coredns: Debug plugin option ignored in CoreDNS 1.8.4

What happened: When CoreDNS version bumped up from 1.6.1 to 1.8.4, debug outputs are not written to the output, although debug option is present in the configuration.

What you expected to happen: Debug outputs written to the output when debug plugin is active.

How to reproduce it (as minimally and precisely as possible): We’ve experienced it with dig @localhost -p 40000 google.com when logging debug outputs with zipkin trace ids in internal logging plugin, but it should be reproducible with any debug logging using coredns@1.8.4/plugin/pkg/log/plugin.go debug() or debugf() function.

Anything else we need to know?: coredns@1.8.4/plugin/pkg/log/plugin.go debug() or debugf() functions use if !D.Value() to ignore debug outputs when debug plugin is not active. Having debug plugin active in Corefile for several ports in the same configuration block is not reflected, D.Value() == false (see Corefile example). Seems like the configuration is not being read properly.

Environment:

• the version of CoreDNS: 1.8.4
• Corefile:

.:8000 { health pprof }

.:40000, .:40001, .:40002, .:40003 { error_trace debug forward . 64.6.64.6 8.8.8.8 log . “plugin/log: [{>id}] ‘{type} {class} {name} {proto}:{port} {size} {>do} {>bufsize}’ {rcode} {>rflags} {rsize} {duration}” ready any }