Greetings folks,

Ever since I implemented the "fixcr" addition to "smtpd", I've noticed that 
over time my qmail servers accumulate a large number (several dozen) pairs 
of processes: "fixcr" and "sh -c fixcr | /usr/qmail/bin/qmail-smtpd".

----------------------------------------------
ps -alxww

  UID   PID  PPID CPU PRI NI   VSZ  RSS WCHAN  STAT  TT       TIME COMMAND

 1002 45088 88130   0  10  0   496  244 wait   I     ??    0:00.00 sh -c 
fixcr | /usr/qmail/bin/qmail-smtpd
 1002 45105 45088   0   2  0   752  380 sbwait S     ??    0:00.08 fixcr
...
88130  p1- S      2:06.75 tcpserver -c 600 -x /etc/tcp.smtp.cdb -u 1002 -g 
1001 0 smtp sh -c fixcr | /usr/qmail/bin/qmail-smtpd
----------------------------------------------

When I kill -9 a "fixcr" process, it ends the corresponding "sh -c fixcr | 
/usr/qmail/bin/qmail-smtpd" process, which makes it appear that for some 
reason "fixcr" is having trouble exiting cleanly. Other than an occasional 
"killall -9 fixcr" that I have to run, it doesn't seem to cause any 
problems except for concern by the mail administrator (me :) that something 
isn't configured properly. Speaking of configuration, here's what I use:

tcpserver -c 600 -x /etc/tcp.smtp.cdb \
-u `id -u qmaild` -g `id -g qmaild` 0 smtp \
sh -c 'fixcr | /usr/qmail/bin/qmail-smtpd' &

Any suggestions? Thanks!

Dave

Reply via email to