On Sat, Apr 24, 2010 at 06:36:48PM -0400, Wietse Venema wrote:
> Here is my reconstruction from this inadequate logging sample.
> 
> > Apr 24 13:20:10 doctor doctor[31]: postfix/smtpd[17053]: 0762D12CFB0D: 
> > client=defout.telus.net[204.209.205.55]
> > Apr 24 13:20:41 doctor doctor[31]: postfix/cleanup[17311]: 0762D12CFB0D: 
> > message-id=<6ef179d262924e5a8c03336971544...@taskercomp>
> 
> At 13:22:00, a different Postfix SMTPD process reports a queue file
> write error. This Postfix SMTPD process is missing from the logfile.
> Also missing is the logging for the postmaster notice delivery.
> 
> > Apr 24 13:22:04 doctor doctor[31]: postfix/qmgr[4921]: 0762D12CFB0D: 
> > from=<sender>, size=5104017, nrcpt=1 (queue active)
> > Apr 24 13:22:04 doctor clamsmtpd: 100B1C: from=sender, to=recipient, 
> > status=CLEAN
> 
> The logging shows a clamsmtpd filter, but the logging is missing
> for the process that sends the message into the clamsmtpd filter.
> 
> Let's assume that this missing process was the same missing Postfix
> SMTPD process that reported a queue file write error. That is
> admittedly optimistic, considering that you left out most of the
> relevant information.
> 
> Then, we have the following timeline:
> 
> 13:20:10 The missing Postfix SMTPD process receives a connection.
> 
> 13:20:10 The missing Postfix SMTPD process connects to a before-queue
>       filter (clamsmtpd).
> 
> 13:20:10 The before-queue filter (clamsmtpd) connects to Postfix
>       SMTPD process 17053.
> 
> 13:20:41 The before-queue filter (clamsmtpd) sends the message
>       header and perhaps the first part of the message body.
> 
> 13:22:00 The missing Postfix SMTPD process times out (smtpd_proxy_timeout)
>       and sends a postmaster notification. I already mentioned
>       that the postmaster notification is missing in the logging.
> 
> 13:22:04 The before-queue filter (clamsmtpd) finishes after two
>       and a half minute. This is long after the missing Postfix
>       SMTPD process has given up.
> 
> If this timeline is correct then you have too many body_checks
> rules, or you have a clamsmtpd setup that is taking too much time.
> 
> But it is more likely that you left out a whole lot of other
> relevant information with your selective logfile reporting.
> 
>       Wietse

Let me see if I can find the thorough transaction.

I have amavisd-ng on 10024 and clamsmtpd on 10125.

Problem is that this is not the only piece of e-mail being inspected
at that time frame.


-- 
Member - Liberal International  This is doc...@nl2k.ab.ca Ici doc...@nl2k.ab.ca
God, Queen and country! Never Satan President Republic! Beware AntiChrist 
rising! 
http://twitter.com/rootnl2k http://www.facebook.com/dyadallee
UK Time for a Common Sense change vote Liberal Democrat / Alliance 

Reply via email to