[Pdns-users] TCP nameserver had error
djones at ena.com
Thu May 18 12:07:12 UTC 2017
I have PowerDNS hidden masters (MASTER) managed by PowerAdmin/PowerDNS API using MySQL replication with PowerDNS public slaves that use NOTIFY/AXFR (SLAVE). Every day or two I get this in my hidden master logs when they have just sent a NOTIFY and the public slave tries to do the AXFR:
May 17 00:00:02 server1 pdns_server: May 17 00:00:02 TCP nameserver had error, cycling backend: GSQLBackend lookup query:Could not execute mysql statement: SELECT content,ttl,prio,type,domain_id,disabled,name,auth FROM records WHERE disabled=0 and type=? and name=?: Lost connection to MySQL server during query
Obviously the zone transfer fails so the updates don't get pushed out to the public slaves until their next SOA REFRESH interval happens. If I resend the same NOTIFY right after this error, then the update happens immediately like normal.
My question is how do I troubleshoot this issue? Do I need to make some adjustments to my MySQL connection settings? Should the pdns daemon try to reconnect to the backend to satisfy the request before closing the AXFR TCP connection?
PowerDNS is installed from the official PowerDNS repo:
Thanks for such excellent software.
More information about the Pdns-users