BIND 9.3.0 coring on SIGHUP (FreeBSD 5.3-RELEASE)

[Robert Lowe]

phn at icke-reklam.ipsec.nu wrote:

> Alex R <alex at deviousmeans.net> wrote:
> 
>>I compiled BIND 9.3.0 from ports on FreeBSD 5.3-RELEASE and every time =
> 
> I
> 
>>SIGHUP the server by doing killall -HUP the server cores. When I run na=
> 
> med
> 
>>in the foreground with -g the server does NOT core but instead I just g=
> 
> et:
> 
>>09-Nov-2004 13:54:44.569 loading configuration from '/etc/namedb/named.=
> 
> conf'
> 
> 
>>09-Nov-2004 13:54:44.601 ignoring config file logging statement due to =
> 
> -g
> 
>>option
> 
> 
>>Has anyone else had this problem or have a solution to it?
> 
> 
> Never tried SIGHUP since bind-4 days. rndc (rndc stop) is there for a rea=
> son, use it!

SIGHUP was the equivalent of rndc reconfig, IIRC.  Stop wasn't even a concern
before DDNS.  Still, the threads issue in FreeBSD 5.3-RELEASE is a bit
disconcerting, assuming the problem did not exist prior to this.  What negative
implications are there in BIND when setting the thread contention scope to
process vs. system?  Worse than not having threads at all?

-Robert