restarting named...takes a long long time ???

[Barry Margolin]

In article <9m65ln$dqn at pub3.rc.vix.com>,
Frank Benady <f.benady at french-connexion.fr> wrote:
>I run Bind 8.2.3-REL on Linux Red Hat 6.0
>
>named serves about 4000 zones on the server.
>
>The pb : it takes 20 up to 30 minutes to restart named with the following
>command
>
>/usr/sbin/named -c /etc/named/named.conf
>
> Is it normal ? Is it because of the 4000 db.files it has to read ?

Yes.  It takes time to read and process thousands of zone files.

>Moreover, when named restarts the syslogd start eating up all the ressources
>on the server.

Unless you have a multiprocessor machine, syslogd and named are fighting
for CPU time and disk bandwidth.

>Do I have to stop syslod before restarting named ?

It would probably be better to adjust what named logs.  For every domain
that it loads, it sends a log message stating what it's doing.  If you
don't need them, you can do:

logging {
  ...
  category load { null; };
};

Also, check your log to see whether you're reporting lots of warnings while
loading.  For instance, if you don't have $TTL statements in your zone
files, it will log "No default TTL set, using SOA Minimum instead" for
every domain.

-- 
Barry Margolin, barmar at genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.