All nameservers unresponsive when master is down

[Simon Waters]

Sam Pointer wrote:
> 
> My domain `hpdsc.com` has 2 "actual" nameservers which we present to the
> world as 4 "logical" nameservers. The 2 boxes are multihomed and all 4
> addresses are given as valid namesevers for the domain (we have 2 leased
> lines and this removes a single point of failure).
> 
> My whois informations is:
> 
>  Domain servers in listed order:
> 
>    NS2.HPDSC.COM                212.158.99.195
>    NS3.HPDSC.COM                195.167.246.3
>    NS4.HPDSC.COM                195.167.246.4
>    NS1.HPDSC.COM                212.158.99.194  The like-typed nameservers
> are actually the same nameserver on the same box (multihomed).

I can't reach ns3 or ns4 from here. Check your routing.

> The relevant part of the Zone data is:
> 
> hpdsc.com.      IN      SOA     ns1.hpdsc.com. dnsadmin.hpdsoftware.com. (
>                                         1018281163      ; Serial
>                                         7200            ; refresh (2 hour)
>                                         3600            ; retry (1 hours)
>                                         604800          ; expire (1 week)
>                                         1               ; neg. cache
>                                         )
> ...
> hpdsc.com.      NS      ns1.hpdsc.com.
> hpdsc.com.      NS      ns2.hpdsc.com.
> hpdsc.com.      NS      ns3.hpdsc.com.
> hpdsc.com.      NS      ns4.hpdsc.com.
> ...
> 
> ns1.hpdsc.com.          IN A            212.158.99.194
> ns2.hpdsc.com.          IN A            212.158.99.195
> ns3.hpdsc.com.          IN A            195.167.246.3
> ns4.hpdsc.com.          IN A            195.167.246.4
> 
> However, if I stop BIND running on ns1 (which is also ns3 if referenced by
> it's alternative IP address) then the whole Zone stops resolving. And I
> can't for the life of me work out why. If you want to tell me to "FAQ off"
> then please do.

I'd start with inserting a line like "$TTL 86400" in the start
of the zone file. (BIND defaults to the negative TTL if you
don't specify it).

A default TTL of 1 second isn't helping things run smoothly.

Still don't see why ns2 should stop if ns1 is down.... ns2
doesn't use ns1 as it's nameserver does it? (i.e. Check
/etc/resolv.conf).

-- 
"Don't get me started on intuitive. You know what's intuitive?
Fear of heights. Everything else we call intuitive, such as
walking or using a pencil took years of practice." - Don Norman