[Mailman-Users] messages stuck in queue after approval

[Richard Barrett]

On 13 Aug 2004, at 13:38, Juan Nin wrote:

> Brad Knowles wrote:
>
>>>  Aug 12 18:28:46 2004 (11220) All recipients refused: Server not 
>>> connected
>>>  Aug 12 18:28:46 2004 (11220) smtp for 338 recips, completed in 
>>> 65.904 seconds
>>     This implies that your MTA is not running, or not accepting 
>> connections.
>>     Check your MTA to make sure it's running and accepting 
>> connections.  Also check your MTA logs to see if they have anything 
>> to say about the incoming connection attempts from Mailman.
>
> Yes, the logs seem to be telling that, but actually the MTA is running 
> just fine, it's the same MTA through which I'm sending these mails..  
> :)
>

Assuming you are using Mailman's SMTPDirect module it is possible that 
your MTA will refuse a connection from Mailman at some point in time or 
even drop a connection but at another time accept and process a 
connection from Mailman OK. Depending on how your MTA is configured and 
how heavily loaded your server is the MTA may refuse a particular 
connection attempt if, for instance, it has reached its concurrent 
connection limit or the load factor on the machine exceeds specified 
limits.

Looking back to your original post which had the following error log 
entry:

>
> and in /var/logs/mailman/error
>  get:
> Aug 12 18:29:33 2004 (11220) Delivery exception:
> Aug 12 18:29:33 2004 (11220) Traceback (innermost last):
>   File "/var/mailman/Mailman/Handlers/HandlerAPI.py", line 82, in 
> do_pipeline
>     func(mlist, msg, msgdata)
>   File "/var/mailman/Mailman/Handlers/SMTPDirect.py", line 74, in 
> process
>     deliver(admin, msgtext, chunk, refused)
>   File "/var/mailman/Mailman/Handlers/SMTPDirect.py", line 215, in 
> deliver
>     conn.quit()
>   File "/var/mailman/Mailman/pythonlib/smtplib.py", line 523, in quit
>     self.docmd("quit")
>   File "/var/mailman/Mailman/pythonlib/smtplib.py", line 307, in docmd
>     return self.getreply()
>   File "/var/mailman/Mailman/pythonlib/smtplib.py", line 281, in 
> getreply
>     line = self.file.readline()

suggests that the smtplib (Python SMTP support library) raised the 
Delivery exception because it found hte TCP/IP connection to the MTA 
had  been closed abruptly when it was trying to complete the protocol 
dialogue with the MTA.

It is not obvious that this is a Mailman problem per se.

> Moreover, several other lists on this server are running just fine, 
> with no problem...
>
> before sending my prevoius mail to the list, I run a "tail -f 
> /var/log/maillog" to see if there was any mailman related error, and 
> no logs appear from the lists which are not running...
>
> # ls /var/mailman/locks/
> qrunner.lock  qrunner.lock.mydomain.tld.24547
>
> # cat /var/log/mailman/qrunner
> Aug 13 09:30:03 2004 (24486) Could not acquire qrunner lock
> Aug 13 09:30:03 2004 (24488) Could not acquire qrunner lock
> Aug 13 09:31:02 2004 (24500) Could not acquire qrunner lock
> Aug 13 09:31:02 2004 (24499) Could not acquire qrunner lock
> Aug 13 09:32:02 2004 (24519) Could not acquire qrunner lock
>

With MM 2.0.13 the qrunner is a cron job. If my recollect is correct, 
when a previously launched instance of the qrunner is running and cron 
launches a new instance the new instance logs the "could not acquire" 
entry in the qrunner log and leaves the older instance to continue 
processing the queues. The only thing these log entries usually tell 
you are that on occasions it is taking longer than the qrunner launch 
interval to process all the queues. But the above entries from the 
qrunner log show that the "could not acquire" entry is appearing twice 
at the same time associated with different process ids, which makes me 
suspect that there may be two crontabs extant on your system which are 
both trying to start the qrunner every minute. Not a good thing.

> # cat /var/log/mailman/smtp
> Aug 13 09:34:41 2004 (24547) All recipients refused: Server not 
> connected
> Aug 13 09:35:17 2004 (24547) All recipients refused: Server not 
> connected
> Aug 13 09:35:17 2004 (24547) smtp for 338 recips, completed in 75.708 
> seconds
>
> # cat /var/log/mailman/smtp-failure | head -3
> Aug 13 09:30:13 2004 (24468) -1 user at domain.com (ignore)
> Aug 13 09:30:13 2004 (24468) -1 user at other.net (ignore)
> Aug 13 09:30:13 2004 (24468) -1 user at another.org (ignore)
>
> any ideas??
>
> should i delete this lists and try again??  :(