On Wednesday, October 27, 2010 11:15 AM, Mike Gerdts wrote:
On Tue, Oct 26, 2010 at 9:43 PM, Albert Lee<tr...@opensolaris.org> wrote:
On Mon, Oct 25, 2010 at 11:53 PM, Christopher Chan
<christopher.c...@bradbury.edu.hk> wrote:
So I ran truss on the master smbd process...
It appears that something fired a SIGTERM arrow at it and then tried to
cancel it with a SIGCONT arrow or something?
Not sure about the cause of the crash, but the truss output appears to
be Samba trying to perform an orderly shutdown after it gets the first
SIGTERM from the SMF stop method.
This is likely a result of SMF detecting that something in the service
is sick (dumped core) and as such is performing a restart. The
default configuration of the apache22 service is smart enough to
delegate this work to httpd, which already knows how to deal with
children that core dump. From
/lib/svc/manifest/network/http-apache22.xml:
<property_group name='startd' type='framework'>
<!-- sub-process core dumps shouldn't restart
session -->
<propval name='ignore_error' type='astring'
value='core,signal' />
</property_group>
Perhaps the samba smf service requires a similar configuration.
Ah, so that's where I can tell SMF to leave things alone. I concur.
Running smbd unmanaged has resulted in zero samba downtime. As for the
core dumps...
In the logs: smbd[22213] setid process, core not dumped:
coreadm
global core file pattern: /var/cores/%d/%f.%p.%n
global core file content: all
init core file pattern: core
init core file content: default
global core dumps: enabled
per-process core dumps: enabled
global setid core dumps: enabled
per-process setid core dumps: enabled
global core dump logging: enabled
What am I missing?
_______________________________________________
OpenIndiana-discuss mailing list
OpenIndiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss
