Re: DNS problem

On Thu, May 13, 2004 at 09:22:52PM +0200, Florent Thoumie wrote:
> Scott Harrison wrote:
> 
> >Hello,
> >
> >    I have a 4.8 system with stock named 8.3.4-REL and I am getting TONS 
> >of messages in /var/log/messages like:
> >
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(I.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(E.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(D.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(A.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(H.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(C.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(G.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(F.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS 
> >(B.ROOT-SERVERS.NET)
> >May 13 14:36:54 mastodon named[74]: sysquery: no addrs found for root NS ()
> 
> 	Is there a problem with your /etc/namedb/named.root ?
> 
> -- 
> Florent Thoumie                        Epita SRS Promo 2005

No solutions, but a "me too." I've had bind8/named in the base system flake
out about once every three months in the same way. It'll spew those messages
repeatedly until /var gets full, then the real fun starts.

If I notice it before it becomes a real problem, generally just running ndc
restart fixes things up. Unfortunately, it only seems to happen at night when
I'm asleep.

The only thing I have noticed that correlates closely with the odd named
behavior I've noticed here is a drop in the machine's network connection. It
doesn't have to be a long drop, just a few seconds or so, and we're off to the
races. Generally the drop is a result of the next gateway up the line flaking
out, etc.

I'm seriously thinking about switching to bind9 from ports because I've never
seen a bind9 machine behave like this.

The namedb.root I'm running is the box-stock one you get from mergemaster. My
boxes are updated fairly religiously (after security advisories, mainly), so I
don't think namedb.root is the problem.

Sorry for not having anything more useful to add,

Josh
-- 
Josh Tolbert
hemi_at_puresimplicity.net  ||  http://www.puresimplicity.net/~hemi/

If your sysadmin's not being fascist, you're paying him too much.
   --Sam Greenfield