[Unbound-users] getting to SERVFAIL faster

[Tim Draegen]

Heyo,

I run a service that performs a lot of concurrent DNS queries, primarily big batches of PTR lookups.  I've prototyped and tested using unbound, and double-pass it's good.

I have one issue, which is that ~95% of lookups I'm performing resolve quickly, but the remaining 5% can take upwards of a minute to return "SERVFAIL".  For my application, I'd like to put in a ceiling of ~dozen seconds before shutting down the batch and returning "SERVFAIL".

I'm happy to dig in and create a patch if necessary.  If this is the route to go, I'd much appreciate some pointers on design/code-placement to get started.

Thank you much!
=- Tim