Help w/specific site after upgrade

[Dawn Lovell]

On Thu, 18 Jul 2002, Simon Waters wrote:
>Serial number 22, any bets they "upgraded" to NT DNS when you
>upgraded BIND, or is that cynical ;)

Heh. Cynical, maybe, but not implausible. :-)

>The delegated servers vary between the parent and child<tsck
>tsck>, and the child zones authoritative name server is
>unreachable (on colo.sonicwall.com which appears to be private
>address space). This typically manifests as domains that resolve
>first time but not again, and other intermittent problems.
>
>The delegation of both the global and colo subdomains appear to
>be completely wrong, in that they both point directly or
>indirectly to servers on private address space, and leak private
>address replies to the global Internet <tsck>.

This is the classic stupid question, but... how does the primary
DNS information (MNAME) in the SOA get used by the resolver?  In
concert somehow with the NS RRs or as more of a comment?

I didn't know if we could pin the issue entirely on their end,
since we are consistently able to resolve the global.sonicwall.com
sites using our test 9.2.1 DNS.  I've restarted, flushed, etc.,
and it resolves.  I thought maybe it was something like the
9.2.1 box keeping the positive result cached longer since it
sees so little traffic, but it always gets the info again after
I flush the cache.

>Oh well looks like the mail address will resolve, even if
>www.sonicwall.com is unreachable - so if your lucky the
>postmaster might check it out.

One of our sonicwall users had a contact in their IT department,
so we've passed along this info.  Hopefully we'll hear back
something from either that or the postmaster.

>Do you want to buy a firewall from a company that understands
>DNS, because I might be able to help you there ;)

That would be supposing anyone asked my opinion about firewall
software/hardware. ;-)

Thanks, again!  I really appreciate your time.

Dawn