On Tue, Apr 29, 2025 at 07:19:22PM +0300, Kapetanakis Giannis wrote:
> On 29/04/2025 18:34, Otto Moerbeek wrote:
> > On Tue, Apr 29, 2025 at 06:23:34PM +0300, Kapetanakis Giannis wrote:
> >
> > > On 29/04/2025 18:11, Otto Moerbeek wrote:
> > > > On Tue, Apr 29, 2025 at 04:32:01PM +0300, Kapetanakis Giannis wrote:
> > > >
> > > > > Can it be related to this: dropped due to no socket?
> > > > IIRC OpenBSD does handle sending to a non-blocking socket in the
> > > > multi-threaded case very well. If multiple threads try to do that
> > > > simulteneously, some of them will get a EAGAIN. To test this
> > > > hypothesis, try running with a single thread and see if the errors
> > > > disappear.
> > > >
> > > > -Otto
> > > I did run one-thread and the errors where gone...
> > >
> > > ok, how I can we go to multiple threads now without the errors?
> > kernel code needs to be written for that. PowerDNS Recursor has an
> > ugly workaround to just retry a few times on EBUSY.
> ok. Why don't I have such errors on another more busy DNS system running
> with 8 threads?
> Is there some other os tuning that could be done?
>
> > > Even with the one thread i still get dropped due to no socket on netstat
> > AFAIK that's incoming UDP traffic without a matching socket to deliver to.
> >
> > -Otto
>
> Does this mean that those drops are not a problem correct?
Depends. Something sending you UDP packets you're not ready to process
can be causd by many things.
>
> is net.inet.udp.sendspace and unbound so-sndbuf: related to this?
> Should I get them back to defaults?
No, both are send related.
>
> Thanks Otto.
>
> G
No comments:
Post a Comment