Problem with Secondary systems with many zones

[Barry Margolin]

In article <cpbs2d$2c3s$1 at sf1.isc.org>, list3 at wwwcrazy.com wrote:

> Quoting Barry Margolin <barmar at alum.mit.edu>:
> > I remember seeing a similar problem with an 8.3 server when I worked at
> > Genuity, maybe 9.x has a similar problem.
> >
> > We had one customer with lots of zones (they owned a class B network and
> <snip>
> > the files' timestamps as the basis.
> 
> So does Bind choke when it exceeds 30k or 40k domains?

At the peak, Genuity's nameservers had over 30K zones, maybe over 40K, 
and I heard about sites with many more than this (although I suspect 
most of those were web hosting companies, where all the zones were 
really tiny, just containing a "www" entry and maybe an MX record).  We 
were running BIND 8.x at the time I worked there.

> Has anyone been able to stop and restart a secondary name server that has over
> 40k domains that have the same master name server (with zone files currently
> in place)?

One important configuration is to make the slaves *not* send notifies.  
Otherwise, when the server starts up, there's a storm of notifies being 
sent to all the other slaves.

-- 
Barry Margolin, barmar at alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***