Hello, when supermaster send NOTIFY for large number of domain I think
some NOTIFYs get re-sent. On first time slave setup that cause errors so
I was looking do pdns have setting to delay re-NOTIFY timeout?
Loglevel 6 doesnt say "no response so I will re-notify" however I see
hint of two notify response in master:
01:19:53 Queued notification of domain 'example.org' to 2.2.2.2:53
01:19:57 Removed from notification list: 'example.org' to 2.2.2.2:53
(was acknowledged)
01:20:13 Received spurious notify answer for 'example.org' from
2.2.2.2:53
01:20:31 AXFR happen successfully
01:20:45 Received unsuccessful notification report for 'example.org'
from 2.2.2.2:53, error: Server Failure
01:20:45 Received spurious notify answer for 'example.org' from
2.2.2.2:53
Slave show in logs 2 NOTIFY, notice time of number 2 is 1sec before than
which master said "removed from notify list (acknowledged)":
01:19:54 Received NOTIFY for example.org from 1.1.1.1 for which we are
not authoritative
01:19:56 Received NOTIFY for example.org from 1.1.1.1 for which we are
not authoritative
01:20:13 Created new slave zone 'example.org' from supermaster 1.1.1.1
01:20:31 AXFR happen successfully
01:20:44 Database error trying to create example.org for potential
supermaster 1.1.1.1: Database error trying to insert new domain
'example.org.': Could not execute mysql statement: insert into domains
(type,name,master,account,last_check,notified_serial)
values(?,?,?,?,NULL,NULL): Duplicate entry 'example.org' for key
'name_index'
So I guess master wait 3sec for NOTIFY acknowledge then send re-NOTIFY.
On slave setup causing database error. Look like operation is normal
after so I can ignore but I wonder if there may be setting to increase
timeout?
_______________________________________________
Pdns-users mailing list
Pdns-users@mailman.powerdns.com
https://mailman.powerdns.com/mailman/listinfo/pdns-users
