[Mimedefang] Seeing weird SMFI errors

[Philip Prindeville]

On Jul 26, 2015, at 11:26 PM, Philip Prindeville <philipp_subx at redfish-solutions.com> wrote:

> 
> On Jul 26, 2015, at 7:42 PM, Dianne Skoll <dfs at roaringpenguin.com> wrote:
> 
>> On Sun, 26 Jul 2015 18:03:38 -0600
>> Philip Prindeville <philipp_subx at redfish-solutions.com> wrote:
>> 
>>> Jul 26 17:44:31 mail mimedefang[31780]: t6QNhOYb027848:
>>> smfi_addheader returned MI_FAILURE
>> 
>> Sendmail timed out (that's in the Milter (mimedefang): timeout before
>> data read log line) so it tempfailed the mail and exited.
>> 
>> MIMEDefang, of course, had no idea that happened.  So when filter
>> processing is finally done and mimedefang tries to call
>> smfi_addheader, the original Sendmail process is long gone and the
>> milter pipe is broken, hence the libmilter call fails.
>> 
>> Regards,
>> 
>> Dianne.
> 
> Okay, thanks.  So the next question… how to fix it?  Would increasing the milter timeouts resolve the issue?
> 
> I’m currently configured as:
> 
> Xmimedefang, S=unix:/var/spool/MIMEDefang/mimedefang.sock, F=T, T=S:1m;R:1m;E:5m
> 
> (I guess ‘C’ is just defaulting to 300, which is hardcoded in sendmail/milter.c …)
> 
> Or could it be that something is just hanging indefinitely, and if so, how do I go about fixing this?  Because 1 minute seems like a REALLY long time to wait on sending or receiving data…
> 
> -Philip

Update: I changed the timeout argument to T=S:2m;R:2m;E:10m and rebuilt sendmail.cf, then restarted the service.

The problem seems to have gone away, but I’m not sure why.

-Philip