SUMMARY: Runaway Sendmail

From: David Graves (admin@graves.com)
Date: Sat Sep 14 1996 - 10:38:52 CDT


Thankyou for the replies.

There was no definitive answer as to what might be causing runaway
sendmails, however there were some suggestions as to possible causes.

Tim Carlson <tim@santafe.edu> wrote that he had experienced this situation
when a huge file was mailed. The answer then, is not to mail huge files.
(>50meg)

sanford@lsil.com (Sanford Whitehouse) wrote that the version of Microsoft
Exchange shipped with win 95 was causing this runaway condition. When the
Exchange was upgraded, all was fixed.

Salvatore Saieva <saieva@ppllc.com> writes that truss(1) could be used to
find out what actually is going on with the runaway sendmail.

Jim Harmon <jim@telecnnct.com> wrote:

I had a bad-mailer on a single user account (pine) while our other users
were all using other mailers (all ontop of sendmail), where we kept
getting messages bounced (excessive hops) from that account. He had a
private alias that directed mail to himself on 2 different machines who
both forwarded to the other machine. (Endless loops)

The only way I found it was by reading the hops headers back to the
source.

I also had a request to see the script that I use to kill the runaway
sendmails. This script I actually stole from another thread, but will
reproduce here:

#!/bin/sh
# a shell script to determine if there are runaway sendmails.

PIDS=`ps -ec | grep -v defunct | grep -v PID | grep -v oninit | grep sendmail |
sed -e 's/:/\. /' | awk ' { if ($5 > 10 ) printf("%s,", $1); } '`

#This kills sendmail if its value increases above 10. Adjust to your own
comfort #level.

echo ${PIDS:=none} >/dev/null

if [ $PIDS = "none" ]
then
exit
fi

#following is solaris script to stop sendmail -- adjust to your system

sh /etc/rc2.d/S88sendmail stop
sleep 5
sh /etc/rc2.d/S88sendmail start

#mail me to let me know
echo "Runaway Sendmail" | mailx -s"sendmail" dgraves

====================================

I'm running Solaris 2.4 on a SPARC 20, 75mhz, 128meg RAM

I've got about 1,700 accounts on this machine

Every now and then -- about every two days, but occasionally more often, I
get a runaway sendmail process. I have a script to sense and kill the
runaway sendmails, but I'm wondering if anyone knows of what is causing it
and if there's anything to be done to keep it going.

Thanks in advance.

-----
NetReach, Inc. -- Pennsylvania Regional Internet Access and Development
support@netreach.net

"In theory, there is no difference between theory and practice, but in
practice, there is." - Author unknown



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:10 CDT