idnits 2.17.1 draft-ietf-dnsop-no-response-issue-01.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (December 1, 2015) is 3062 days in the past. Is this intentional? Checking references for intended status: Best Current Practice ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Obsolete normative reference: RFC 5966 (Obsoleted by RFC 7766) Summary: 1 error (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Andrews 3 Internet-Draft ISC 4 Intended status: Best Current Practice December 1, 2015 5 Expires: June 3, 2016 7 A Common Operational Problem in DNS Servers - Failure To Respond. 8 draft-ietf-dnsop-no-response-issue-01 10 Abstract 12 The DNS is a query / response protocol. Failure to respond or to 13 respond correctly to queries causes both immediate operational 14 problems and long term problems with protocol development. 16 This document identifies a number of common classes of queries to 17 which some servers either fail to respond or else respond 18 incorrectly. This document also suggests procedures for TLD and 19 other similar zone operators to apply to help reduce / eliminate the 20 problem. 22 The document does not look at the DNS data itself, just the structure 23 of the responses. 25 Status of This Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on June 3, 2016. 42 Copyright Notice 44 Copyright (c) 2015 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 60 2. Common queries class that result in non responses. . . . . . 3 61 2.1. EDNS Queries - Version Independent . . . . . . . . . . . 4 62 2.2. EDNS Queries - Version Specific . . . . . . . . . . . . . 4 63 2.3. EDNS Options . . . . . . . . . . . . . . . . . . . . . . 4 64 2.4. EDNS Flags . . . . . . . . . . . . . . . . . . . . . . . 4 65 2.5. DNS Flags . . . . . . . . . . . . . . . . . . . . . . . . 5 66 2.6. Unknown / Unsupported Type Queries . . . . . . . . . . . 5 67 2.7. Unknown DNS opcodes . . . . . . . . . . . . . . . . . . . 5 68 2.8. TCP Queries . . . . . . . . . . . . . . . . . . . . . . . 5 69 3. Remediating . . . . . . . . . . . . . . . . . . . . . . . . . 6 70 4. Firewalls and Load Balancers . . . . . . . . . . . . . . . . 7 71 5. Scrubbing Services . . . . . . . . . . . . . . . . . . . . . 8 72 6. Whole Answer Caches . . . . . . . . . . . . . . . . . . . . . 9 73 7. Response Code Selection . . . . . . . . . . . . . . . . . . . 9 74 8. Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 75 8.1. Testing - Basic DNS . . . . . . . . . . . . . . . . . . . 10 76 8.2. Testing - Extended DNS . . . . . . . . . . . . . . . . . 12 77 9. Security Considerations . . . . . . . . . . . . . . . . . . . 15 78 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 15 79 11. Normative References . . . . . . . . . . . . . . . . . . . . 16 80 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 16 82 1. Introduction 84 The DNS [RFC1034], [RFC1035] is a query / response protocol. Failure 85 to respond to queries or to respond incorrectly causes both immediate 86 operational problems and long term problems with protocol 87 development. 89 Failure to respond to a query is indistinguishable from a packet loss 90 without doing a analysis of query response patterns and results in 91 unnecessary additional queries being made by DNS clients and 92 unnecessary delays being introduced to the resolution process. 94 Due to the inability to distinguish between packet loss and 95 nameservers dropping EDNS [RFC6891] queries, packet loss is sometimes 96 misclassified as lack of EDNS support which can lead to DNSSEC 97 validation failures. 99 Allowing servers which fail to respond to queries to remain results 100 in developers being afraid to deploy implementations of recent 101 standards. Such servers need to be identified and corrected / 102 replaced. 104 The DNS has response codes that cover almost any conceivable query 105 response. A nameserver should be able to respond to any conceivable 106 query using them. 108 Unless a nameserver is under attack, it should respond to all queries 109 directed to it as a result of following delegations. Additionally 110 code should not assume that there isn't a delegation to the server 111 even if it is not configured to serve the zone. Broken delegations 112 are a common occurrence in the DNS and receiving queries for zones 113 that the server is not configured for is not necessarily an 114 indication that the server is under attack. Parent zone operators 115 are supposed to regularly check that the delegating NS records are 116 consistent with those of the delegated zone and to correct them when 117 they are not [RFC1034]. If this was being done regularly, the 118 instances of broken delegations would be much lower. 120 When a nameserver is under attack it may wish to drop packets. A 121 common attack is to use a nameserver as a amplifier by sending 122 spoofed packets. This is done because response packets are bigger 123 than the queries and big amplification factors are available 124 especially if EDNS is supported. Limiting the rate of responses is 125 reasonable when this is occurring and the client should retry. This 126 however only works if legitimate clients are not being forced to 127 guess whether EDNS queries are accept or not. While there is still a 128 pool of servers that don't respond to EDNS requests, clients have no 129 way to know if the lack of response is due to packet loss, EDNS 130 packets not being supported or rate limiting due to the server being 131 under attack. Mis-classifications of server characteristics are 132 unavoidable when rate limiting is done. 134 2. Common queries class that result in non responses. 136 There are three common query classes that result in non responses 137 today. These are EDNS queries, queries for unknown (unallocated) or 138 unsupported types, and filtering of TCP queries. 140 2.1. EDNS Queries - Version Independent 142 Identifying servers that fail to respond to EDNS queries can be done 143 by first identifying that the server responds to regular DNS queries, 144 followed by a series of otherwise identical responses using EDNS, 145 then making the original query again. A series of EDNS queries is 146 needed as at least one DNS implementation responds to the first EDNS 147 query with FORMERR but fails to respond to subsequent queries from 148 the same address for a period until a regular DNS query is made. The 149 EDNS query should specify a UDP buffer size of 512 bytes to avoid 150 false classification of not supporting EDNS due to response packet 151 size. 153 If the server responds to the first and last queries but fails to 154 respond to most or all of the EDNS queries, it is probably faulty. 155 The test should be repeated a number of times to eliminate the 156 likelihood of a false positive due to packet loss. 158 Firewalls may also block larger EDNS responses but there is no easy 159 way to check authoritative servers to see if the firewall is 160 misconfigured. 162 2.2. EDNS Queries - Version Specific 164 Some servers respond correctly to EDNS version 0 queries but fail to 165 respond to EDNS queries with version numbers that are higher than 166 zero. Servers should respond with BADVERS to EDNS queries with 167 version numbers that they do not support. 169 Some servers respond correctly to EDNS version 0 queries but fail to 170 set QR=1 when responding to EDNS versions they do not support. Such 171 answers are discarded or treated as requests. 173 2.3. EDNS Options 175 Some servers fail to respond to EDNS queries with EDNS options set. 176 Unknown EDNS options are supposed to be ignored by the server 177 [RFC6891]. 179 2.4. EDNS Flags 181 Some servers fail to respond to EDNS queries with EDNS flags set. 182 Server should ignore EDNS flags there do not understand and should 183 not add them to the response [RFC6891]. 185 2.5. DNS Flags 187 Some servers fail to respond to DNS queries with various DNS flags 188 set, regardless of whether they are defined or still reserved. At 189 the time of writing there are servers that fail to respond to queries 190 with the AD bit set to 1 and servers that fail to respond to queries 191 with the last reserved flag bit set. 193 2.6. Unknown / Unsupported Type Queries 195 Identifying servers that fail to respond to unknown or unsupported 196 types can be done by making an initial DNS query for an A record, 197 making a number of queries for an unallocated type, then making a 198 query for an A record again. IANA maintains a registry of allocated 199 types. 201 If the server responds to the first and last queries but fails to 202 respond to the queries for the unallocated type, it is probably 203 faulty. The test should be repeated a number of times to eliminate 204 the likelihood of a false positive due to packet loss. 206 2.7. Unknown DNS opcodes 208 The use of previously undefined opcodes is to be expected. Since the 209 DNS was first defined two new opcodes have been added, UPDATE and 210 NOTIFY. 212 NOTIMP is the expected rcode to an unknown / unimplemented opcode. 214 Note: while new opcodes will most probably use the current layout 215 structure for the rest of the message there is no requirement than 216 anything other than the DNS header match. 218 2.8. TCP Queries 220 All DNS servers are supposed to respond to queries over TCP 221 [RFC5966]. Firewalls that drop TCP connection attempts rather that 222 resetting the connect attempt or send a ICMP/ICMPv6 administratively 223 prohibited message introduce excessive delays to the resolution 224 process. 226 Whether a server accepts TCP connections can be tested by first 227 checking that it responds to UDP queries to confirm that it is up and 228 operating, then attempting the same query over TCP. An additional 229 query should be made over UDP if the TCP connection attempt fails to 230 confirm that the server under test is still operating. 232 3. Remediating 234 While the first step in remediating this problem is to get the 235 offending nameserver code corrected, there is a very long tail 236 problem with DNS servers in that it can often take over a decade 237 between the code being corrected and a nameserver being upgraded with 238 corrected code. With that in mind it is requested that TLD, and 239 other similar zone operators, take steps to identify and inform their 240 customers, directly or indirectly through registrars, that they are 241 running such servers and that the customers need to correct the 242 problem. 244 TLD operators are being asked to do this as they, due to the nature 245 of running a TLD and the hierarchical nature of the DNS, have access 246 to a large numbers of nameserver names as well as contact details for 247 the registrants of those nameservers. While it is possible construct 248 lists of nameservers from other sources and that has been done to 249 survey the state of the Internet, that doesn't give the tester the 250 contact details necessary to inform the operators. The SOA RNAME is 251 often invalid and whois data is obscured and / or not available which 252 makes it infeasible for others to do this. 254 While this section talks about TLD operators performing this work, it 255 may be done by registrars on behalf of the TLD operator. The intent 256 is to ensure that the testing happens and that operators of 257 noncompliant nameservers be informed, rather than to prescribe who 258 does the actual testing and communication. Note: having registrars 259 perform this testing and reporting is likely to result in duplicate 260 reports for the same server being issued by multiple registrars. 262 TLD operators should construct a list of servers child zones are 263 delegated to along with a delegated zone name. This name shall be 264 the query name used to test the server as it is supposed to exist. 266 For each server the TLD operator shall make an SOA query of the 267 delegated zone name. This should result in the SOA record being 268 returned in the answer section. If the SOA record is not returned 269 but some other response is returned, this is a indication of a bad 270 delegation and the TLD operator should take whatever steps it 271 normally takes to rectify a bad delegation. If more that one zone is 272 delegated to the server, it should choose another zone until it finds 273 a zone which responds correctly or it exhausts the list of zones 274 delegated to the server. 276 If the server fails to get a response to a SOA query, the TLD 277 operator should make an A query as some nameservers fail to respond 278 to SOA queries but respond to A queries. If it gets no response to 279 the A query, another delegated zone should be queried for as some 280 nameservers fail to respond to zones they are not configured for. If 281 subsequent queries find a responding zone, all delegation to this 282 server need to be checked and rectified using the TLD's normal 283 procedures. 285 Having identified a working tuple the TLD 286 operator should now check that the server responds to EDNS, Unknown 287 Query Type and TCP tests as described above. If the TLD operator 288 finds that server fails any of the tests, the TLD operator shall take 289 steps to inform the operator of the server that they are running a 290 faulty nameserver and that they need to take steps to correct the 291 matter. The TLD operator shall also record the 292 for follow-up testing. 294 If repeated attempts to inform and get the customer to correct / 295 replace the faulty server are unsuccessful the TLD operator shall 296 remove all delegations to said server from the zone. 298 It will also be necessary for TLD operators to repeat the scans 299 periodically. It is recommended that this be performed monthly 300 backing off to bi-annually once the numbers of faulty servers found 301 drops off to less than 1 in 100000 servers tested. Follow-up tests 302 for faulty servers still need to be performed monthly. 304 Some operators claim that they can't perform checks at registration 305 time. If a check is not performed at registration time, it needs to 306 be performed within a week of registration in order to detect faulty 307 servers swiftly. 309 Checking of delegations by TLD operators should be nothing new as 310 they have been required from the very beginnings of DNS to do this 311 [RFC1034]. Checking for compliance of nameserver operations should 312 just be a extension of such testing. 314 It is recommended that TLD operators setup a test web page which 315 performs the tests the TLD operator performs as part of their regular 316 audits to allow nameserver operators to test that they have correctly 317 fixed their servers. Such tests should be rate limited to avoid 318 these pages being a denial of service vector. 320 4. Firewalls and Load Balancers 322 Firewalls and load balancers can affect the externally visible 323 behaviour of a nameserver. Tests for conformance need to be done 324 from outside of any firewall so that the system as a whole is tested. 326 Firewalls and load balancers should not drop DNS packets that they 327 don't understand. They should either pass through the packets or 328 generate an appropriate error response. 330 Requests for unknown query types is normal client behaviour and 331 should not be construed as an attack. Nameservers have always been 332 expected to be able to handle such queries. 334 Requests with unassigned flags set (DNS or EDNS) is expected client 335 behaviour and should not be construed as an attack. The behaviour 336 for unassigned is to ignore them in the request and to not set them 337 in the response. All dropping DNS / EDNS packets with unassigned 338 flags does is make it harder to deploy extensions that make use of 339 them due to the need to reconfigure / update firewalls. 341 Requests with unknown EDNS options is expected client behaviour 342 should not be construed as an attack. The correct behaviour for 343 unknown EDNS options is to ignore there presence when constructing a 344 reply. 346 Requests with unknown EDNS versions is expected client behaviour 347 should not be trued as an attack The correct behaviour for unknown 348 EDNS versions is to return BADVERS along with the highest EDNS 349 version the server supports. All dropping EDNS packets does is break 350 EDNS version negotiation. 352 Firewalls should not assume that there will only be a single response 353 message to a requests. There have been proposals to use EDNS to 354 signal that multiple DNS messages be returned rather than a single 355 UDP message that is fragmented at the IP layer. 357 5. Scrubbing Services 359 Scrubbing services, like firewalls, can affect the externally visible 360 behaviour of a nameserver. If a operator uses a scrubbing service, 361 they should check that legitimate queries are not being blocked. 363 Scrubbing services, unlike firewalls, are also turned on and off in 364 response to denial of service attacks. One needs to take care when 365 choosing a scrubbing service and ask questions like: 367 Do they pass unknown DNS query types? 369 Do they pass unknown EDNS versions? 371 Do they pass unknown EDNS options? 373 Do they pass unknown EDNS flags? 374 Do they pass requests with unknown DNS opcodes? 376 Do they pass requests with the remaining reserved DNS header flag 377 bit set? 379 All of these are not attack vectors but some scrubbing services treat 380 them as such. 382 6. Whole Answer Caches 384 Whole answer caches take a previously constructed answer and return 385 it to a subsequent query for the same name, type and class, just 386 updating the query id field and possibly the qname to match the 387 incoming query to avoid constructing each response individually. 389 Whole answer caches can return the wrong response to a query if they 390 do not take all of the attributes of the query into account, rather 391 than just some of them e.g. qname, qtype and qclass. This has 392 implications when testing and with overall protocol compliance. 394 e.g. There are whole answer caches that ignore the EDNS version 395 field which results in incorrect answers to non EDNS version 0 396 queries being returned if they were preceded by a EDNS version 0 397 query for the same name and type. 399 e.g. There are caches that ignore the EDNS options in the query 400 resulting in options only working some of the time and/or options 401 being returned when not requested. 403 7. Response Code Selection 405 Choosing the correct response code when fixing a nameserver is 406 important. Just because a type is not implemented does not mean that 407 NOTIMP is the correct response code to return. Response codes need 408 to be chosen considering how clients will handle them. 410 For unimplemented opcodes NOTIMP is the expected response code. 411 Additionally a new opcode could change the message format by 412 extending the header or changing the structure of the records etc. 413 This may result in FORMERR being returned though NOTIMP would be more 414 correct. 416 In general, for unimplemented type codes Name Error (NXDOMAIN) and 417 NOERROR (no data) are the expected response codes. A server is not 418 supposed to serve a zone which contains unsupported types ([RFC1034]) 419 so the only thing left is return if the QNAME exists or not. NOTIMP 420 and REFUSED are not useful responses as they force the clients to try 421 all the authoritative servers for a zone looking for a server which 422 will answer the query. 424 Meta queries type may be the exception but these need to be thought 425 about on a case by case basis. 427 If the server supports EDNS and get a query with an unsupported EDNS 428 version, the correct response is BADVERS [RFC6891]. 430 If the server do not support EDNS at all, FORMERR and NOTIMP are the 431 expected error codes. That said a minimal EDNS server implementation 432 just requires parsing the OPT records and responding with an empty 433 OPT record. There is no need to interpret any EDNS options present 434 in the request as unsupported options are expected to be ignored 435 [RFC6891]. 437 8. Testing 439 Testing is divided into two sections. Basic DNS which all servers 440 should meet and Extended DNS which should be met by all servers that 441 support EDNS. If a server does not support EDNS it should still 442 respond to all the tests. 444 It is advisable to run all of the tests below in parallel so as to 445 minimise the delays due to multiple timeouts when the servers do not 446 respond. 448 The tests below use dig from BIND 9.11.0 which is still in 449 development. 451 8.1. Testing - Basic DNS 453 This first set of tests cover basic DNS server behaviour and all 454 servers should pass these tests. 456 Verify the server is configured for the zone: 458 dig +noedns +noad +norec soa $zone @$server 460 expect: status: NOERROR 461 expect: SOA record 462 expect: flag: aa to be present 464 Check that TCP queries work: 466 dig +noedns +noad +norec +tcp soa $zone @$server 468 expect: status: NOERROR 469 expect: SOA record 470 expect: flag: aa to be present 472 The requirement that TCP be supported is defined in [RFC5966]. 474 Check that queries for an unknown type work: 476 dig +noedns +noad +norec type1000 $zone @$server 478 expect: status: NOERROR 479 expect: an empty answer section. 480 expect: flag: aa to be present 482 That new types are to be expected is specified in Section 3.6, 483 [RFC1035]. Servers that don't support a new type are expected to 484 reject a zone that contains a unsupported type as per Section 5.2, 485 [RFC1035]. This means that a server that does load a zone can answer 486 questions for unknown types with NOERROR or NXDOMAIN as per 487 Section 4.3.2, [RFC1034]. [RFC6895] later reserved distinct ranges 488 for meta and data types which allows servers to be definitive about 489 whether a query should be answerable from zone content or not. 491 Check that queries with CD=1 work: 493 dig +noedns +noad +norec +cd soa $zone @$server 495 expect: status: NOERROR 496 expect: SOA record to be present 497 expect: flag: aa to be present 499 CD use in queries is defined in [RFC4035]. 501 Check that queries with AD=1 work: 503 dig +noedns +norec +ad soa $zone @$server 505 expect: status: NOERROR 506 expect: SOA record to be present 507 expect: flag: aa to be present 509 AD use in queries is defined in [RFC6840]. 511 Check that queries with the last unassigned DNS header flag work and 512 that the flag bit is not copied to the response: 514 dig +noedns +noad +norec +zflag soa $zone @$server 516 expect: status: NOERROR 517 expect: SOA record to be present 518 expect: MBZ to not be in the response 519 expect: flag: aa to be present 521 MBZ (Must Be Zero) presence indicates the flag bit has been 522 incorrectly copied. See Section 4.1.1, [RFC1035] "Z Reserved for 523 future use. Must be zero in all queries and responses." 525 Check that new opcodes are handled: 527 dig +noedns +noad +opcode=15 +norec +header-only @$server 529 expect: status: NOTIMP 530 expect: SOA record to not be present 531 expect: flag: aa to NOT be present 533 As unknown opcodes have no definition, including packet format other 534 than there must be a DNS header present, there is only one possible 535 rcode that make sense to return to a request with a unknown opcode 536 and that is NOTIMP. 538 8.2. Testing - Extended DNS 540 The next set of test cover various aspects of EDNS behaviour. If any 541 of these tests succeed, then all of them should succeed. There are 542 servers that support EDNS but fail to handle plain EDNS queries 543 correctly so a plain EDNS query is not a good indicator of lack of 544 EDNS support. 546 Check that plain EDNS queries work: 548 dig +nocookie +edns=0 +noad +norec soa $zone @$server 550 expect: status: NOERROR 551 expect: SOA record to be present 552 expect: OPT record to be present 553 expect: EDNS Version 0 in response 554 expect: flag: aa to be present 556 +nocookie disables sending a EDNS COOKIE option in which is on by 557 default. 559 Check that EDNS version 1 queries work (EDNS supported): 561 dig +nocookie +edns=1 +noednsneg +noad +norec soa $zone @$server 563 expect: status: BADVERS 564 expect: SOA record to not be present 565 expect: OPT record to be present 566 expect: EDNS Version 0 in response 567 expect: flag: aa to NOT be present 569 Only EDNS Version 0 is currently defined so the response should 570 always be a 0 version. This will change when EDNS version 1 is 571 defined. BADVERS is the expected rcode if EDNS is supported as per 572 Section 6.1.3, [RFC6891]. 574 Check that EDNS queries with an unknown option work (EDNS supported): 576 dig +nocookie +edns=0 +noad +norec +ednsopt=100 soa $zone @$server 578 expect: status: NOERROR 579 expect: SOA record to be present 580 expect: OPT record to be present 581 expect: OPT=100 to not be present 582 expect: EDNS Version 0 in response 583 expect: flag: aa to be present 585 Unknown EDNS options are supposed to be ignored, Section 6.1.2, 586 [RFC6891]. 588 Check that EDNS queries with unknown flags work (EDNS supported): 590 dig +nocookie +edns=0 +noad +norec +ednsflags=0x40 soa $zone @$server 592 expect: status: NOERROR 593 expect: SOA record to be present 594 expect: OPT record to be present 595 expect: MBZ not to be present 596 expect: EDNS Version 0 in response 597 expect: flag: aa to be present 599 MBZ (Must Be Zero) presence indicates the flag bit has been 600 incorrectly copied as per Section 6.1.4, [RFC6891]. 602 Check that EDNS version 1 queries with unknown flags work (EDNS 603 supported): 605 dig +nocookie +edns=1 +noednsneg +noad +norec +ednsflags=0x40 soa \ 606 $zone @$server 608 expect: status: BADVERS 609 expect: SOA record to NOT be present 610 expect: OPT record to be present 611 expect: MBZ not to be present 612 expect: EDNS Version 0 in response 613 expect: flag: aa to NOT be present 615 +noednsneg disables EDNS version negotiation in DiG; MBZ (Must Be 616 Zero) presence indicates the flag bit has been incorrectly copied. 618 Check that EDNS version 1 queries with unknown options work (EDNS 619 supported): 621 dig +nocookie +edns=1 +noednsneg +noad +norec +ednsopt=100 soa \ 622 $zone @$server 624 expect: status: BADVERS 625 expect: SOA record to NOT be present 626 expect: OPT record to be present 627 expect: OPT=100 to NOT be present 628 expect: EDNS Version 0 in response 629 expect: flag: aa to be present 631 +noednsneg disables EDNS version negotiation in DiG. 633 Check that a DNSSEC queries work (EDNS supported): 635 dig +nocookie +edns=0 +noad +norec +dnssec soa $zone @$server 637 expect: status: NOERROR 638 expect: SOA record to be present 639 expect: OPT record to be present 640 expect: DO=1 to be present if a RRSIG is in the response 641 expect: EDNS Version 0 in response 642 expect: flag: aa to be present 644 DO=1 should be present if RRSIGs are returned as they indicate that 645 the server supports DNSSEC. Servers that support DNSSEC are supposed 646 to copy the DO bit from the request to the response as per [RFC3225]. 648 Check that EDNS version 1 DNSSEC queries work (EDNS supported): 650 dig +nocookie +edns=1 +noednsneg +noad +norec +dnssec soa \ 651 $zone @$server 653 expect: status: BADVERS 654 expect: SOA record to not be present 655 expect: OPT record to be present 656 expect: DO=1 to be present if the EDNS version 0 DNSSEC query test 657 returned DO=1 658 expect: EDNS Version 0 in response 659 expect: flag: aa to NOT be present 661 +noednsneg disables EDNS version negotiation in DiG. 663 Check that EDNS queries with multiple defined EDNS options work. 665 dig +edns=0 +noad +norec +cookie +nsid +expire +subnet=0.0.0.0/0 \ 666 soa $zone @$server 668 expect: status: NOERROR 669 expect: SOA record to be present 670 expect: OPT record to be present 671 expect: EDNS Version 0 in response 672 expect: flag: aa to be present 674 If EDNS is not supported by the nameserver, we expect a response to 675 all the above queries. That response may be a FORMERR or NOTIMP 676 error response or the OPT record may just be ignored. 678 9. Security Considerations 680 Testing protocol compliance can potentially result in false reports 681 of attempts to break services from Intrusion Detection Services and 682 firewalls. None of the tests listed above should break nominally 683 EDNS compliant servers. None of the tests above should break non 684 EDNS servers. All the tests above are well formed, though not 685 necessarily common, DNS queries. 687 Relaxing firewall settings to ensure EDNS compliance could 688 potentially expose a critical implementation flaw in the nameserver. 689 Nameservers should be tested for conformance before relaxing firewall 690 settings. 692 10. IANA Considerations 694 IANA / ICANN needs to consider what tests, if any, from above that it 695 should add to the zone maintenance procedures for zones under its 696 control including pre-delegation checks. Otherwise this document has 697 no actions for IANA. 699 11. Normative References 701 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 702 STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987, 703 . 705 [RFC1035] Mockapetris, P., "Domain names - implementation and 706 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 707 November 1987, . 709 [RFC3225] Conrad, D., "Indicating Resolver Support of DNSSEC", 710 RFC 3225, DOI 10.17487/RFC3225, December 2001, 711 . 713 [RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S. 714 Rose, "Protocol Modifications for the DNS Security 715 Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005, 716 . 718 [RFC5966] Bellis, R., "DNS Transport over TCP - Implementation 719 Requirements", RFC 5966, DOI 10.17487/RFC5966, August 720 2010, . 722 [RFC6840] Weiler, S., Ed. and D. Blacka, Ed., "Clarifications and 723 Implementation Notes for DNS Security (DNSSEC)", RFC 6840, 724 DOI 10.17487/RFC6840, February 2013, 725 . 727 [RFC6891] Damas, J., Graff, M., and P. Vixie, "Extension Mechanisms 728 for DNS (EDNS(0))", STD 75, RFC 6891, 729 DOI 10.17487/RFC6891, April 2013, 730 . 732 [RFC6895] Eastlake 3rd, D., "Domain Name System (DNS) IANA 733 Considerations", BCP 42, RFC 6895, DOI 10.17487/RFC6895, 734 April 2013, . 736 Author's Address 738 M. Andrews 739 Internet Systems Consortium 740 950 Charter Street 741 Redwood City, CA 94063 742 US 744 Email: marka@isc.org