This belated posting is for the benefit of those who experience this
problem later.
If the service is called say foo and you have a stop method other than
:kill which does "pkill foo", then your stop method will fail with
error 208 if foo is a substring of the script that does the stopping.

Example: service name=smbd, stop method name=/lib/svc/method/arl_smbd
if arl_smbd pkills smbd then the signal will terminate both smbd and
srl_smbd thereby not allowing exit 0 or whatever the last line is to be
reached.

Jay G. Scott wrote:
> i tried doing this the right way, ie, with manifests and methods.
> didn't work. so, i copied the methods to /etc/init.d,
> and manually started them there, to prove that the methods, at least
> weren't broken. this does get the daemons to start, and i DO get
> access from my peecee.
>
> /etc/init.d/arl_smbd start
> /etc/init.d/arl_nmbd start
> /etc/init.d/arl_winbindd start
>
> i did join the domain.
>
> well, using those SAME methods and--well, i __hope__--appropriate
> manifests, i can't get them to start. i __believe__ if they'll
> start then i can join the domain later, and... it'll be okay.
>
> but trying to figure out why they aren't starting isn't going well.
>
> dismayed# more site-smbd:smbd.log
> [ May 23 13:10:34 Disabled. ]
> [ May 23 13:10:34 Rereading configuration. ]
> [ Jun 28 12:13:46 Disabled. ]
> [ Jun 28 12:13:47 Rereading configuration. ]
> [ Jun 28 12:30:07 Enabled. ]
> [ Jun 28 13:21:20 Disabled. ]
> [ Jun 28 13:21:25 Enabled. ]
> [ Jun 28 13:21:25 Executing start method ("/lib/svc/method/arl_smbd start") ]
> SAMBA (smbd) server
> [ Jun 28 13:21:26 Method "start" exited with status 0 ]
> [ Jun 28 13:21:27 Stopping because service restarting. ]
> [ Jun 28 13:21:27 Executing stop method ("/lib/svc/method/arl_smbd stop 90") ]
> [ Jun 28 13:21:27 Method "stop" exited with status 208 ]
> [ Jun 28 13:21:27 Executing stop method ("/lib/svc/method/arl_smbd stop 90") ]
> [ Jun 28 13:21:27 Method "stop" exited with status 208 ]
> [ Jun 28 13:21:27 Executing stop method ("/lib/svc/method/arl_smbd stop 90") ]
> [ Jun 28 13:21:27 Method "stop" exited with status 208 ]
>
> tried looking up the ...208 msg on sunsolve, got a bunch of "stop-a" hits.
>
> and it looks like it gets started, and something comes along to kill it.
> ie, it says start exited with status zero.
>
> i presume whatever blunder i've made is common to all three, since i did them
> as close to one another as i could.
>
> dismayed# !ps
> ps -ef | grep mb
> dismayed#
>
> so nothing is running.
> dismayed# svcadm restart smbd
> dismayed# svcs -vx smbd
> svc:/site/smbd:smbd (smbd)
> State: maintenance since Tue Jun 28 13:21:27 2005
> Reason: Method failed repeatedly.
> See: http://sun.com/msg/SMF-8000-8Q
> See: /var/svc/log/site-smbd:smbd.log
> Impact: This service is not running.
>
> dismayed# svcadm clear smbd
> dismayed# svcadm restart smbd
> ismayed# svcs -vx smbd
> svc:/site/smbd:smbd (smbd)
> State: maintenance since Tue Jun 28 13:52:16 2005
> Reason: Method failed repeatedly.
> See: http://sun.com/msg/SMF-8000-8Q
> See: /var/svc/log/site-smbd:smbd.log
> Impact: This service is not running.
> [ Jun 28 13:52:09 Executing start method ("/lib/svc/method/arl_smbd start") ]
> SAMBA (smbd) server
> [ Jun 28 13:52:09 Method "start" exited with status 0 ]
> [ Jun 28 13:52:16 Stopping because service restarting. ]
> [ Jun 28 13:52:16 Executing stop method ("/lib/svc/method/arl_smbd stop 98") ]
> [ Jun 28 13:52:16 Method "stop" exited with status 208 ]
> [ Jun 28 13:52:16 Executing stop method ("/lib/svc/method/arl_smbd stop 98") ]
> [ Jun 28 13:52:16 Method "stop" exited with status 208 ]
> [ Jun 28 13:52:16 Executing stop method ("/lib/svc/method/arl_smbd stop 98") ]
> [ Jun 28 13:52:16 Method "stop" exited with status 208 ]
>
> ummm. does this tell anybody anything?
>
> if i edit the manifest, do i have to re-import every time?
>
> j.
> --
> Jay Scott 512-835-3553 gl@arlut.utexas.edu
> Head of Sun Support, Sr. Operating Systems Specialist
> Applied Research Labs, Computer Science Div. S224
> University of Texas at Austin

exlover@ml1.net <exlover@ml1.net> wrote:
> This belated posting is for the benefit of those who experience this
> problem later.
> If the service is called say foo and you have a stop method other than
> :kill which does "pkill foo", then your stop method will fail with
> error 208 if foo is a substring of the script that does the stopping.
>
> Example: service name=smbd, stop method name=/lib/svc/method/arl_smbd
> if arl_smbd pkills smbd then the signal will terminate both smbd and
> srl_smbd thereby not allowing exit 0 or whatever the last line is to be
> reached.

This is why pkill has a -x option :

-x Considers only processes whose argument
string or executable file name exactly
matches the specified pattern to be matching
processes. The pattern match is considered
to be exact when all characters in the pro-
cess argument string or executable file name
match the pattern.

Scott