Re: flexfax: mail to fax gateway: mailer died with signal 13

["Edward S. Marshall" <emarshal@xnet.com>]

On 6 Apr 1998, Eloy A. Paris wrote:
> So, to me it looks like faxmail is dumping core because of a
> segmentation fault (13 in octal is 11 in decimal and it corresponds to
> SIGSEGV).

This sounds like faxmail segfaulting. By any chance was the message which
caused the problem MIME-encoded? I've noticed that faxmail segfaults with
some MIME-encoded mail (wonderful denial-of-service attack possibilities
there ;-).

> Yes, I know about the infamous signal 11 problem and that it almost
> always relates to hardware problems. However, this machine has been up

Nah, that's generally when you're dealing with GCC, not already-running
programs. Segfaults are normal in those situations. ;-)

> :0
> * Subject.*fax
> | /usr/bin/faxmail -n -d eparis@9412863 root
> 
> However, I have just found out that calling faxmail from this
> procmail recipe also causes the "mailer died with signal 13" error.

Yeppers; its probably the exact same problem as I've been seeing here.
Plain messages seem to work, but faxmail goes into a tailspin when
it decodes a MIME message (actually, it looks like it translates it just
fine, but then goes into an infinite loop at the end of the decoding
process).

> The problem is that I can't find any core dump file, "find / -name
> core -type f" gives nothing.

Sendmail might be ulimiting faxmail. Alternatively, you may have an
implicit ulimit on daemon processes started from your system startup
scripts. Try starting sendmail with a coredump size of a reasonable value,
and possibly check sendmail's sources to see if it's putting resource
limits on the programs it runs.

-- 
Edward S. Marshall <emarshal@xnet.com>   -+-   BOFH, UNIX admin, Linux advocate

 Progress (n.):  The process through which the Internet has evolved from smart
 people in front of dumb terminals to dumb people in front of smart terminals.