[Pdns-users] powerdns master, bind slave, slave not updating...

[Mathew Hennessy]

Hi,
     I'm in the process of evaluating PowerDNS (2.9.21 src tarball built on
Solaris 10) and while I've gotten my bind zones imported into a gmysql
default-template backend and it responds correctly, and I've been able to
point a slave BIND (9.4.1-P1) to it and have it acquire initial zones, it
never seems to update based on serial changes when left alone or HUPd.  The
original config was pointing to a BIND master, all I did was do a vi
search/replace in its named.conf to change the BIND master IP to the
PowerDNS master IP.

The errors I'm getting in syslog on the powerdns master are:
Feb 29 16:40:20 powerdns-dev pdns[13444]: [ID 702911 daemon.warning] Not
authoritative for 'hfnmag.com', sending servfail to 10.202.108.22

On the slave side:
29-Feb-2008 16:42:55.680 general: info: zone hfnmag.com/IN: refresh:
unexpected rcode (SERVFAIL) from master 10.202.108.20#53 (source 0.0.0.0#0)

I know PowerDNS doesn't set authoritative flags, and this box is an eval
(dev) server at this point so nobody in whois will ever know about it.  I am
not trying to run recursion in any way, I just want a proof-of-concept so I
can show my management that PowerDNS will be able to take over for our
master DNS and pass zones onto BIND slaves until such time as we migrate the
slaves to PowerDNS at our leisure.

How can I get this BIND slave to actually refresh its zones off the master
without wiping the zone files and start/stopping named in cron? 

Any help appreciated!
-- Matt

config files (with some edits)
http://www.nabble.com/file/p15768835/named.conf named.conf 
http://www.nabble.com/file/p15768835/pdns.conf pdns.conf 
-- 
View this message in context: http://www.nabble.com/powerdns-master%2C-bind-slave%2C-slave-not-updating...-tp15768835p15768835.html
Sent from the PowerDNS mailing list archive at Nabble.com.