[Mimedefang] mimedefang lockup - help needed

[Phil Eschallier]

On Tue, 2003-05-20 at 00:19, Bill Randle wrote:
> I've got a script run by cron every two minutes that checks for an
> abnormally high number of sendmail processes and restarts mimedefang.
> To this I add a "ps -ax|grep mime" and a tail of the mail info file.
> The output shows the mimedefang process itself is still around, although
> it is sleeping:
> 

Bill;

I was something similar happen before I moved /var/spool/MIMEDefang to a
tmpfs (memory) file system -- mimedefang could just not keep up would
essentially hang.  Are you using a tmpfs file system for that spool
directory?

Though in your case, I remember reading about this type of problem when
there was an incorrect path to the milter socket file in the sendmail.mc
INPUT_MAIL_FILTER macro.  This seems unlikely here as things work at the
start ... so it could be that something is changing the permissions on
the socket file of the directories at some point after launch. 
Personally, again before using a tmpfs file system, I experienced a
situation where the the server got so I/O bound that stopping and
restarting mimedefang actually made things worse (it took so long for
mimedefang to remove the old socket file -- the new processes had
started and the shutdown of the old actually removed the new socket
file).

If nothing is apparent after checking path names and permissions, try
running sendmail with -d44.4 to trace the safefile processing, it may
lead to a more descriptive error message.

... Phil