printing, locking, etc in 2.2.0

[Andrew Bartlett]

Jeremy Allison wrote:
> 
> On Tue, May 22, 2001 at 07:44:05PM +1000, Andrew Bartlett wrote:
> >
> > Just a quick 'me too', with a CVS snapshot from over the weekend, I have
> > SIG11s (I'll investigate seperately) and the resultant dead deamon
> > giving logfiles like this:
> >
> > Sorry its stuck as a level 2 syslog, I'll look into turning up the
> > logging.
> >
> > May 22 14:47:54 jessie smbd[12512]:
> > ===============================================================
> > May 22 14:47:54 jessie smbd[12512]: INTERNAL ERROR: Signal 11 in pid
> > 12512 (2.2.0)
> > May 22 14:47:54 jessie smbd[12512]: Please read the file BUGS.txt in the
> > distribution
> > May 22 14:47:54 jessie smbd[12512]:
> > ===============================================================
> > May 22 14:47:54 jessie smbd[12512]: PANIC: internal error
> > May 22 14:47:54 jessie smbd[12512]:
> 
> Ok - that would explain why a lock record is left (exceptional
> exit condition). Andrew, can you add one of the following
> to your smb.conf
> 
> panic action = /usr/bin/X11/xterm -display :0 -e gdb /usr/local/samba/bin/smbd %d
> 
> (replace :0 with a desktop display you've enabled access to)
> 
> or even use :
> 
> panic action = /bin/sleep 9000000
> 
> This will get a process you can attach to with gdb and
> get a stack backtrace. This is very important to get - I
> don't want to ship 2.2.1 with a crash bug.
> 

I've had last nights CVS (and my pam mods) installed today and the
logfiles show no crashing deamons.  I have since installed this mornings
CVS (ie the same thing) but by a normal 'make install', not by RPM as I
normally do.  (I am assuming here that RPM would have stripped all the
debugging info). I will report back if I catch any crashes.

Andrew Bartlett
abartlet at pcug.org.au

-- 
Andrew Bartlett
abartlet at pcug.org.au