More observations on "Mailscanner stopped delivering to outgoing queue"

[Mordur Ingolfsson]

Glenn Steen wrote:
> On 01/07/06, Mordur Ingolfsson <ms at 1984.is> wrote:
>> I apologize for not having been thorough enough in the first place, but
>> I have been looking better into the problem described in my previous
>> mail to this list under the subject ."Mailscanner stopped delivering to
>> outgoing queue"  a few hours ago.
>>
>> Below is an excerpt from syslog. Everything is fine and messages get
>> nicely processed and delivered, as the first five lines indicate.  Then,
>> at 12:53:13, something happens and Debian-exim (the exim user and the
>> username under which mailscanner operates) starts to complain about
>> something (lines in syslog excerpt below marked with"--------->" ) . And
>> the message gets scanned and rescanned.
>>
>> After this, nothing gets delivered. If you read further into the syslog
>> excerpt you will see that this messages, and messages delivered to the
>> host subsequently, do not get delivered. Messages are piling up on the
>> incoming queue and MailScanner refuses to deliver. The child processes
>> are zombies <defunct> and I simply cannot find a way arount this.
>>
>>
>> Thank you,
>>
>> Mordur
>>
> (snip)
>> Jun 30 12:53:21 mx0 dccproc[25688]: socket(UDP): Address family not
>> supported by protocol
> (snip)
>> --------->Jun 30 12:53:25 mx0 Debian-exim: Process did not exit cleanly,
>> returned 28 with signal 0
> (snip)
>
> I'm not sure the dcc error is the culprit, but... well, something
> exits with an error, probably "agitating" exim... you should look into
> that:)
>
The dcc error is not the culprit. I have solved that by setting ipv6 to
off in dcc config and it changed nothing.