1.9.4: TCP queries when some threads are full
he at uninett.no
Mon Nov 25 20:48:46 UTC 2019
Back in October, Wouter replied:
> You should probably increase that count to 100 or better 1000.
> incoming-num-tcp: 1000
> And perhaps also increase the upstream (outgoing-num-tcp) if you have
> upstream TCP or TLS configured.
> This allocates more buffers and that is useful for a server with more
> clients on it.
> The setting is actually similar to a setting for NSD that controls
> similarly the number of buffers for client streams.
Someone directed me to this thread after my recent battle with
unbound for dns-over-tls.
Coming from a Unix background, I must say I'm slightly puzzled by the
use of the term "buffer" here. I think a more appropriate term to use
would be "file descriptor". In my understanding, a buffer is usually
a data structure in the application used to temporarily store data,
possibly coming from a file descriptor (but not necessarily), and
there *is* something called "unbuffered read". So a "buffer" does not
necessarily map 1:1 to "file descriptor", at least not to my ears.
Also, a "buffer" isn't an OS resource, whereas a "file descriptor"
definately *is* an OS resource, and what's important to describe here
is the (potential) OS resource use.
Therefore, I *think* what is being described here is "file
I'm going to suggest a few updates to the unbound.conf man page
via a pull request shortly...
More information about the Unbound-users