[Pdns-users] Interpreting server variables

[Jan-Piet Mens]

I'd appreciate a bit of help in interpreting some server variables for
a performance comparison chapter.

With the following pdns.conf

        default-ttl=3600
        distributor-threads=3
        launch=gmysql
        log-dns-details=no

        cache-ttl=3600
        negquery-cache-ttl=3600
        query-cache-ttl=3600

I'm getting these values:

        corrupt-packets         0    Number of corrupt packets received
        deferred-cache-inserts  62746    Amount of cache inserts that were deferred because of maintenance
        deferred-cache-lookup   6350    Amount of cache lookups that were deferred because of maintenance
        latency                 2    Average number of microseconds needed to answer a question
        packetcache-hit         2036029    
        packetcache-miss         410882    
        packetcache-size         501350    
        qsize-q                   0    Number of questions waiting for database attention
        query-cache-hit         158314    Number of hits on the query cache
        query-cache-miss         690664    Number of misses on the query cache
        servfail-packets         0    Number of times a server-failed packet was sent out
        timedout-packets    0    Number of packets which weren't answered within timeout set
        udp-answers         2450270    Number of answers sent out over UDP
        udp-queries         2450270    Number of UDP queries received
        udp4-answers         2450270    Number of IPv4 answers sent out over UDP
        udp4-queries         2450270    Number of IPv4UDP queries received

I'm hitting the server with 2.5 million synthetic queries which remain
constant. There are queries for CNAME, A, MX, NS, SOA, etc. which are
being sent in constant fashion to the PowerDNS server..

If I've studied the documentation correctly, it is good that qsize-q
is very low.

I don't understand why packetcache-miss / query-cache-miss are so
high. I'd have thought that the ttl of the query/negative-query would
answer the packet from the cache. Isn't that the case?

Sending the server a constant stream of identical queries (also
queries which result in NXDOMAIN or SERVFAIL) show expected results:

        packetcache-hit 199997  
        packetcache-miss        6       
        packetcache-size        20      
        qsize-q 0       
        query-cache-hit 12      
        query-cache-miss        16

Can someone kindly help me with this, perhaps also pointing out
whether the cache parameters could be sensible?

BTW, Is there a method of configuring a packet cache size?

Thanks,
        -JP