EDNS Compliance

[Warren Kumari]

On Fri, Jan 18, 2019 at 12:07 PM Ben Croswell <ben.croswell at gmail.com>
wrote:

> As long as all 4 DNS servers are running the same version, my first
> suggestion would be to check firewalls for dropped packets.
>
> Some FW/IPS drop packets with edns versions other 0 because they see it as
> an attack.
>

This can be generalized to "Some FW/IPS drop packets".
A huge number of nameservers are running with their nameserver software
directly exposed on the Internet (and the rest of their services protected
by iptables / stateless ACLs) - this leads to better stability,
performance, and predictability - the simplification usually also leads to
better security - being able to understand the system and what the (lack
of) firewall is doing make it simpler and easier to protect.

Either your "firewall" is doing really deep inspection and understanding of
the DNS protocol (in which case you are relying on the ALG to be fully
compliant with all behaviors), or you have disabled all ALG work, in which
case the firewall is simply adding another point of failure (and likely
building state, making troubleshooting harder,etc).

Roland Dobbins had some good articles about the fragility and security
decrease caused by stateful devices in front of Internet service type
protocols (such as DNS,etc).

Warren "Fully expecting FW vendor flames" Kumari.



> On Fri, Jan 18, 2019, 12:02 PM N. Max Pierson <nmaxpierson at gmail.com
> wrote:
>
>> Hi List,
>>
>> I am trying to ensure our Bind servers comply with EDNS for the upcoming
>> Flag Day (https://dnsflagday.net/). I am somewhat ignorant to EDNS but
>> from what I have read, the information is somewhat conflicting as some
>> documentation states EDNS is not a record that you configure in your zone
>> file then other sites refer to some sort of OPT record you can configure.
>> So my first question is which of the documentation is correct from what I
>> have read? Is it DNS server functionality that supports EDNS or do you also
>> have to configure something in the zone files?
>>
>> Also, I have 4 (well 5 counting the master that isn't queryable)
>> nameservers with multiple domains served on them. When I run one of my
>> primary domains through the ISC EDNS tool, it comes back as 2 out of the 4
>> are failing EDNS queries.They are all on the same version of Bind
>> (9.8.2rc1) and they are all slaves of the master so they should all have
>> the same records. Can anyone please explain what I need to do to resolve
>> the timeouts listed on the ISC testing tool?
>>
>> Here is what the tool says ...
>>
>>
>> venyu.com. @208.79.48.30 (ns4.venyu.com.): dns=ok edns=ok *edns1=timeout*
>>  edns at 512=ok ednsopt=ok *edns1opt=timeout* do=ok ednsflags=ok
>> docookie=ok edns512tcp=ok *optlist=timeout*
>>
>> venyu.com. @69.2.33.250 (ns1.venyu.com.): dns=ok edns=ok edns1=ok
>> edns at 512=ok ednsopt=ok edns1opt=ok do=ok ednsflags=ok docookie=ok
>> edns512tcp=ok optlist=ok
>> venyu.com. @2604:d800:12::250 (ns1.venyu.com.): dns=ok edns=ok edns1=ok
>> edns at 512=ok ednsopt=ok edns1opt=ok do=ok ednsflags=ok docookie=ok
>> edns512tcp=ok optlist=ok
>>
>> venyu.com. @69.2.63.250 (ns3.venyu.com.): dns=ok edns=ok edns1=ok
>> edns at 512=ok ednsopt=ok edns1opt=ok do=ok ednsflags=ok docookie=ok
>> edns512tcp=ok optlist=ok
>> venyu.com. @2604:d800:13::250 (ns3.venyu.com.): dns=ok edns=ok edns1=ok
>> edns at 512=ok ednsopt=ok edns1opt=ok do=ok ednsflags=ok docookie=ok
>> edns512tcp=ok optlist=ok
>>
>> venyu.com. @208.79.48.26 (ns2.venyu.com.): dns=ok edns=ok *edns1=timeout*
>>  edns at 512=ok ednsopt=ok *edns1opt=timeout* do=ok ednsflags=ok
>> docookie=ok edns512tcp=ok *optlist=timeout*
>>
>>
>> TIA!!
>>
>> Regards,
>>
>> Max
>> _______________________________________________
>> Please visit https://lists.isc.org/mailman/listinfo/bind-users to
>> unsubscribe from this list
>>
>> bind-users mailing list
>> bind-users at lists.isc.org
>> https://lists.isc.org/mailman/listinfo/bind-users
>>
> _______________________________________________
> Please visit https://lists.isc.org/mailman/listinfo/bind-users to
> unsubscribe from this list
>
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users
>


-- 
I don't think the execution is relevant when it was obviously a bad idea in
the first place.
This is like putting rabid weasels in your pants, and later expressing
regret at having chosen those particular rabid weasels and that pair of
pants.
   ---maf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/bind-users/attachments/20190118/31b41b3a/attachment.html>