Network was down only while reboot.
Indeed those broadcast drops are also strange. I temporarily pf blocked
224/4 as well as all (local) broadcast addresses but still the numbers
of "dropped due to no socket" continued to increase (both lines).
I've also played with net.inet.udp.sendspace but nothing changed.
BTW is this the default limit or the max allowed?
I believe it's just the default and the max is SB_MAX on
/usr/include/sys/socketvar.h going into kernel.
That's why even with the default of 9216, unbound is able to reach up to
to 2m of so-sndbuf
I've disabled outgoing-range since this probably controls outgoing to
remote DNS servers as well as num-queries-per-thread but also nothing
changed :-/
G
On 29/04/2025 17:00, Dan wrote:
> I found instead interesting the "broadcast/multicast datagrams" that let me think that
> that your dest network was probably missing for a while. I dont think it like an unbound problem.
> As if the resource, ie. the dns server, is unreachable unbound doesn't cache.
>
> Apr 29, 2025 15:34:48 Kapetanakis Giannis <bilias@edu.physics.uoc.gr>:
>
>> Can it be related to this: dropped due to no socket?
>>
>> # netstat -ss|grep -A7 ^udp
>> udp:
>> 819120 datagrams received
>> 15494 with no checksum
>> 22621 dropped due to no socket
>> 170569 broadcast/multicast datagrams dropped due to no socket
>> 625930 delivered
>> 6497322 datagrams output
>> 341480 missed PCB cache
No comments:
Post a Comment