[Samba] Workaround for bind9 reload bug : samba_dlz Ignoring duplicate zone

[Harry Jede]

Am Montag, 12. März 2018, 18:13:09 CET schrieb Prunk Dump via samba:
> Hi samba team !
> 
> I'm face with a new problem on a new Samba PDC install (Debian 9).
> 
> I don't know why, but systemd run multiples "rndc reconfig" commands
> during the init script. So the bind9 log file show :
> -> A successful start
> -> A failed reconfig (samba_dlz Ignoring duplicate zone)
> at each boot/reboot.
> 
> So I need to restart bind9 each time manually. I created a wrapper
> script and moved rndc to rndc.ORIG to intercept the reload and
> reconfig commands. But I don't like this solution as rndc is a binary
> and moving it prevent security updates to be applied on it.
> 
> Is there a way to make "rndc reload" and "rndc reconfig" working with
> samba ?
I believe this comes from /etc/init.d/bind9.
Change this init script to avoid playing with rndc.

comment the reload part:
#    reload|force-reload)
#        log_daemon_msg "Reloading domain name service..." "bind9"
#        if ! check_network; then
#            log_action_msg "no networks configured"
#            log_end_msg 1
#        fi
#
#        /usr/sbin/rndc reload >/dev/null && log_end_msg 0 || log_end_msg 1
#    ;;

switch reload to restart:
    restart|reload|force-reload)

Yes, it is slower. Yes, it interrupts the bind service for a short time.
So what. Boot error is gone.
 
> Is someone have an idea whose service send these "rndc reconfig"
> command ? I can't find the origin.
> 
> Thanks !
> 
> Baptiste.


-- 

Gruss
	Harry Jede