Bug 231237

Summary:	Sendmail queue runner does not start
Product:	[openSUSE] openSUSE 10.2	Reporter:	Hubertus Haniel <hubba>
Component:	Network	Assignee:	Dr. Werner Fink <werner>
Status:	RESOLVED DUPLICATE	QA Contact:	E-mail List <qa-bugs>
Severity:	Normal
Priority:	P5 - None	CC:	seth
Version:	RC 5
Target Milestone:	---
Hardware:	i586
OS:	Linux
Whiteboard:
Found By:	Customer	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---

Description Hubertus Haniel 2007-01-02 03:30:40 UTC

The Sendmail queue runner does not start in 10.2 and therefore messages in /var/spool/clientmqueue ever gt processed - this bug is somewhere in the start script as I have now copied the start script from my old 9.2 box which seems to start up everything correctly.

Comment 1 Hubertus Haniel 2007-01-02 10:15:34 UTC

I have now compared the two startup scripts and it looks like the fix fot this bug would be to change line 61 to:
startproc -f -p $msppid -i $srvpid /usr/sbin/sendmail $SENDMAIL_CLIENT_ARGS

At least that works for me :)

Comment 2 Seth Chaiklin 2007-01-03 02:05:06 UTC

Can confirm same problem (and the solution works for me as well)

And while we are at it, there is also a problem with reload.

Demonstration:

27284 ?        Ss     0:00 sendmail: accepting connections
27295 ?        Ss     0:00 sendmail: Queue control
27296 ?        S      0:00 sendmail: running queue: /var/spool/clientmqueue

/etc/init.d/sendmail reload results in the the sendmail: accepting connections
to die.

However, if I issue only the following command from the command line, 
#  killproc -v  -p /var/run/sendmail.pid -i /var/spool/clientmqueue/sm-client.pid -HUP /usr/sbin/sendmail

SIGHUP sendmail(27295) SIGHUP sendmail(27284)

then the queue control dies.

This leads me to believe that killproc is not doing its job in relation to -i

Is it possible that it gets confused by the fact that the sendmail pid files, unlike more pid files, also have a second line with the initiating arguments?
(sorry no more time for experiments)

Comment 3 Hubertus Haniel 2007-01-03 02:17:49 UTC

So maybe this bug is really a sysvinit bug and the behaviour of startproc and killproc has changed? - I have just checked on SuSE 10.1 where the -f flag is not present in the start script but everything starts up fine on 10.1

Comment 4 Dr. Werner Fink 2007-01-08 11:00:41 UTC

Please install the fixed sysvinit package which should fix the startproc
utility.

*** This bug has been marked as a duplicate of bug 228115 ***