FIXED: sendmail sm-client
I've got this weird problem with sendmail, and I've got the bat book sitting in my lap, but I just can't figure this out.
[root@hostname mail]# service sendmail restart
Shutting down sendmail: [ OK ]
Shutting down sm-client: [FAILED]
Starting sendmail: [ OK ]
Starting sm-client: [ OK ]
[root@hostname mail]# tail /var/log/messages
Nov 15 13:39:47 dhgsrv17 sendmail: sendmail shutdown succeeded
Nov 15 13:39:47 dhgsrv17 sendmail: sm-client shutdown failed
Nov 15 13:39:47 dhgsrv17 sendmail: sendmail startup succeeded
Nov 15 13:39:47 dhgsrv17 sendmail: sm-client startup succeeded
Also, everytime I restart it, I get an extra sendmail queue runner process:
[root@hostname mail]# ps aux | grep sendmail
smmsp 1011 0.0 0.1 3968 1424 ? S 13:52 0:00 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue
smmsp 1056 0.0 0.1 5276 1424 ? S 13:52 0:00 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue
root 1087 0.0 0.1 4680 1544 ? S 13:56 0:00 sendmail: accepting connections
smmsp 1096 0.0 0.1 4996 1428 ? S 13:56 0:00 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue
If I were to restart sendmail again, I'd have 4 queue runner's running.
Should I bump up the debug level in my sendmail config to see if I can get some more info. I'm working totally blind here, I don't have any idea whats causing this.
EDIT: and its not a problem w/ resolution, the /etc/hosts file is perfectly fine. It stops and starts very quickly, just fails the sm-client shutdown and leaves it hanging, just to start another one when the service starts again.
EDIT 2:
Found this in /var/spool/maillog:
Nov 15 13:35:22 dhgsrv17 sendmail[1047]: starting daemon (8.13.1): SMTP+queueing@01:00:00
Nov 15 13:35:22 dhgsrv17 sm-msp-queue[1056]: starting daemon (8.13.1): queueing@01:00:00
Nov 15 13:35:22 dhgsrv17 sm-msp-queue[1056]: unable to write pid to /var/spool/clientmqueue/sm-client.pid: file in use by another process
Nov 15 13:39:47 dhgsrv17 sendmail[1082]: alias database /etc/aliases rebuilt by sysadmin
Nov 15 13:39:47 dhgsrv17 sendmail[1082]: /etc/aliases: 0 aliases, longest 0 bytes, 0 bytes total
Apparently the pid file is in use by another process?
Worth noting:
[root@dhgsrv17 mail]# cd /var/spool/clientmqueue/
[root@dhgsrv17 clientmqueue]# ls -l
total 4
-rw------- 1 smmsp smmsp 50 Nov 15 13:35 sm-client.pid
EDIT:
Just changed my /etc/mail/submit.cf to write the pid file out to /var/run and the problem seems to be solved now. Oh well, at least anyone searching on this problem later will see how I fixed it.
If anyone runs into this, really you should check your /etc/submit.mc, and make sure you have a line like this:
define(`confPID_FILE', `/var/run/sm-client.pid')dnl
Then, in /etc/mail just do a:
make submit.cf
then, just do a "service sendmail restart" and it should clear it up.
Last edited by jon3k; 11-15-2004 at 03:20 PM.
|