[Mimedefang] Slaves dying!

[Nels Lindquist]

> I'm running into an odd problem with MIMEDefang.  A particular thread 
> from the SA-Talk mailing list is causing slaves to segfault (signal 
> 11) while all other mail traffic is flowing smoothly.  I've made no 
> changes to the filter.

I thought maybe the problem was due to an older version of MD (I was 
running 2.33).  I've now upgraded to 2.39, but running with or 
without the embedded perl interpreter, I get the same results:

> Nov 26 17:22:28 aerodrome mimedefang-multiplexor: Slave 3 died
> prematurely -- check your filter rules 
> Nov 26 17:22:28 aerodrome mimedefang-multiplexor: Reap: Idle slave 3
> (pid 11548) exited due to signal 11 (SLAVE DIED UNEXPECTEDLY) 
> Nov 26 17:22:28 aerodrome mimedefang-multiplexor: Slave 3 resource
> usage: req=1, scans=1, user=3.260, sys=0.150, nswap=0, majflt=357,
> minflt=3591, maxrss=0, bi= 0, bo=0 
> Nov 26 17:22:28 aerodrome mimedefang[11683]: Error from multiplexor:
> ERR No response from slave 
> Nov 26 17:22:28 aerodrome mimedefang[11683]: hAR0MOLn011682: Filter
> failed.  Message kept in /var/spool/MIMEDefang/mdefang-hAR0MOLn011682 

I tried snagging the failed directories again; I now receive the 
following error message if I run mimedefang.pl directly on one of the 
failed directories:

> Cannot chdir(mdefang-hAR0MOLh011682/): No such file or directory at
> /usr/bin/mimedefang.pl line 412. 

This has really got me stumped; the four messages are currently 
queued on our secondary MX which is still running version 2.33--the 
problem didn't occur there.

What else can I look at?

----
Nels Lindquist <*>
Information Systems Manager
Morningstar Air Express Inc.