>; Sat, 17 Sep 2011 16:50:49 +0200 (CEST)
>
> and
>
> Received: from albatross.python.org (localhost [127.0.0.1])
> 	by mail.python.org (Postfix) with ESMTP id 3S15yp2hlbzNvQ
> 	for <mailman-users at python.org>; Sat, 17 Sep 2011 10:20:50 +0200 (CEST)
>
> The bottom header indicates the post was delivered to mail.python.org
> atSat, 17 Sep 2011 10:20:50 +0200, and the top header indicates the
> message to me was received from Mailman at Sat, 17 Sep 2011 16:50:49
> +0200. Thus, there was a delay of about 6 hours and 30 minutes between
> delivery to Mailman and delivery from Mailman. In this case, this
> delay is explained by the fact the you are a new list member and your
> post was moderated as noted by the header
>
> X-Mailman-Approved-At: Sat, 17 Sep 2011 16:50:04 +0200
>
> Thus, there was actually only a 45 second delay between moderator
> approval and delivery from Mailman.
>
> Anyway, if the above indicates the delay is between delivery to Mailman
> and delivery from Mailman, you need to examine Mailman's 'smtp' and
> 'smtp-failure' logs for 'long' processing times and or temporary
> failures to get additional clues.
>
> [...]
>> However, I wanted to ask this community whether there are any other
>> known bugs or issues to take into account that might impair list
>> performance and delay the distribution of messages,
>
>
> There are no known bugs but there are several potential issues that
> could affect this, but we need more information to say more.
>
> --
> Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
> San Francisco Bay Area, California    better use your sense - B. Dylan
>
>



-- 
Dr. Lluis Montoliu
Investigador Cientifico - Research Scientist CSIC
Centro Nacional de Biotecnologia (CNB-CSIC)
Campus de Cantoblanco
C/ Darwin, 3
28049 Madrid (Spain)
Tel. +34-91-5854844  / Fax +34-91-5854506
e-mail: montoliu at cnb.csic.es
WEB: http://www.cnb.csic.es/~montoliu/



More information about the Mailman-Users mailing list