Maintained by: NLnet Labs

Unable to resolv 1 domain

Ondřej Surý
Mon Apr 10 15:18:34 CEST 2017


Perhaps this could be added to things controlled by:

harden-algo-downgrade: yes/no?

I don't think there's any security risk from using SHA1 for DS record
verification even if SHA-2 is available.

Ultimately, it's your call and decision.

Cheers,
-- 
Ondřej Surý <ondrej at sury.org>
Knot DNS (https://www.knot-dns.cz/) – a high-performance DNS server
Knot Resolver (https://www.knot-resolver.cz/) – secure, privacy-aware,
fast DNS(SEC) resolver
Vše pro chleba (https://vseprochleba.cz) – Mouky ze mlýna a potřeby pro
pečení chleba všeho druhu

On Mon, Apr 10, 2017, at 15:10, W.C.A. Wijngaards via Unbound-users
wrote:
> Hi Ondrej,
> 
> On 10/04/17 14:57, Ondřej Surý wrote:
> > I see - the 31653 DS is only algo 1, but the other one is 1,2, but
> > 
> > But RFC 4509 says:
> > 
> > 3.  Implementation Requirements
> > 
> >    Implementations MUST support the use of the SHA-256 algorithm in DS
> >    RRs.  Validator implementations SHOULD ignore DS RRs containing SHA-1
> >    digests if DS RRs with SHA-256 digests are present in the DS RRset.
> > 
> > So perhaps Unbound is too strict here?  There are no known usable
> > attacks on SHA-1 for use in DNSSEC, so I don't think it's necessary to
> > ignore it right _now_.
> 
> But unbound clearly implements the SHOULD and thus should be
> interoperable?  That is what the 'SHOULD' is there for, right?
> So, I am doing this because I think it is the standard.  And I think so
> should you.
> 
> I didn't do this out of strictness, but out of trying to implement
> exactly what the standard said.
> 
> Best regards, Wouter
> 
> > 
> > O.
> > 
> 
> 
> Email had 1 attachment:
> + signature.asc
>   1k (application/pgp-signature)