How does Bind's SDB API handle slow backends, e.g. which require a lot of
network latency? I see that it allows a backend to declare that it is
thread safe, but Bind 9 (by default) only starts one thread per CPU and
this is not enough concurrency. Will Bind start up more threads if
necessary?

What if two queries come in for the same SDB domain in quick succession,
so that the response to the first is not available when the second
arrives? Will the second query stall so that it can re-use the result from
the first? Or would I have to run a cacheing server in front of the master
server in order to get this behaviour?

Tony.
--
f.a.n.finch http://dotat.at/
MALIN HEBRIDES: NORTHEAST 4 OR 5 INCREASING 6. RAIN LATER. GOOD BECOMING
MODERATE.