Broken DNS QNAME Recovery

Mark Andrews marka at isc.org
Mon Apr 22 21:04:45 UTC 2024


No. “Forward zones” are not DNS zones. They are overrides to the DNS resolution processes that just happened to be configured in named by overloading the zone syntax element.  Similarly stub and static stub are not zones. The are other things. 

-- 
Mark Andrews

> On 23 Apr 2024, at 01:24, John Thurston <john.thurston at alaska.gov> wrote:
> 
>  On 4/21/2024 10:05 PM, Mark Andrews wrote:
>> On 19 Apr 2024, at 16:12, Crist Clark <cjc+bind-users at pumpky.net> wrote:
>>> 
>>> First, yes, I know. Their DNS is broken. They should fix their DNS. We shouldn't need to make QNAME-minimization work around broken DNS.
>>> 
>>> Name and shame a domain name in question,
>>> 
>>>     e1083.d.akamaiedge.akamai.csd.disa.mil
>>> 
>>> The problem I see: akamai.csd.disa.mil is a delegated zone. All four name servers for the zone are in the zone. All four of the addresses in the parent's glue are unresponsive. It's actually the same for d.akamaiedge.akamai.csd.disa.mil too.
>>> 
>>> That is breaking resolution for BIND 9.18 servers with default qname-minimization. If qname-minimization is set "off", it works. That's because the disa.mil NSes will respond with the answer for that full name. We never go farther up the name to try to find the non-responsive NS servers.
>>> 
>>> (And yes, the DNS "authoritative" servers here are questionable too. The TTLs look like they are caching answers, but all of the responses have AA set.)
>>> 
>>> Does that assessment look correct? I know BIND defaults to "relaxed" QNAME minimization. It works around certain cases of brokeness. I guess this is not one of them? Should it be? It's a case where things work without minimization. The brokeness is hidden for non-minimizing resolvers.
>>> 
>>> Again, yeah, they are broken. They should fix it, but it broke someone's Very Important Work at our shop. And it used to work and it works from home and for other customers so it must be our DNS that's broken. So we end up setting "qname-minimization off" globally despite the fact they are really the broken ones. We'd rather keep minimization on, but it's the only reasonable work around we could find.
>> Just use a forward zone in forward only mode.  When the parent servers give you non working nameservers for child zones there isn’t a sensible automatic solution.
>> 
>> zone disa.mil {
>>         type forward;
>>         forward only;
>>         forwarders { 152.229.110.235; 214.3.125.231; 131.77.60.235; };
>> };
> 
> Can such forward-zones be defined in catalog-zones?
> 
> 
> 
> --
> Do things because you should, not just because you can. 
> 
> John Thurston    907-465-8591
> John.Thurston at alaska.gov
> Department of Administration
> State of Alaska
> -- 
> Visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list
> 
> ISC funds the development of this software with paid support subscriptions. Contact us at https://www.isc.org/contact/ for more information.
> 
> 
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/bind-users/attachments/20240423/7d637032/attachment-0001.htm>


More information about the bind-users mailing list