Re: Python Error with Postfix queue filling up

Mat Cantin schreef op 08-03-2013 2:19:
> On 2013-03-06 0:38, Paul Klos wrote:
>> 
>> The message contains a 0xbb character that wallace is choking on.
>> 
>> I've had some similar errors, and just for laughs changed these
>> characters into something else using hexedit, after which the mails
>> went
>> through without problems. For now, I have disabled wallace on my 
>> Kolab
>> instance.
>> 
>> This is bug https://issues.kolab.org/show_bug.cgi?id=1627.
>> 
> 
> This is exactly my problem. Interestingly, as described in the 
> comments
> of this bug report, I tried to disable wallace in the master.cf file 
> and
> the flush the postfix queue, but the wallace python error still showed
> up and the error in the log still showed the email being delivered to
> port 10026. Are these emails destined to go to wallace now that they 
> are
> in the queue?
> 
> I guess my lack of postfix knowledge is showing. :-P
> 
> --
> Mat
> 
Disabling Wallace will indeed only help for new messages. If you look 
at the messages in the deferred queue (/var/spool/postfix/deferrred), 
you'll see that it contains a complete destination address, including 
port 10026 (somewhere at the top of the file). I don't know if there's a 
'proper' way to inspect these messages, but a simple less will show you 
this, despite a warning about a possible binary file.

So, I was expecting these messages to remain in the queue, and I was 
planning to leave them there and use them as test cases when I enable 
wallace again after this bug has been fixed. The mysterious thing is 
that my queue is now empty. I checked after this problem was posted to 
the list.

I did at some point restart postfix, maybe that has something to do 
with it. Or maybe postfix just gives up on these messages after a while.

Cheers,

Paul