[dnsdist] Handling lack of caching of TC responses

[Dan McCombs]

Hello,

I've noticed some queries recently for hostnames with large sets of A
records which return a NOERROR response with no answers and the TC flag set
when queried over UDP. There's no TTL on anything, so nothing is cached.
The client then retries over TCP, and gets the full answer, which is cached
by dnsdist.

The client then queries again starting over UDP, which doesn't hit cache,
and goes through to the downstream, before retrying over TCP. This means
that while the full response is cached and is served the next time the
client queries over TCP, the UDP queries keep going all the way through to
the downstream.

Is that something that's expected to happen once the full response has been
returned from a downstream over TCP? Is there some way to force TC
responses to have at least some minimal TTL? Or some way to have dnsdist
use its cached response rather than going back to the downstream in that
case on UDP queries?

An example hostname I was seeing this happen with is: grid-use.bidswitch.net

Anyone have any clever tricks to ease the load on the downstreams?

Thanks,

-Dan



Dan McCombs
Senior Engineer I - DNS
dmccombs at digitalocean.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.powerdns.com/pipermail/dnsdist/attachments/20221117/9017e4a0/attachment.htm>