Re: Named reporting port in use on 12.4-STABLE r372838 GENERIC arm

On Thu, Jan 05, 2023 at 09:19:58AM +0000, John F Carr wrote:
> Do you have both unbound and named enabled?
> 

Certainly not intentionally, and seemingly not at all:

root@ns1:/usr/ports # ps -aux | grep -i unbound
root  7597   0.0  0.2  4788  1884  0  S+   05:44      0:00.01 grep -i unbound

while /etc/defaults/rc.conf contains

local_unbound_enable="NO"       # Local caching resolver
local_unbound_tls="NO"          # Use DNS over TLS

with no reference to unbound in /etc/rc.conf .

Perhaps the error message was spurious, though it certainly
seemed that named wasn't working immediately after reboot
last night when tested with nslookup. Rndc status now reports
the server is up and running, nslookup gets answers.

Named seems to be working, at least for now. Maybe I'm somehow
starting named twice and only noticed the fact yesterday.

Thanks for writing!

bob prohaska






> > On Jan 5, 2023, at 00:18, bob prohaska <fbsd@www.zefox.net> wrote:
> > 
> > After an upgrade to 12.4-STABLE r372838 GENERIC arm on a Pi2 (armv7)
> > BIND 9.16.1 (Stable Release) <id:d497c32>  fails on start with
> > Jan  4 20:41:25 ns1 named[973]: creating TCP socket: address in use
> > Jan  4 20:41:25 ns1 syslogd: last message repeated 2 times
> > Jan  4 20:41:25 ns1 named[973]: unable to listen on any configured interfaces
> > Jan  4 20:41:25 ns1 named[973]: loading configuration: failure
> > Jan  4 20:41:25 ns1 named[973]: exiting (due to fatal error)
> > 
> > Apart from the OS upgrade nothing was changed and it had been
> > working flawlessly for about six months. Re-using the old kernel
> > didn't fix it, has something in userland changed? A slave machine
> > updated yesterday without trouble so I thought it was safe to 
> > update the master.
> > 
> > Thanks for reading, any hints appreciated!
> > 
> > bob prohaska
> > 
> > 
> 
>