0

I'm setting up a new PowerDNS server version 2.9.22 running on Debian Lenny in a Xen instance, and for some reason it is reporting that the packetcache size is 0 when I run a /etc/init.d/pdns dump:

corrupt-packets=2,deferred-cache-inserts=232,deferred-cache-lookup=59,latency=1,packetcache-hit=6616,packetcache-miss=4111,packetcache-size=0,qsize-q=1,query-cache-hit=5938,query-cache-miss=8516,recursing-answers=0,recursing-questions=0,servfail-packets=0,tcp-answers=0,tcp-queries=0,timedout-packets=0,udp-answers=10702,udp-queries=10748,udp4-answers=10702,udp4-queries=10748,udp6-answers=0,udp6-queries=0

This was taken in the middle of a dns benchmark run so the packet cache should be populated with something also given that packetcache-hit is increasing shows that its hitting against something but I don't know what. cache-ttl,negquery-ttl and query-cache-ttl are all set to 60 seconds.

This always wasn't a problem, only popped up today when I restarted the instance and it just stopped working, I'm drawing a blank what I might have changed in the instance to cause this.

Can anybody point me in the right direction on how to fix this issue?

Malvane
  • 1
  • 2

1 Answers1

0

I am not sure what your benchmark tests for but if you have packet cache hits and a packet cache size of 0 that might imply that all of the queries so far were for non existent records i.e. those that will return an NXDOMAIN response. I also see that you have about 75% of your packet cache hits in packet cache miss so these must all be queries for things the packet cache has not previously seen.

I think you are running a benchmark intended for recursive DNS on your authoritative PowerDNS Server. Testing for NXDOMAIN responses is essential to authoritative DNS Operation too but it will be equally worth testing for qnames that your server actually knows about i.e. that will result in NOERROR answers.

ZaphodB
  • 653
  • 3
  • 9