Network Working Group W. Kumari
Internet-Draft Google
Intended status: Standards Track E. Hunt
Expires: April 19, 2018 ISC
R. Arends
W. Hardaker
D. Lawrence
Akamai Technologies
October 16, 2017

Extended DNS Errors


This document defines an extensible method to return additional information about the cause of DNS errors. The primary use case is to extend SERVFAIL to provide additional information about the cause of DNS and DNSSEC failures.

[ Open question: The document currently defines a registry for errors. It has also been suggested that the option also carry human readable (text) messages, to allow the server admin to provide additional debugging information (e.g: " pointed their NS at us. No idea why...", "We don't provide recursive DNS to Please stop asking...", "Have you tried Acme Anvil and DNS? We do DNS right..." (!). Please let us know if you think text is needed, or if a 16bit FCFS registry is expressive enough. ]

[ Open question: This document discusses extended *errors*, but it has been suggested that this could be used to also annotate *non-error* messages. The authors do not think that this is a good idea, but could be persuaded otherwise. ]

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

This Internet-Draft will expire on April 19, 2018.

Copyright Notice

Copyright (c) 2017 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents ( in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

1. Introduction and background

There are many reasons that a DNS query may fail, some of them transient, some permanent; some can be resolved by querying another server, some are likely best handled by stopping resolution. Unfortunately, the error signals that a DNS server can return are very limited, and are not very expressive. This means that applications and resolvers often have to "guess" at what the issue is - e.g the answer was marked REFUSED because of a lame delegation, or because of a lame delegation or because the nameserver is still starting up and loading zones? Is a SERVFAIL a DNSSEC validation issue, or is the nameserver experiencing a bad hair day?

A good example of issues that would benefit by additional error information is an error caused by a DNSSEC validation issue. When a stub resolver queries a DNSSEC bogus name (using a validating resolver), the stub resolver receives only a SERVFAIL in response. Unfortunately, SERVFAIL is used to signal many sorts of DNS errors, and so the stub resolver simply asks the next configured DNS resolver. The result of trying the next resolver is one of two outcomes: either the next resolver also validates, a SERVFAIL is returned again, and the user gets an (largely) incomprehensible error message; or the next resolver is not a validating resolver, and the user is returned a potentially harmful result.

This document specifies a mechanism to extend (or annotate) DNS errors to provide additional information about the cause of the error. This information can be used by the resolver to make a decision regarding whether or not to retry, or by technical users attempting to debug issues.

Here is a reference to an "external" (non-RFC / draft) thing: ([IANA.AS_Numbers]). And this is a link to an ID:[I-D.ietf-sidr-iana-objects].

1.1. Requirements notation

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

2. Extended Error EDNS0 option format

This draft uses an EDNS0 ([RFC2671]) option to include extended error (ExtError) information in DNS messages. The option is structured as follows:

                                             1   1   1   1   1   1
     0   1   2   3   4   5   6   7   8   9   0   1   2   3   4   5  
0: |                            OPTION-CODE                        |
2: |                           OPTION-LENGTH                       |
4: | R |                          RESERVED                         |
6: |                                CODE                           |

Currently the only defined flag is the R flag.

R - Retry
The R (or Retry) flag provides a hint to the receiver that it should retry the query, probably by querying another server. If the R bit is set (1), the sender believes that retrying the query may provide a successful answer next time; if the R bit is clear (0), the sender believes that it should not ask another server.

The remaining bits in the RESERVED field are reserved for future use and MUST be set to 0 by the sender and SHOULD be ignored by the receiver.

Code: A code point into the IANA "Extended DNS Errors" registry.

3. Use of the Extended DNS Error option

The Extended DNS Error (EDE) is an EDNS option. It can be included in any error response (SERVFAIL, NXDOMAIN, REFUSED, etc) to a query that includes an EDNS option. This document includes a set of initial codepoints (and requests to the IANA to add them to the registry), but is extensible via the IANA registry to allow additional error codes to be defined in the future.

The R (Retry) flag provides a hint (or suggestion) as to what the receiver may want to do with this annotated error. The mechanism is specifically designed to be extensible, and so implementations may receive EDE codes that it does not understand. The R flag allows implementations to make a decision as to what to do if it receives a response with an unknown code - retry or drop the query. Note that this flag is only a suggestion or hint. Receivers can choose to ignore this hint.

4. Defined Extended DNS Errors

This document defines some initial EDE codes. The mechanism is intended to be extensible, and additional codepoints will be registered in the "Extended DNS Errors" registry. This document provides suggestions for the R flag, but the originating server may ignore these recommendations if it knows better.

4.1. Extended DNS Error Code 100 - DNSSEC Bogus

The resolver attempted to perform DNSSEC validation, but validation ended in the Bogus state. The R flag should not be set.

4.2. Extended DNS Error Code 2 - DNSSEC Indeterminate

The resolver attempted to perform DNSSEC validation, but validation ended in the Indeterminate state.

Usually attached to SERVFAIL messages. The R flag should not be set.

4.3. Extended DNS Error Code 3 - Lame

An authoritative resolver that receives a query (with the RD bit clear) for a domain for which it is not authoritative SHOULD include this EDE code in the REFUSED response.

Implementations should set the R flag in this case (another nameserver might not be lame).

4.4. Extended DNS Error Code 4 - Prohibited

An authoritative or recursive resolver that receives a query from an "unauthorized" client can annotate its REFUSED message with this code. Examples of "unauthorized" clients are recursive queries from IP addresses outside the network, blacklisted IP addresses, etc.

Implementations SHOULD allow operators to define what to set the R flag to in this case.

4.5. Extended DNS Error Code 5 - TooBusy

[ Ed: This might be a bad idea. It is intended to allow servers under a DoS (for example a random subdomain attack) to signal to recursive clients that they are being abusive and should back off. This may be a bad idea -- it may "complete the attack", it may be spoofable (by anyone who could also do a MITM style attack), etc. ]

A nameserver which is under excessive load (for example, because it is experiencing a DoS) may annotate any answer with this code.

It is RECOMMENDED that implementations set the R flag in this case, but may allow operators to define what to set the R flag to.

[ agreed: bad idea -wjh ]

5. IANA Considerations

[This section under construction, beware. ]

This document defines a new EDNS(0) option, entitled "Extended DNS Error", assigned a value of TBD1 from the "DNS EDNS0 Option Codes (OPT)" registry [to be removed upon publication: []

Value  Name                 Status    Reference
-----  ----------------     ------    ------------------
 TBD   Extended DNS Error    TBD       [ This document ]

Data Tag Name Length Meaning ---- ---- ------ ------- TBD1 FooBar N FooBar server

The IANA is requested to create and maintain the "Extended DNS Error codes" registry. The codepoint space is broken into 3 ranges:

The codepoints 0, 65535 are reserved.

6. Open questions

Can this be included in *any* response or only responses to requests that included an EDNS option? Resolvers are supposed to ignore additional. EDNS capable ones are supposed to simply ignore unknown options. I know the spec says you can only include EDNS0 in a response if in a request -- it is time to reevaluate this?
Can this be applied to *any* response, or only error responses?
Should textual information be allowed as well? What if the only thing allowed is a domain name, e.g to point at where validation began failing?

7. Security Considerations

DNSSEC is being deployed - unfortunately a significant number of clients (~11% according to [GeoffValidation]), when receiving a SERVFAIL from a validating resolver because of a DNSSEC validaion issue simply ask the next (non-validating) resolver in their list, and don't get any of the protections which DNSSEC should provide. This is very similar to a kid asking his mother if he can have another cookie. When the mother says "No, it will ruin your dinner!", going off and asking his (more permissive) father and getting a "Yes, sure, cookie!".

8. Acknowledgements

The authors wish to thank Geoff Huston and Bob Harold, Carlos M. Martinez, Peter DeVries, George Michelson, Mark Andrews, Ondrej Sury, Edward Lewis, Paul Vixie, Shane Kerr. They also vaguely remember discussing this with a number of people over the years, but have forgotten who all they were -- if you were one of them, and are not listed, please let us know and we'll acknowledge you.

I also want to thank the band "Infected Mushroom" for providing a good background soundtrack (and to see if I can get away with this!) Another author would like to thank the band "Mushroom Infectors". This was funny at the time we wrote it, but I cannot remember why...

We would like to especially thank Peter van Dijk, who sent GitHub pull requests.

9. References

9.1. Normative References

[IANA.AS_Numbers] IANA, "Autonomous System (AS) Numbers"
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997.

9.2. Informative References

[GeoffValidation] IANA, "A quick review of DNSSEC Validation in today’s Internet", June 2016.
[I-D.ietf-sidr-iana-objects] Manderson, T., Vegoda, L. and S. Kent, "RPKI Objects issued by IANA", Internet-Draft draft-ietf-sidr-iana-objects-03, May 2011.

Appendix A. Changes / Author Notes.

[RFC Editor: Please remove this section before publication ]

From -02 to -03:

From -00 to -01;

Authors' Addresses

Warren Kumari Google 1600 Amphitheatre Parkway Mountain View, CA, 94043 US EMail:
Evan Hunt ISC 950 Charter St Redwood City, CA, 94063 US EMail:
Roy Arends Nominet UK EMail: TBD
Wes Hardaker USC/ISI P.O. Box 382 Davis, VA, 95617 US
David C Lawrence Akamai Technologies 150 Broadway Cambridge, MA, 02142-1054 US EMail: