Hi Brian,
Yes, that's exactly what I've done. It does try both masters,
eventually, but seems to get stuck on the "missing" one, and (from the
log messages) seems to back off from doing anything due to the errors.
I'd have expected it to query both masters in quick succession on
receipt of a NOTIFY, and treat an unresponsive master independently of a
responsive one.
This is with the latest Centos 7 RPMs on the 4.1 branch. Also note that
I'm using TSIG and IXFR (I saw the same behaviour with AXFR).
-- don
On 15/12/18 8:03 PM, Brian Candler wrote:
On 15/12/2018 06:29, Don Stokes wrote:
I'm looking to deploy PowerDNS in a slave configuration that has
multiple (BIND) masters. In a test rig I'm finding that if I have
dual masters configured on a PDNS slave, and one of the two masters
is unavailable, the slave is querying the absent master and not
failing over to the other.
And even if it receives and accepts a NOTIFY from the functional
master, it still sends the SOA query to the non-functional master.
The first question is: what version of PDNS Authoritative are you using?
The second is just to confirm how you have configured each zone with
multiple masters in SQL: is it a single row with a comma-separated
list of masters?
INSERT INTO domains (name, master, type) VALUES ('example.com',
'192.0.2.1, 192.0.2.2', 'SLAVE');
--
Don Stokes, d...@nz.net <mailto:d...@nz.net>, 021 796 072
_______________________________________________
Pdns-users mailing list
Pdns-users@mailman.powerdns.com
https://mailman.powerdns.com/mailman/listinfo/pdns-users
