PowerDNS manual PowerDNS BV It is a book about a Spanish guy called Manual. You shou ld read it. -- Dilbert __________________________________________________________ Table of Contents 1. The PowerDNS dynamic nameserver 1.1. Function & design of PDNS 1.2. About this document 1.3. Release notes 1.3.1. Authoritative Server version 2.9.22 1.3.2. Authoritative Server version 2.9.21.2 1.3.3. Authoritative Server version 2.9.21.1 1.3.4. Recursor version 3.1.7 1.3.5. Recursor version 3.1.6 1.3.6. Recursor version 3.1.5 1.3.7. PowerDNS Authoritative Server version 2.9.21 1.3.8. Recursor version 3.1.4 1.3.9. Recursor version 3.1.3 1.3.10. Recursor version 3.1.2 1.3.11. Recursor version 3.1.1 1.3.12. Recursor version 3.0.1 1.3.13. Recursor version 3.0 1.3.14. Version 2.9.20 1.3.15. Version 2.9.19 1.3.16. Version 2.9.18 1.3.17. Version 2.9.17 1.3.18. Version 2.9.16 1.3.19. Version 2.9.15 1.3.20. Version 2.9.14 1.3.21. Version 2.9.13 1.3.22. Version 2.9.12 1.3.23. Version 2.9.11 1.3.24. Version 2.9.10 1.3.25. Version 2.9.8 1.3.26. Version 2.9.7 1.3.27. Version 2.9.6 1.3.28. Version 2.9.5 1.3.29. Version 2.9.4 1.3.30. Version 2.9.3a 1.3.31. Version 2.9.2 1.3.32. Version 2.9.1 1.3.33. Version 2.9 1.3.34. Version 2.8 1.3.35. Version 2.7 and 2.7.1 1.3.36. Version 2.6.1 1.3.37. Version 2.6 1.3.38. Version 2.5.1 1.3.39. Version 2.5 1.3.40. Version 2.4 1.3.41. Version 2.3 1.3.42. Version 2.2 1.3.43. Version 2.1 1.3.44. Version 2.0.1 1.3.45. Version 2.0 1.3.46. Version 2.0 Release Candidate 2 1.3.47. Version 2.0 Release Candidate 1 1.3.48. Version 1.99.12 Prerelease 1.3.49. Version 1.99.11 Prerelease 1.3.50. Version 1.99.10 Prerelease 1.3.51. Version 1.99.9 Early Access Prerelease 1.3.52. Version 1.99.8 Early Access Prerelease 1.3.53. Version 1.99.7 Early Access Prerelease 1.3.54. Version 1.99.6 Early Access Prerelease 1.3.55. Version 1.99.5 Early Access Prerelease 1.3.56. Version 1.99.4 Early Access Prerelease 1.3.57. Version 1.99.3 Early Access Prerelease 1.3.58. Version 1.99.2 Early Access Prerelease 1.3.59. Version 1.99.1 Early Access Prerelease 1.4. Security 1.5. PowerDNS Security Advisory 2006-01: Malformed TCP queries can lead to a buffer overflow which might be exploitable 1.6. PowerDNS Security Advisory 2006-02: Zero second CNAME TTLs can make PowerDNS exhaust allocated stack space, and crash 1.7. PowerDNS Security Advisory 2008-01: System random generator can be predicted, leading to the potential to 'spoof' PowerDNS Recursor 1.8. PowerDNS Security Advisory 2008-02: By not responding to certain queries, domains become easier to spoof 1.9. PowerDNS Security Advisory 2008-02: Some PowerDNS Configurations can be forced to restart remotely 1.10. Acknowledgements 2. Installing on Unix 2.1. Possible problems at this point 2.2. Testing your install 2.2.1. Typical errors 2.3. Running PDNS on unix 3. Installing on Microsoft Windows 3.1. Configuring PDNS on Microsoft Windows 3.2. Running PDNS on Microsoft Windows 4. Basic setup: configuring database connectivity 4.1. Example: configuring MySQL 4.1.1. Common problems 5. Dynamic resolution using the PipeBackend 5.1. Deploying the PipeBackend with the BindBackend 6. Logging & Monitoring Authoritative Server performance 6.1. Webserver 6.2. Via init.d commands 6.3. Operational logging using syslog 7. Security settings & considerations 7.1. Settings 7.1.1. Running as a less privileged identity 7.1.2. Jailing the process in a chroot 7.2. Considerations 8. Virtual hosting 9. Performance 9.1. General advice 9.2. Native Posix Thread Library vs LinuxThreads 9.3. Performance related settings 9.3.1. Packet Cache 9.3.2. Query Cache 10. Migrating to PDNS 10.1. Zone2sql 11. Recursion 11.1. Details 12. PowerDNS resolver/recursing nameserver 12.1. pdns_recursor settings 12.2. Controlling and querying the recursor 12.3. PowerDNS Recursor performance 12.4. Details 12.4.1. Anti-spoofing 12.4.2. Throttling 12.5. Statistics 12.6. Scripting 12.6.1. Configuring Lua scripts 12.6.2. Writing Lua PowerDNS Recursor scripts 12.7. Design and Engineering of the PowerDNS Recursor 12.7.1. The PowerDNS Recursor 12.7.2. Synchronous code using MTasker 12.7.3. MPlexer 12.7.4. MOADNSParser 12.7.5. The C++ Standard Library / Boost 12.7.6. Actual DNS Algorithm 12.7.7. The non-cached case 12.7.8. Some of the things we glossed over 12.7.9. The Recursor Cache 12.7.10. Some small things 13. Master/Slave operation & replication 13.1. Native replication 13.2. Slave operation 13.2.1. Supermaster automatic provisioning of slaves 13.3. Master operation 14. Fancy records for seamless email and URL integration 15. Index of all Authoritative Server settings 16. Index of all Authoritative Server metrics 16.1. Counters & variables 16.1.1. Counters 16.1.2. Ring buffers 17. Supported record types and their storage 18. HOWTO & Frequently Asked Questions 18.1. Getting support, free and paid FAQ 18.2. Using and Compiling PowerDNS FAQ 18.3. Backend developer HOWTO 18.4. About PowerDNS.COM BV, 'the company' 19. Other tools included with PowerDNS 19.1. Notification proxy (nproxy) 20. Tools to analyse DNS traffic A. Backends in detail A.1. PipeBackend A.1.1. PipeBackend protocol A.2. MySQL backend A.2.1. Configuration settings A.2.2. Notes A.3. Random Backend A.4. MySQL PDNS backend A.4.1. Notes A.5. Generic MySQL and PgSQL backends A.5.1. MySQL specifics A.5.2. PostgresSQL specifics A.5.3. Oracle specifics A.5.4. Basic functionality A.5.5. Master/slave queries A.5.6. Fancy records A.5.7. Settings and specifying queries A.5.8. Native operation A.5.9. Slave operation A.5.10. Superslave operation A.5.11. Master operation A.6. Oracle backend A.6.1. Setting up Oracle for use with PowerDNS A.7. Generic SQLite backend (2 and 3) A.7.1. Compiling the SQLite backend A.7.2. Setting up the database A.7.3. Using the SQLite backend A.8. DB2 backend A.9. Bind zone file backend A.9.1. Operation A.9.2. Pdns_control commands A.9.3. Performance A.9.4. Master/slave configuration A.9.5. Commands A.10. ODBC backend A.11. XDB Backend A.12. LDAP backend A.13. OpenDBX backend A.14. Geo backend B. PDNS internals B.1. Controlsocket B.1.1. pdns_control B.2. Guardian B.3. Modules & Backends B.4. How PDNS translates DNS queries into backend queries C. Backend writers' guide C.1. Simple read-only native backends C.1.1. A sample minimal backend C.1.2. Interface definition C.2. Reporting errors C.3. Declaring and reading configuration details C.4. Read/write slave-capable backends C.4.1. Supermaster/Superslave capability C.5. Read/write master-capable backends D. Compiling PowerDNS D.1. Compiling PowerDNS on Unix D.1.1. AIX D.1.2. FreeBSD D.1.3. Linux D.1.4. MacOS X D.1.5. OpenBSD D.1.6. Solaris D.2. Compiling PowerDNS on Windows D.2.1. Assumptions D.2.2. Prequisites D.2.3. Nullsoft Installer D.2.4. Setting up the build-environment D.2.5. Compilation D.2.6. Miscellaneous E. PowerDNS license (GNU General Public License version 2) F. Further copyright statements F.1. AES implementation by Brian Gladman List of Tables 1-1. PowerDNS Security Advisory 1-2. PowerDNS Security Advisory 1-3. PowerDNS Security Advisory 1-4. PowerDNS Security Advisory 1-5. PowerDNS Security Advisory 17-1. SOA fields A-1. PipeBackend capabilities A-2. MySQL backend capabilities A-3. Random Backend capabilities A-4. MySQL backend capabilities A-5. Generic PgSQL and MySQL backend capabilities A-6. Oracle backend capabilities A-7. Generic SQLite backend capabilities A-8. DB2 backend capabilities A-9. Bind zone file backend capabilities A-10. ODBC backend capabilities A-11. LDAP backend capabilities A-12. OpenDBX backend capabilities A-13. Geo backend capabilities C-1. DNSResourceRecord class C-2. SOAData struct C-3. DomainInfo struct __________________________________________________________ Chapter 1. The PowerDNS dynamic nameserver The PowerDNS daemon is a versatile nameserver which supports a large number of backends. These backends can either be plain zonefiles or be more dynamic in nature. Additionally, through use of clever programming techniques, PowerDNS offers very high domain resolution performance. Prime examples of backends include relational databases, but also (geographical) loadbalancing and failover algorithms. The company is called PowerDNS.COM BV, the nameserver daemon is called PDNS. __________________________________________________________ 1.1. Function & design of PDNS PowerDNS consists of two parts: the Authoritative Server and the Recursor. Other nameservers fully combine these functions, PowerDNS offers them separately, but can mix both authoritative and recursive usage seamlessly. The Authoritative Server will answer questions about domains it knows about, but will not go out on the net to resolve queries about other domains. However, it can use a recursing backend to provide that functionality. Depending on your needs, this backend can either be the PowerDNS recursor or an external one. When the Authoritative Server answers a question, it comes out of the database, and can be trusted as being authoritative. There is no way to pollute the cache or to confuse the daemon. The Recursor, conversely, by default has no knowledge of domains itself, but will always consult other authoritative servers to answer questions given to it. PDNS has been designed to serve both the needs of small installations by being easy to setup, as well as for serving very large query volumes on large numbers of domains. Another prime goal is security. By the use of language features, the PDNS source code is very small (in the order of 10.000 lines) which makes auditing easy. In the same way, library features have been used to mitigate the risks of buffer overflows. Finally, PDNS is able to give a lot of statistics on its operation which is both helpful in determining the scalability of an installation as well as for spotting problems. __________________________________________________________ 1.2. About this document If you are reading this document from disk, you may want to check http://doc.powerdns.com for updates. The PDF version is available on http://doc.powerdns.com/pdf, a text file is on http://doc.powerdns.com/txt/. __________________________________________________________ 1.3. Release notes Before proceeding, it is advised to check the release notes for your PDNS version, as specified in the name of the distribution file. Beyond PowerDNS 2.9.20, the Authoritative Server and Recursor are released separately. __________________________________________________________ 1.3.1. Authoritative Server version 2.9.22 Warning Released on the 27th of January 2009. This is a huge release, spanning almost 20 months of development. Besides fixing a lot of bugs, of note is the addition of the so called 'Notification Proxy', which allows PowerDNS to function as a master server behind a firewall, plus the huge performance improvement of the internal caches. This work has been made possible by UPC Broadband and Directi, respectively. Finally, the release candidates of this version have been tested & improved by Jorn Ekkelenkamp, Ton van Rosmalen, Jeff Sipek, Tyler Hall, Christof Meerwald and Stefan Schmidt. Fixed between rc1 and rc2, but not an issue in 2.9.21. * pdns_control ccounts again outputs proper cache statistics. Implemented in commit 1304. * Negative query caching was reinstated, leading to 6 times fewer backend queries than rc1 on the Express.powerdns.com servers. * Packetcache no longer needlessly parses outgoing packets before sending them. * Fancy records work again. This work has been sponsored by ISP Services. Implemented in commit 1302 and commit 1299. New features: * pdns_control can now also work over TCP/IP. Sponsored by Directi. Commits 1246, 1251, 1254, 1255. * Implemented a notification proxy, see Section 19.1. This work was sponsored by UPC Broadband. Implemented in commits 1075, 1077, 1082, 1083, 1085 and 1086. * IXFR queries are now supported in the sense that we treat them as AXFR queries, silencing warnings in other nameservers. Suggested in ticket 131. * The PIPE backend has been extended by David Apgar to allow the reporting of errors using the 'FAIL' command, plus support for responses with whitespace. Implemented in commit 1114. * PowerDNS Authoritative server now parses incoming EDNS options, like maximum allowed packet size. Implemented in commit 1123 and commit 1281. * Added support for DHCID, IPSECKEY and KX records, thanks Norbert Sendetzky for the hint. Implemented in commit 1144. * Norbert Sendetzky has has added support for all record types supported by PowerDNS to the LDAPBackend. Furthermore, the detection of OpenLDAP in autoconf has been improved. Finally, debian has supplied some fixes to PowerLDAP. Implemented in commit 1152 and commit 1153. * Implemented EDNS NSID option for retrieving the nameserver ID out of band. Defaults to hostname, can be specified using the server-id setting. Code in commit 1232. * Implemented experimental EDNS PING for enhanced forgery resilience. Code in commit 1232. Performance: * Improve packet generation performance, in some cases by 25%. Code in 1258, 1259. * Improved access list checking performance. commit 1261. * PowerDNS Authoritative caches were completely redone, and are now based on the same cache that is in the resolver. This work has been sponsored by Directi. In large benchmarks, PowerDNS performance has improved by an order of magnitude or more. This new version allows for near-instantaneous cache purging, plus very rapid purging based on suffix. Purge commands can also be batched. This work is partially based on an innovative reverse-string comparison function authored by Aki Tuomi. * Installations which run with very high cache hitrates can now benefit from multiple CPUs by setting receiver-threads to the number of desired CPUs to utilize in cache operations. Implemented in commit 1316. * BIND backend speedups in commit 1108, measured at around a 20% improvement, possibly more on very large setups. Bugs fixed: * Tyler Hall discovered the PowerDNS configuration file parser had problems with trailing tabs. This turned out to be a wider problem in PowerDNS. Buggy code replaced by a library call in commit 1237 and commit 1240. * David Apgar of Yahoo discovered that our 'guardian' method of restarting PowerDNS in case of problems was not fool proof, and submitted a fix. A variation of this fix can be found in commit 1323. Also reported by Directi. * Connection reset by peer events in the TCP nameserver no longer lead to the cycling of database connections. Code in commit 1241. * FreeBSD compilation with Generic PostgreSQL backend was fixed. Reported by Wouter de Jong of WideXS, fixed in commit 1305, closes ticket 95. * Webserver no longer prints '1e2%'. Finally closes ticket 26. Much friendly nagging for over 3 years by Jeff Sipek, code in commit 1303. * PowerDNS used to ignore certain queries it could not answer. These queries are no longer ignored, but get a SERVFAIL response. Implemented in commit 1239. * Fix subtle CNAME and wildcard interactions reported by 'zzyzz', implemented in commit 1147. * The generic backends did not honour the default-ttl setting. Spotted and implemented by Matti Hiljanen. * Matti Hiljanen discovered that the OpenDBX backend did not fill out the SOA ttl value properly. Matti also improved the SQL statements for better compatibility. Implemented in commit 1181. * Treat invalid WWW requests better. Spotted by Maikel Verheijen, implemented in commit 1092. * Documentation errors and typos, spotted by Marco Davids (commit 1097) and Rejo Zengers (commit 1119) * Properly fill out the 'recursion available'-flag. Spotted by Augie Schwer in ticket 167. * Several memory leaks on bad data in the database or other errors have been fixed. Addressed in 1078 and 1079. * In contravention to the documentation, the domain type as specified in the database ('MASTER', 'SLAVE' or 'NATIVE') was interpreted case sensitively. 1084. * BIND backend could crash on processing information about slave zones to be checked. Spotted by Stefan Schmidt, fixed in 1089. * Jelte Jansen of Stichting NLNetLabs discovered PowerDNS in BIND mode couldn't operate as a root-server! Fixed in 1057. * 'DPS' discovered there was a rare opportunity for PowerDNS to lock up waiting for new data. Addressed in 1076. * Make singlethreaded mode more resilient against errors. commit 1272. * DNSSEC records were part of 2.9.21, but were not actually hooked up. Please note that while PowerDNS can serve most DNSSEC records, it does not do DNSSEC processing. Implemented in 1046. * Shawn Starr migrated all his domains to PowerDNS in one evening, from an installation that had been used since BIND4. In doing so, he found 3 bugs in as many hours. An IN statement in the BIND named.conf with a zone with a trailing dot was misparsed, fixed in commit 1233. Secondly, the zonefile parser tripped over a line consisting of nothing but comments in the wrong place. Finally '$ORIGIN .' was misparsed. Last two issues fixed in commit 1234. * Our statistics counters did not wrap correctly after the 2.15 billion mark. Spotted by Stefan Schmidt, reported in ticket 179, fixed in commit 1284. * Bindbackend could sometimes generate very strange error messages while processing a malformed zone file. Sometimes such error messages could cause a crash (reported on HP-UX). Addressed by commit 1279. This could not be triggered remotely. Closes ticket ticket 203. * Pipe backend did not clean up killed coprocesses. Found and fixed by Daniel Drown * Installations with tens of thousands of slave domains would never complete the cycle to check the freshness of all zones as each incoming notification disrupted this cycle. Addressed in cooperation with Tyler Hall of EditDNS. Improvements: * Zoneparser improvements mean $TTL and $INCLUDES now work a lot better. Implemented in 1056, 1062. * No longer report temporary recvfrom errors, which used to spam the log on many systems. Addressed in commit 1320. * Direct queries for 'fancy records' would lead to errors, such queries now fail early. Spotted by Jorn Ekkelenkamp, implemented in 1051. * Fix typo in geobackend, closing ticket 157, implemented in 1090. * Initial work on TSIG support - not done yet. Spurred on by Marco Davids. * Embarrassingly, the 'master' configuration setting was not documented in the list of all settings! * Norbert has updated OpenDBX so that SQLite reads and writes no longer deadlock, plus compliation fixes on Solaris, plus the addition of autoserials to backends that support triggers. Implemented in commit 1154. * Random generator is now based on AES, improving the security of certain proxy operations. This is the same random generator that is in the recursor. Implemented in commit 1256. * Documentation for 'supermaster' mode was improved due to popular demand. * When binding to a UDP port failed, supply a more precise error message (commit 1245) * The zoneparser error messages were vastly improved, partially inspired by Shawn's cowboy migration. Code in commit 1235. * Labels are compressed more efficiently (case-insensitively), leading to smaller packets. Implemented in commit 1156. * Fix handling of TCP timeouts to not cause a reload of the backends. Implemented in commit 1092. * TCP Receiver no longer spams the log with common network errors. Implemented in commit 1306. * Move from select() to poll()-based multiplexing, allowing PowerDNS to listen on more than 1024 sockets simultaneously. One big PowerDNS user needs this. Implemented in 1072. * Zone2sql now reads source files in performance enhancing inode order. Additionally, zone2sql no longer dies on a missing zone file if --on-error-resume-next was specified. Finally, statistics of zone2sql conversion have been improved. Implemented in 1055. * Address issues found by more recent g++ versions. Spotted and/or fixed by Jorn Ekkelenkamp (commit 1051), Marcus Rueckert (commit 1094), Norbert Sendetzky (commit 1107), Serge Belyshev (commit 1171). * The Intel C Compiler implements certain things differently, causing the master/slave communicator to malfunction. Spotted by Marcus Rueckert, implemented in 1052, plus fallout in 1105. * PowerDNS can now be compiled with Boost 1.37.0. * Andre Lorbach of Adiscon discovered the microsoft windows 2003 nameserver adds out of zone data to zonetransfers, which we need to ignore, instead of rejecting the entire zone. Implemented in 1048. * PowerDNS now skips remote master servers which consistently generate timeout messages, improving the master checking cycle time tremendously. Developed in cooperation with Tyler Hall. Implemented in commit 1278. * When binding to a UDP port failed, supply a more precise error message (commit 1245) * dnsreplay now waits for the final answers to arrive, making it possible to process even small pcap files and get meaningful statistics. commit 1268. * dnsreplay has a more sane default timeout now, which can be configured too. Suggested by Augie Schwer in ticket 163, implemented in commit 1287. __________________________________________________________ 1.3.2. Authoritative Server version 2.9.21.2 Released on the 18th of November 2008. This release consists of a single patch to PowerDNS Authoritative Server version 2.9.21.1. In some configurations, notably with configuration option 'distributor-threads=1', the PowerDNS Authoritative Server crashes easily in some error conditions. All users are urged to upgrade. Even though PowerDNS restarts itself on encountering such error conditions, and even though most PowerDNS configurations do not run in single threaded mode, an upgrade is recommended. More detail can be found in Section 1.9. __________________________________________________________ 1.3.3. Authoritative Server version 2.9.21.1 Released on the 6th of August 2008. This release consists of a single patch to PowerDNS Authoritative Server version 2.9.21. Brian J. Dowling of Simplicity Communications has discovered a security implication of the previous PowerDNS behaviour to drop queries it considers malformed. We are grateful that Brian notified us quickly about this problem. This issue has been assigned CVE-2008-3337. The single patch is in commit 1239. More detail can be found in Section 1.8. The implication is that while the PowerDNS Authoritative server itself does not face a security risk because of dropping these malformed queries, other resolving nameservers run a higher risk of accepting spoofed answers for domains being hosted by PowerDNS Authoritative Servers before 2.9.21.1. While the dropping of queries does not aid sophisticated spoofing attempts, it does facilitate simpler attacks. It may be good to know that several large sites already run with this patch applied, as it has been in the public codebase for some weeks already. __________________________________________________________ 1.3.4. Recursor version 3.1.7 Released the 25th of June 2008. This version contains powerful scripting abilities, allowing operators to modify DNS responses in many interesting ways. Among other things, these abilities can be used to filter out malware domains, to perform load balancing, to comply with legal and other requirements and finally, to implement 'NXDOMAIN' redirection. It is hoped that the addition of Lua scripting will enable responsible DNS modification for those that need it. For more details about the Lua scripting, which can be modified, loaded and unloaded at runtime, see Section 12.6. Many thanks are due to the #lua irc channel, for excellent near-realtime Lua support. In addition, a number of PowerDNS users have been enthousiastically testing prereleases of the scripting support, and have found and solved many issues. In addition, 3.1.7 fixes a number of bugs: * In 3.1.5 and 3.1.6, an authoritative server could continue to renew its authority, even though a domain had been delegated to other servers in the meantime. In the rare cases where this happened, and the old servers were not shut down, the observed effect is that users were fed outdated data. Bug spotted and analysed by Darren Gamble, fix in commit 1182 and commit 1183. * Thanks to long time PowerDNS contributor Stefan Arentz, for the first time, Mac OS X 10.5 users can compile and run the PowerDNS Recursor! Patch in commit 1185. * Sten Spans spotted that for outgoing TCP/IP queries, the query-local-address setting was not honored. Fixed in commit 1190. * rec_control wipe-cache now also wipes domains from the negative cache, hurrying up the expiry of negatively cached records. Suggested by Simon Kirby, implemented in commit 1204. * When a forwarder server is configured for a domain, using the forward-zones setting, this server IP address was filtered using the dont-query setting, which is generally not what is desired: the server to which queries are forwarded will often live in private IP space, and the operator should be trusted to know what he is doing. Reported and argued by Simon Kirby, fix in commit 1211. * Marcus Rueckert of OpenSUSE reported that very recent gcc versions emitted a (correct) warning on an overly complicated line in syncres.cc, fixed in commit 1189. * Stefan Schmidt discovered that the netmask matching code, used by the new Lua scripts, but also by all other parts of PowerDNS, had problems with explicit '/32' matches. Fixed in commit 1205. __________________________________________________________ 1.3.5. Recursor version 3.1.6 Released on the 1st of May 2008. This version fixes two important problems, each on its own important enough to justify a quick upgrade. * Version 3.1.5 had problems resolving several slightly misconfigured domains, including for a time 'juniper.net'. Nameserver timeouts were not being processed correctly, leading PowerDNS to not update the internal clock, which in turn meant that any queries immediately following an error would time out as well. Because of retries, this would usually not be a problem except on very busy servers, for domains with different nameservers at different levels of the DNS-hierarchy, like 'juniper.net'. This issue was fixed rapidly because of the help of XS4ALL (Eric Veldhuyzen, Kai Storbeck), Brad Dameron and Kees Monshouwer. Fix in commit 1178. * The new high-quality random generator was not used for all random numbers, especially in source port selection. This means that 3.1.5 is still a lot more secure than 3.1.4 was, and its algorithms more secure than most other nameservers, but it also means 3.1.5 is not as secure as it could be. A quick upgrade is recommended. Discovered by Thomas Biege of Novell (SUSE), fixed in commit 1179. __________________________________________________________ 1.3.6. Recursor version 3.1.5 Released on the 31st of March 2008. Much like 3.1.4, this release does not add a lot of major features. Instead, performance has been improved significantly (estimated at around 20%), and many rare and not so rare issues were addressed. Multi-part TXT records now work as expected - the only significant functional bug found in 15 months. One of the oldest feature requests was fulfilled: version 3.1.5 can finally forward queries for designated domains to multiple servers, on differing port numbers if needed. Previously only one forwarder address was supported. This lack held back a number of migrations to PowerDNS. We would like to thank Amit Klein of Trusteer for bringing a serious vulnerability to our attention which would enable a smart attacker to 'spoof' previous versions of the PowerDNS Recursor into accepting possibly mallicious data. Details can be found on this Trusteer page. It is recommended that all users of the PowerDNS Recursor upgrade to 3.1.5 as soon as practicable, while we simultaneously note that busy servers are less susceptible to the attack, but not immune. The PowerDNS Security Advisory can be found in Section 1.7. This version can properly benefit from all IPv4 and IPv6 addresses in use at the root-servers as of early February 2008. In order to implement this, changes were made to how the Recursor deals internally with A and AAAA queries for nameservers, see below for more details. Additionally, newer releases of the G++ compiler required some fixes (see ticket 173). This release was made possible by the help of Wichert Akkerman, Winfried Angele, Arnoud Bakker (Fox-IT), Niels Bakker (no relation!), Leo Baltus (Nederlandse Publieke Omroep), Marco Davids (SIDN), David Gavarret (Neuf Cegetel), Peter Gervai, Marcus Goller (UPC), Matti Hiljanen (Saunalahti/Elisa), Ruben Kerkhof, Alex Kiernan, Amit Klein (Trusteer), Kenneth Marshall (Rice University), Thomas Rietz, Marcus Rueckert (OpenSUSE), Augie Schwer (Sonix), Sten Spans (Bit), Stefan Schmidt (Freenet), Kai Storbeck (xs4all), Alex Trull, Andrew Turnbull (No Wires) and Aaron Thompson, and many more who filed bugs anonymously, or who we forgot to mention. Security related issues: * Amit Klein has informed us that System random generator output can be predicted based on its past behaviour, allowing a smart attacker to 'spoof' our nameserver. Full details in Section 1.7. * The Recursor will by default no longer query private-space nameservers. This closes a slight security risk and simultaneously improves performance and stability. For more information, see dont-query in Section 12.1. Implemented in commit 923. * Applied fix for ticket 110 ('PowerDNS should change directory to '/' in chroot), implemented in commit 944. Performance: * The DNS packet writing and parsing infrastructure performance was improved in several ways, see commits 925, 926, 928, 931, 1021, 1050. * Remove multithreading overhead from the Recursor (commit 999). Bug fixes: * Built-in authoritative server now properly derives the TTL from the SOA record if not specified. Implemented in commit 1165. Additionally, even when TTL was specified for the built-in authoritative server, it was ignored. Reported by Stefan Schmidt, closing ticket 147. * Empty TXT record components can now be served. Implemented in commit 1166, closing ticket 178. Spotted by Matti Hiljanen. * The Recursor would not properly override old data with new, sometimes serving old and new data concurrently. Fixed in commit 1137. * SOA records with embedded carriage-return characters are now parsed correctly. Implemented in commit 1167, closing ticket 162. * Some routing conditions could cause UDP connected sockets to generate an error which PowerDNS did not deal with properly, leading to a leaked file descriptor. As these run out over time, the recursor could crash. This would also happen for IPv6 queries on a host with no IPv6 connectivity. Thanks to Kai of xs4all and Wichert Akkerman for reporting this issue. Fix in commit 1133. * Empty unknown record types can now be stored without generating a scary error (commit 1129) * Applied fix for ticket 111, ticket 112 and ticket 153 - large (multipart) TXT records are now retrieved and served properly. Fix in commit 996. * Solaris compilation instructions in Recursor documentation were wrong, leading to an instant crash on startup. Luckily nobody reads the documentation, except for Marcus Goller who found the error. Fixed in commit 1124. * On Solaris, finally fix the issue where queries get distributed strangely over CPUs, or not get distributed at all. Much debugging and analysing performed by Alex Kiernan, who also supplied fixes. Implemented in commit 1091, commit 1093. * Various fixes for modern G++ versions, most spotted by Marcus Rueckert (commits 964, 965, 1028, 1052), and Ruben Kerkhof (commit 1136, closing ticket 175). * Recursor would not properly clean up pidfile and control socket, closing ticket 120, code in commit 988, commit 1098 (part of fix by Matti Hiljanen, spotted by Leo Baltus) * Recursor can now serve multi-line records from its limited authoritative server (commit 1014). * When parsing zones, the 'm' time specification stands for minutes, not months! Closing Debian bug 406462 (commit 1026) * Authoritative zone parser did not support '@' in the content of records. Spotted by Marco Davids, fixed in commit 1030. * Authoritative zone parser could be confused by trailing TABs on record lines (commit 1062). * EINTR error code could block entire server if received at the wrong time. Spotted by Arnoud Bakker, fix in commit 1059. * Fix crash on NetBSD on Alpha CPUs, might improve startup behaviour on empty caches on other architectures as well (commit 1061). * Outbound TCP queries were being performed sub-optimally because of an interaction with the 'Mplexer'. Fixes in commit 1115, commit 1116. New features: * Implemented rec_control command get uptime, as suggested by Niels Bakker (commit 935). Added to default rrdtool scripts in commit 940. * The Recursor Authorative component, meant for having the Recursor serve some zones authoritatively, now supports $INCLUDE and $GENERATE. Implemented in commit 951 and commit 952, commit 967 (discovered by Thomas Rietz), * Implemented forward-zones-file option in order to support larger amounts of zones which should be forwarded to another nameserver (commit 963). * Both forward-zones and forward-zones-file can now specify multiple forwarders per domain, implemented in commit 1168, closing ticket 81. Additionally, both these settings can also specify non-standard port numbers, as suggested in ticket ticket 122. Patch authored by Aaron Thompson, with additional work by Augie Schwer. * Sten Spans contributed allow-from-file, implemented in commit 1150. This feature allows the Recursor to read access rules from a (large) file. General improvements: * Ruben Kerkhof fixed up weird permission bits as well as our SGML documentation code in commit 936 and commit 937. * Full IPv6 parity. If configured to use IPv6 for outgoing queries (using query-local-address6=::0 for example), IPv6 and IPv4 addresses are finally treated 100% identically, instead of 'mostly'. This feature is implemented using 'ANY' queries to find A and AAAA addresses in one query, which is a new approach. Treat with caution. * Now perform EDNS0 root refreshing queries, so as to benefit from all returned addresses. Relevant since early February 2008 when the root-servers started to respond with IPv6 addresses, which made the default non-EDNS0 maximum packet length reply no longer contain all records. Implemented in commit 1130. Thanks to dns-operations AT mail.oarc.isc.org for quick suggestions on how to deal with this change. * rec_control now has a timeout in case the Recursor does not respond. Implemented in commit 945. * (Error) messages are now logged with saner priorities (commit 955). * Outbound query IP interface stemmed from 1997 (!) and was in dire need of a cleanup (commit 1117). * L.ROOT-SERVERS.NET moved (commit 1118). __________________________________________________________ 1.3.7. PowerDNS Authoritative Server version 2.9.21 Released the 21st of April 2007. This is the first release the PowerDNS Authoritative Server since the Recursor was split off to a separate product, and also marks the transfer of the new technology developed specifically for the recursor, back to the authoritative server. This move has reduced the amount of code of the Authoritative server by over 2000 lines, while improving the quality of the program enormously. However, since so much has been changed, care should be taken when deploying 2.9.21. To signify the magnitude of the underlying improvements, the next release of the PowerDNS Authoritative Server will be called 3.0. This release would not have been possible without large amounts of help and support from the PowerDNS Community. We specifically want to thank Massimo Bandinelli of Italy's Register.it, Dave Aaldering of Aaldering ICT, True BV, XS4ALL, Daniel Bilik of Neosystem, EasyDNS, Heinrich Ruthensteiner of Siemens, Augie Schwer, Mark Bergsma, Marco Davids, Marcus Rueckert of OpenSUSE, Andre Muraro of Locaweb, Antony Lesuisse, Norbert Sendetzky, Marco Chiavacci, Christoph Haas, Ralf van der Enden and Ruben Kerkhof. Security issues: * The previous packet parsing and generating code contained no known bugs, but was however very lengthy and overly complex, and might have had security problems. The new code is 'inherently safe' because it relies on bounds-checking C++ constructs. Therefore, a move to 2.9.21 is highly recommended. * Pre-2.9.21, communication between master and server nameservers was not checked as rigidly as possible, possibly allowing third parties to disrupt but not modify such communications. Warning The 'bind1' legacy version of our BIND backend has been dropped! There should be no need to rely on this old version anymore, as the main BIND backend has been very well tested recently. Bugs: * Multi-part TXT records weren't supported. This has been fixed, and regression tests have been added. Code in commits 1016, 996, 994. * Email addresses with embedded dots in SOA records were not parsed correctly, nor were other embedded dots. Noted by 'Bastiaan', fixed in commit 1026. * BIND backend treated the 'm' TTL modifier as 'months' and not 'minutes'. Closes Debian bug 406462. Addressed in commit 1026. * Our snapshots were built against a static version of PosgreSQL that was incompatible with many Linux distributions, leading to instant crashes on startup. Fixed in 1022 and 1023. * CNAME referrals to child zones gave improper responses. Noted by Augie Schwer in ticket 123, fixed in commit 992. * When passing a port number with the recursor setting, this would sometimes generate errors during additional processing. Switched off overly helpful additional processing for recursive queries to remove this problem. Implemented in commit 1031, spotted by Ralf van der Enden. * NS to a nameserver with the name of the zone itself generated problems. Spotted by Augie Schwer, fixed in commit 947. * Multi-line records in the BIND backend were not always parsed correctly. Fixed in commit 1014. * The LOC-record had problems operating outside of the eastern hemisphere of the northern part of the world! Fixed in commit 1011. * Backends were compiled without multithreading preprocessor flags. As far as we can determine, this would only cause problems for the BIND backend, but we cannot rule out this caused instability in other backends. Fixed in commit 1001. * The BIND backend was highly unstable under reloads, and leaked memory and file descriptors. Thanks to Mark Bergsma and Massimo Bandinelli for respectively pointing this out to us and testing large amounts of patches to fix the problem. The fixes have resulted in better performance, less code, and a remarkable simplification of this backend. Commits 1039, 1034, 1035, 1006, 999, 905 and previous. * BIND backend gave convincing NXDOMAINS on unloaded zones in some cases. Spotted and fixed by Daniel Bilik in commit 984. * SOA records in zone transfers sometimes contained the wrong SOA TTL. Spotted by Christian Kuehn, fixed in commit 902. * PowerDNS could get confused by very high SOA serial numbers. Spotted and fixed by Dan Billik, fixed in commit 626. * Some versions of FreeBSD perform very strict checks on socket address sizes passed to 'connect', which could lead to problems retrieving zones over AXFR. Fixed in commit 891. * Some versions of FreeBSD perform very strict checks on IPv6 socket addresses, leading to problems. Discovered by Sten Spans, fixed in commit 885 and commit 886. * IXFR requests were not logged properly. Noted by Ralf van der Enden, fixed in commit 990. * Some NAPTR records needed an additional space character to encode correctly. Spotted by Heinrich Ruthensteiner, fixed in commit 1029. * Many bugs in the TCP nameserver, leading to a PowerDNS process that did not respond to TCP queries over time. Many fixes provided by Dan Bilik, other problems were fixed by rewriting our TCP handling code. Commits 982 and 980, 950, 924, 889, 874, 869, 685, 684. * Fix crashes on the ARM processor due to alignment errors. Thanks to Sjoerd Simons. Closes Debian bug 397031. * Missing data in generic SQL backends would sometimes lead to faked SOA serial data. Spotted by Leander Lakkas from True. Fix in commit 866. * When receiving two quick notifications in succession, the packet cache would sometimes "process" the second one, leading PowerDNS to ignore it. Spotted by Dan Bilik, fixed in commit 686. * Geobackend (by Mark Bergsma) did not properly override the getSOA method, breaking non-overlay operation of this fine backend. The geobackend now also skips '.hidden' configuration files, and now properly disregards empty configuration files. Additionally, the overlapping abilities were improved. Details available in commit 876, by Mark. Features: * Thanks to EasyDNS, PowerDNS now supports multiple masters per domain. For configuration details, see Section 13.2. Implemented in commit 1018, commit 1017. * Thanks to EasyDNS, PowerDNS now supports the KEY record type, as well the SPF record. In commit 976. * Added support for CERT, SSHFP, DNSKEY, DS, NSEC, RRSIG record types, as part of the move to the new DNS parsing/generating code. * Support for the AFSDB record type, as requested by 'Bastian'. Implemented in commit 978, closing ticket 129. * Support for the MR record type. Implemented in commit 941 and commit 1019. * Gsqlite3 backend was added by Antony Lesuisse in commit 942; * Added the ability to send out light-weight root-referrals that save bandwidth yet still placate mediocre resolver implementations. Implemented in commit 912, enable with 'root-referral=lean'. Improvements: * Miscellaneous OpenDBX and LDAP backend improvements by Norbert Sendetzky. Applied in commit 977 and commit 1040. * SGML source of the documentation was cleaned up by Ruben Kerkhof in commit 936. * Speedups in core DNS label processing code. Implemented in commit 928, commit 654, commit 1020. * When communicating with master servers and encountering errors, more useful details are logged. Reported by Stefan Arentz in ticket 137, closed by commit 1015. * Database errors are now logged with more details. Addressed in commit 1004. * pdns_control problems are now logged more verbosely. Change in commit 910. * Erroneous address configuration was logged unclearly. Spotted by River Tarnell, fixed in commit 888. * Example configuration shipped with PowerDNS was very old. Noted by Leen Besselink, fixed in commit 946. * PowerDNS neglected to chdir to the root when chrooted. This closes ticket 110, fixed in commit 944. * Microsoft resolver had problems with responses we generated for CNAMEs pointing out of our bailiwick. Fixed in commit 983 and expedited by Locaweb.com.br. * Built-in webserver logs errors more verbosely. Closes ticket 82, gixed in commit 991. * Queries containing '@' no longer flood the logs. Addressed in commit 1014. * The build process now looks for PostgreSQL in more places. Implemented in commit 998, closes ticket 90. * Speedups in the BIND backend now mean large installations enjoy startup times up to 30 times faster than with the original BIND nameserver. Many thanks to Massimo Bandinelli. * BIND backend now offers full support for query logging, implemented in commit 1026, commit 1029. * BIND backend named.conf parsing is now fully case-insensitive for domain names. This closes Debian bug 406461, fixed in commit 1027. * IPv6 and IPv4 address parsing routines have been replaced, which should result in prettier output in some cases. commit 962, commit 1012 and others. * 5 new regression tests have been added to insure old bugs do not return. * Fix small issues with very modern compilers and BOOST snapshots. Noted by Marcus Rueckert, addressed in commit 954, commit 964 commit 965, commit 1003. __________________________________________________________ 1.3.8. Recursor version 3.1.4 Released the 13th of November 2006. This release contains almost no new features, but consists mostly of minor and major bug fixes. It also addresses two major security issues, which makes this release a highly recommended upgrade. Security issues: * Large TCP questions followed by garbage could cause the recursor to crash. This critical security issue has been assigned CVE-2006-4251, and is fixed in commit 915. More information can be found in Section 1.5. * CNAME loops with zero second TTLs could cause crashes in some conditions. These loops could be constructed by malicious parties, making this issue a potential denial of service attack. This security issue has been assigned CVE-2006-4252 and is fixed by commit 919. More information can be found in Section 1.6. Many thanks to David Gavarret for helping pin down this problem. Bugs: * On certain error conditions, PowerDNS would neglect to close a socket, which might therefore eventually run out. Spotted by Stefan Schmidt, fixed in commits 892, 897, 899. * Some nameservers (including PowerDNS in rare circumstances) emit a SOA record in the authority section. The recursor mistakenly interpreted this as an authoritative "NXRRSET". Spotted by Bryan Seitz, fixed in commit 893. * In some circumstances, PowerDNS could end up with a useless (not working, or no longer working) set of nameserver records for a domain. This release contains logic to invalidate such broken NSSETs, without overloading authoritative servers. This problem had previously been spotted by Bryan Seitz, 'Cerb' and Darren Gamble. Invalidations of NSSETs can be plotted using the "nsset-invalidations" metric, available through rec_control get. Implemented in commit 896 and commit 901. * PowerDNS could crash while dumping the cache using rec_control dump-cache. Reported by Wouter of WideXS and Stefan Schmidt and many others, fixed in commit 900. * Under rare circumstances (depleted TCP buffers), PowerDNS might send out incomplete questions to remote servers. Additionally, on big-endian systems (non-Intel and non-AMD generally), sending out large TCP answers questions would not work at all, and possibly crash. Brought to our attention by David Gavarret, fixed in commit 903. * The recursor contained the potential for a dead-lock processing an invalid domain name. It is not known how this might be triggered, but it has been observed by 'Cerb' on #powerdns. Several dead-locks where PowerDNS consumed all CPU, but did not answer questions, have been reported in the past few months. These might be fixed by commit 904. * IPv6 'allow-from' matching had problems with the least significant bits, sometimes allowing disallowed addresses, but mostly disallowing allowed addresses. Spotted by Wouter from WideXS, fixed in commit 916. Improvements: * PowerDNS has support to drop answers from so called 'delegation only' zones. A statistic ("dlg-only-drops") is now available to plot how often this happens. Implemented in commit 890. * Hint-file parameter was mistakenly named "hints-file" in the documentation. Spotted by my Marco Davids, fixed in commit 898. * rec_control quit should be near instantaneous now, as it no longer meticulously cleans up memory before exiting. Problem spotted by Darren Gamble, fixed in commit 914, closing ticket 84. * init.d script no longer refers to the Recursor as the Authoritative Server. Spotted by Wouter of WideXS, fixed in commit 913. * A potentially serious warning for users of the GNU C Library version 2.5 was fixed. Spotted by Marcus Rueckert, fixed in commit 920. __________________________________________________________ 1.3.9. Recursor version 3.1.3 Released the 12th of September 2006. Compared to 3.1.2, this release again consists of a number of mostly minor bug fixes, and some slight improvements. Many thanks are again due to Darren Gamble who together with his team has discovered many misconfigured domains that do work with some other name servers. DNS has long been tolerant of misconfigurations, PowerDNS intends to uphold that tradition. Almost all of the domains found by Darren now work as well in PowerDNS as in other name server implementations. Thanks to some recent migrations, this release, or something very close to it, is powering over 40 million internet connections that we know of. We appreciate hearing about succesful as well as unsuccesful migrations, please feel free to notify pdns.bd@powerdns.com of your experiences, good or bad. Bug-fixes: * The MThread default stack size was too small, which led to problems, mostly on 64-bit platforms. This stack size is now configurable using the stack-size setting should our estimate be off. Discovered by Darren Gamble, Sten Spans and a number of others. Fixed in commit 868. * Plug a small memory leak discovered by Kai and Darren Gamble, fixed in commit 870. * Switch from the excellent nedmalloc to dlmalloc, based on advice by the nedmalloc author. Nedmalloc is optimised for multithreaded operation, whereas the PowerDNS recursor is single threaded. The version of nedmalloc shipped contained a number of possible bugs, which are probably resolved by moving to dlmalloc. Some reported crashes on hitting 2G of allocated memory on 64 bit systems might be solved by this switch, which should also increase performance. See commit 873 for details. Improvements: * The cache is now explicitly aware of the difference between authoritative and unauthoritative data, allowing it to deal with some domains that have different data in the parent zone than in the authoritative zone. Patch in commit 867. * No longer try to parse DNS updates as if they were queries. Discovered and fixed by Jan Gyselinck, fix in commit 871. * Rebalance logging priorities for less log cluttering and add IP address to a remote server error message. Noticed and fixed by Jan Gyselinck (commit 877). * Add logging-facility setting, allowing syslog to send PowerDNS logging to a separate file. Added in commit 871. __________________________________________________________ 1.3.10. Recursor version 3.1.2 Released Monday 26th of June 2006. Compared to 3.1.1, this release consists almost exclusively of bug-fixes and speedups. A quick update is recommended, as some of the bugs impact operators of authoritative zones on the internet. This version has been tested by some of the largest internet providers on the planet, and is expected to perform well for everybody. Many thanks are due to Darren Gamble, Stefan Schmidt and Bryan Seitz who all provided excellent feedback based on their large-scale tests of the recursor. Bug-fixes: * Internal authoritative server did not differentiate between 'NXDOMAIN' and 'NXRRSET', in other words, it would answer 'no such host' when an AAAA query came in for a domain that did exist, but did not have an AAAA record. This only affects users with auth-zones configured. Discovered by Bryan Seitz, fixed in commit 848. * ANY queries for hosts where nothing was present in the cache would not work. This did not cause real problems as ANY queries are not reliable (by design) for anything other than debugging, but did slow down the nameserver and cause unnecessary load on remote nameservers. Fixed in commit 854. * When exceeding the configured maximum amount of TCP sessions, TCP support would break and the nameserver would waste CPU trying to accept TCP connections on UDP ports. Noted by Bryan Seitz, fixed in commit 849. * DNS queries come in two flavours: recursion desired and non-recursion desired. The latter is not very useful for a recursor, but is sometimes (erroneously) used by monitoring software or loadbalancers to detect nameserver availability. A non-rd query would not only not recurse, but also not query authoritative zones, which is confusing. Fixed in commit 847. * Non-standard DNS TCP queries, that did occur however, could drive the recursor to 100% CPU usage for extended periods of time. This did not disrupt service immediately, but does waste a lot of CPU, possibly exhausting resources. Discovered by Bryan Seitz, fixed in commit 858, which is post-3.1.2-rc1. * The PowerDNS recursor did not honour the rare but standardised 'ANY' query class (normally 'ANY' refers to the query type, not class), upsetting the Wildfire Jabber server. Discovered and debugged by Daniel Nauck, fixed in commit 859, which is post-3.1.2-rc1. * Everybody's favorite, when starting up under high load, a bogus line of statistics was sometimes logged. Fixed in commit 851. * Remove some spurious debugging output on dropping a packet by an unauthorized host. Discovered by Kai. Fixed in commit 854. Improvements: * Misconfigured domains, with a broken nameserver in the parent zone, should now work better. Changes motivated and suggested by Darren Gamble. This makes PowerDNS more compliant with RFC 2181 by making it prefer authoritative data over non-authoritative data. Implemented in commit 856. * PowerDNS can now listen on multiple ports, using the local-address setting. Added in commit 845. * A number of speedups which should have a noticeable impact, implemented in commits 850, 852, 853, 855 * The recursor now works around an issue with the Linux kernel 2.6.8, as shipped by Debian. Fixed by Christof Meerwald in commit 860, which is post 3.1.2-rc1. __________________________________________________________ 1.3.11. Recursor version 3.1.1 Warning 3.1.1 is identical to 3.1 except for a bug in the packet chaining code which would mainly manifest itself for IPv6 enabled Konqueror users with very fast connections to their PowerDNS installation. However, all 3.1 users are urged to upgrade to 3.1.1. Many thanks to Alessandro Bono for his quick aid in solving this problem. Released on the 23rd of May 2006. Many thanks are due to the operators of some of the largest internet access providers in the world, each having many millions of customers, who have tested the various 3.1 pre-releases for suitability. They have uncovered and helped fix bugs that could impact us all, but are only (quickly) noticeable with such vast amounts of DNS traffic. After version 3.0.1 has proved to hold up very well under tremendous loads, 3.1 adds important new features: * Ability to serve authoritative data from 'BIND' style zone files (using auth-zones statement). * Ability to forward domains so configured to external servers (using forward-zones). * Possibility of 'serving' the contents of /etc/hosts over DNS, which is very well suited to simple domestic router/DNS setups. Enabled using export-etc-hosts. * As recommended by recent standards documents, the PowerDNS recursor is now authoritative for RFC-1918 private IP space zones by default (suggested by Paul Vixie). * Full outgoing IPv6 support (off by default) with IPv6 servers getting equal treatment with IPv4, nameserver addresses are chosen based on average response speed, irrespective of protocol. * Initial Windows support, including running as a service ('NET START "POWERDNS RECURSOR"'). rec_channel is still missing, the rest should work. Performance appears to be below that of the UNIX versions, this situation is expected to improve. Bug fixes: * No longer send out SRV and MX record priorities as zero on big-endian platforms (UltraSPARC). Discovered by Eric Sproul, fixed in commit 773. * SRV records need additional processing, especially in an Active Directory setting. Reported by Kenneth Marshall, fixed in commit 774. * The root-records were not being refreshed, which could lead to problems under inconceivable conditions. Fixed in commit 780. * Fix resolving domain names for nameservers with multiple IP addresses, with one of these addresses being lame. Other nameserver implementations were also unable to resolve these domains, so not a big bug. Fixed in commit 780. * For a period of 5 minutes after expiring a negative cache entry, the domain would not be re-cached negatively, leading to a lot of duplicate outgoing queries for this short period. This fix has raised the average cache hit rate of the recursor by a few percent. Fixed in commit 783. * Query throttling was not aggressive enough and not all sorts of queries were throttled. Implemented in commit 786. * Fix possible crash during startup when parsing empty configuration lines (commit 807). * Fix possible crash when the first query after wiping a cache entry was for the just deleted entry. Rare in production servers. Fixed in commit 820. * Recursor would send out differing TTLs when receiving a misconfigured, standards violating, RRSET with different TTLs. Implement fix as mandated by RFC 2181, paragraph 5.2. Reported by Stephen Harker (commit 819). * The top-remotes would list remotes duplicately, once per source port. Discovered by Jorn Ekkelenkamp, fixed in commit 827, which is post 3.1-pre1. * Default allow-from allowed queries from fe80::/16, corrected to fe80::/10. Spotted by Niels Bakker, fixed in commit 829, which is post 3.1-pre1. * While PowerDNS blocks failing queries quickly, multiple packets could briefly be in flight for the same domain and nameserver. This situation is now explicitly detected and queries are chained to identical queries already in flight. Fixed in commit 833 and commit 834, post 3.1-pre1. Improvements: * ANY queries are now implemented as in other nameserver implementations, leading to a decrease in outgoing queries. The RFCs are not very clear on desired behaviour, what is implemented now saves bandwidth and CPU and brings us in line with existing practice. Previously ANY queries were not cached by the PowerDNS recursor. Implemented in commit 784. * rec_control was very sparse in its error reporting, and user unfriendly as well. Reported by Erik Bos, fixed in commit 818 and commit 820. * IPv6 addresses were printed in a non-standard way, fixed in commit 788. * TTLs of records are now capped at two weeks, commit 820. * allow-from IPv4 netmasks now automatically work for IP4-to-IPv6 mapper IPv4 addresses, which appear when running on the wildcard :: IPv6 address. Lack of feature noted by Marcus 'darix' Rueckert. Fixed in commit 826, which is post 3.1-pre1. * Errors before daemonizing are now also sent to syslog. Suggested by Marcus 'darix' Rueckert. Fixed in commit 825, which is post 3.1-pre1. * When launching without any form of configured network connectivity, all root-servers would be cached as 'down' for some time. Detect this special case and treat it as a resource-constraint, which is not accounted against specific nameservers. Spotted by Seth Arnold, fixed in commit 835, which is post 3.1-pre1. * The recursor now does not allow authoritative servers to keep supplying its own NS records into perpetuity, which causes problems when a domain is redelegated but the old authorative servers are not updated to this effect. Noticed and explained at length by Darren Gamble of Shaw Communications, addressed by commit 837, which is post 3.1-pre2. * Some operators may want to follow RFC 2181 paragraph 5.2 and 5.4. This harms performance and does not solve any real problem, but does make PowerDNS more compliant. If you want this, enable auth-can-lower-ttl. Implemented in commit 838, which is post 3.1-pre2. __________________________________________________________ 1.3.12. Recursor version 3.0.1 Released 25th of April 2006, download. This release consists of nothing but tiny fixes to 3.0, including one with security implications. An upgrade is highly recommended. * Compilation used both cc and gcc, leading to the possibility of compiling with different compiler versions (commit 766). * rec_control would leave files named lsockXXXXXX around in the configured socket-dir. Operators may wish to remove these files from their socket-dir (often /var/run), quite a few might have accumulated already (commit 767). * Certain malformed packets could crash the recursor. As far as we can determine these packets could only lead to a crash, but as always, there are no guarantees. A quick upgrade is highly recommended (commits 760, 761). Reported by David Gavarret. * Recursor would not distinguish between NXDOMAIN and NXRRSET (commit 756). Reported and debugged by Jorn Ekkelenkamp. * Some error messages and trace logging statements were improved (commits 756, 758, 759). * stderr was closed during daemonizing, but not dupped to /dev/null, leading to slight chance of odd behaviour on reporting errors (commit 757) Operating system specific fixes: * The stock Debian sarge Linux kernel, 2.6.8, claims to support epoll but fails at runtime. The epoll self-testing code has been improved, and PowerDNS will fall back to a select based multiplexer if needed (commit 758) Reported by Michiel van Es. * Solaris 8 compilation and runtime issues were addressed. See the README for details (commit 765). Reported by Juergen Georgi and Kenneth Marshall. * Solaris 10 x86_64 compilation issues were addressed (commit 755). Reported and debugged by Eric Sproul. __________________________________________________________ 1.3.13. Recursor version 3.0 Released 20th of April 2006, download. This is the first separate release of the PowerDNS Recursor. There are many reasons for this, one of the most important ones is that previously we could only do a release when both the recursor and the authoritative nameserver were fully tested and in good shape. The split allows us to release new versions when each part is ready. Now for the real news. This version of the PowerDNS recursor powers the network access of over two million internet connections. Two large access providers have been running pre-releases of 3.0 for the past few weeks and results are good. Furthermore, the various pre-releases have been tested nearly non-stop with DNS traffic replayed at 3000 queries/second. As expected, the 2 million househoulds shook out some very rare bugs. But even a rare bug happens once in a while when there are this many users. We consider this version of the PowerDNS recursor to be the most advanced resolver publicly available. Given current levels of spam, phishing and other forms of internet crime we think no recursor should offer less than the best in spoofing protection. We urge all operators of resolvers without proper spoofing countermeasures to consider PowerDNS, as it is a Better Internet Nameserver Daemon. A good article on DNS spoofing can be found here. Some more information, based on a previous version of PowerDNS, can be found on the PowerDNS development blog. Warning Because of recent DNS based denial of service attacks, running an open recursor has become a security risk. Therefore, unless configured otherwise this version of PowerDNS will only listen on localhost, which means it does not resolve for hosts on your network. To fix, configure the local-address setting with all addresses you want to listen on. Additionally, by default service is restricted to RFC 1918 private IP addresses. Use allow-from to selectively open up the recursor for your own network. See Section 12.1 for details. Important new features of the PowerDNS recursor 3.0: * Best spoofing protection and detection we know of. Not only is spoofing made harder by using a new network address for each query, PowerDNS detects when an attempt is made to spoof it, and temporarily ignores the data. For details, see Section 12.4.1. * First nameserver to benefit from epoll/kqueue/Solaris completion ports event reporting framework, for stellar performance. * Best statistics of any recursing nameserver we know of, see Section 12.5. * Last-recently-used based cache cleanup algorithm, keeping the 'best' records in memory * First class Solaris support, built on a 'try and buy' Sun CoolThreads T 2000. * Full IPv6 support, implemented natively. * Access filtering, both for IPv4 and IPv6. * Experimental SMP support for nearly double performance. See Section 12.3. Many people helped package and test this release. Jorn Ekkelenkamp of ISP-Services helped find the '8000 SOAs' bug and spotted many other oddities and XS4ALL internet funded a lot of the recent development. Joaquín M López Muñoz of the boost::multi_index_container was again of great help. __________________________________________________________ 1.3.14. Version 2.9.20 Released the 15th of March 2006 Besides adding OpenDBX, this release is mostly about fixing problems and speeding up the recursor. This release has been made possible by XS4ALL and True. Thanks! Furthermore, we are very grateful for the help of Andrew Pinski, who hacks on gcc, and of Joaquín M López Muñoz, the author of boost::multi_index_container. Without their near-realtime help this release would've been delayed a lot. Thanks! Bugs fixed in the recursor: * Possible stability issues in the recursor on encountering errors (commit 532, commit 533) * Memory leaks in recursor fixed (commit 534, commit 572). In a test 800 million real life DNS packets have been sent to the recursor, representing several days of traffic from a major ISP, memory use was high (500MB), but stable. * Prune all data in PowerDNS - previously per-nameserver and per-query performance statistics were kept around forever (commit 535) * IPv6 additional processing was broken. Reported by Lionel Elie Mamane, who also provided a fix. The problem was fixed differently in the end. commit 562. * pdns_recursor did not shuffle answers since 2.9.19, leading to problems sending mail to the Hotmail servers. Reported in ticket 54, fixed in commit 567. * If a single nameserver had multiple IP addresses listed, PowerDNS would only use one of them. Noted by Mark Martin, fixed in commit 570, who depends on a domain with 4 nameserver IP addresses of which 2 are broken. Improvements to the recursor: * Commits 535, 540, 541, 542, 543, 544, 545, 547 and 548, 574 all speed up the recursor by a large factor, without altering the DNS algorithm. * Move recursor to the incredible boost::multi_index_container (commit 580). This brings a huge improvement in cache pruning times. * commit 549 and commit 550 work around gcc bug 24704 if requested, which speeds up the recursor a lot, but involves a dirty hack. Enable with ./configure --enable-gcc-skip-locking. No guarantees! Bugs fixed in the authoritative nameserver: * PowerDNS would no longer allow a '/' in domain names, fixed by commit 537, reported in ticket 48. * Parameters to pdns_control notify-host were not checked, leading to possible crashes. Reported in ticket 24, fixed in commit 565. * On some compilers, processing of NAPTR records could cause the server to crash. Reported by Bernd Froemel in ticket 29, fixed in commit 538. * Backend errors could make the whole nameserver exit under some circumstances, notably using the LDAP backend. Fixed in commit 583, reported in ticket 62. * Referrals were subtly broken by recent CNAME/Wildcard improvements, fixed in commit 539. Fix and other improvements sponsored by True. * PowerDNS would try to insert records it has no knowledge about in slave zones, which did not work. Reported in ticket 60, fixed in commit 566. A superior fix would be to implement the relevant unknown record standard. Improvements to the authoritative nameserver: * Pipebackend did not properly propagate the ABI version to its children, fixed in commit 546, reported by kickdaddy@gmail.com in ticket 45. * OpenDBX backend added (commit 559, commit 560, commit 561) by Norbert Sendetzky. From the website: " The OpenDBX backend enables it to fetch DNS information from every DBMS supported by the OpenDBX library and combines the power of one of the best DNS server implementations with the flexibility of the OpenDBX library. " OpenDBX adds some other features like database failover. Thanks Norbert! * LDAP fixes as reported in ticket 37, fixed in commit 558, which maked pdns_control notify work. * Arjo Hooimeijer added support for soa-refresh-default, soa-retry-default, soa-expire-default, which were previously hardcoded. commit 563 and fallout in commit 573 (thanks to Wolfram Schlich). Miscellaneous: * Fixes for g++ 4.1. Compiling with 4.1 realizes notable speedups. commit 568, commit 569. * PowerDNS now reports if it is running in 32 or 64 bit mode, useful for bi-arch users that need to know if they are benefitting from their great processor. commit 571. * dnsscope compiles again, commit 551, commit 564 (FreeBSD 64-bit time_t). * dnsreplay_mindex compiles again, fixed by commit 572. Its performance, and the performance of the recursor was improved by commit 559. * Build scripts were added, mostly for internal use but we know some PowerDNS users build their own packages too. commit 553, commit 554, commit 555, commit 556, commit 557. * bootstrap script was not included in release. Thanks to Stefan Arentz for noticing. Fixed in commit 574. __________________________________________________________ 1.3.15. Version 2.9.19 Released 29th of October 2005. As with other recent releases, the usage of PowerDNS appears to have skyrocketed. Informal, though strict, measurements show that PowerDNS now powers around 50% of all German domains, and somewhere in the order of 10-15% of the rest of the world. Furthermore, DNS is set to take a central role in connecting Voice over IP providers, with PowerDNS offering a very good feature set for these ENUM deployments. PowerDNS is already powering the E164.info ENUM zone and also acts as the backend for a major VoIP provisioning platform. Included in this release is the now complete packet parsing/generating, record parsing/generating infrastructure. Furthermore, this framework is used by the recursor, hopefully making it very fast, memory efficient and robust. Many records are now processed using a single line of code. This has made the recursor a lot stricter in packet parsing, you will see some error messages which did not appear before. Rest assured however that these only happen for queries which have no valid answer in any case. Furthermore, support for DNSSEC records is available in the new infrastructure, although is should be emphasised that there is more to DNSSEC than parsing records. There is no real support for DNSSEC (yet). Additionally, the BIND Backend has been replaced by what was up to now known as the 'Bind2Backend'. Initial benchmarking appears to show that this backend is faster, uses less memory and has shorter startup times. The code is also shorter. This release fixes a number of embarassing bugs and is a recommended upgrade. Thanks are due to XS4ALL who are supporting continuing development of PowerDNS, the fruits of which can be found in this release already. Furthermore, a remarkable number of people have helped report bugs, validate solutions or have submitted entire patches. Many thanks! Improvements: * dnsreplay now has a help message and has received further massive updates, making the code substantially faster. It turns out that dnsreplay is often 'heavier' than the PowerDNS process being benchmarked. * PowerDNS recursor no longer prints out its queries by default as most recursor deployments have too much traffic for this to be useful. * PowerDNS recursor is now able to read its root-hints from disk, which is useful to operate with alternate roots, like the Open Root Server Network. See Chapter 12. * PowerDNS can now send out old-fashioned root-referrals when queried for domains for which it is not authoritative. Wastes some bandwidth but may solve incoming query floods if domains are delegated to you for which you are not authoritative, but which are queried by broken recursors. * PowerDNS now prints out a warning when running with legacy LinuxThreads implementation instead of the high performance NPTL library, see Section 9.2. commit 455. * A lot of superfluous calls to gettimeofday() have been removed, making PowerDNS and especially the recursor faster. Suggested by Kai. * SPF records are now supported natively. commit 472, closing ticket 22. * Improved IPv6 'bound to' messages. Thanks to Niels Bakker, Wichert Akkerman and Gerty de Wolf for suggestions. * Separate graphs can now be made of IPv6 queries and answers. commit 485. * Out of zone additional processing is now on by default to better comply with standards. commit 487. * Regression tests have been expanded to deal with more record types (SRV, NAPTR, TXT, duplicate SRV). * Improved query-logging in Bindbackend, which can be used for debugging purposes. * Dropped libpcap dependency, making compilation easier * pdns_control now has a help message. * Add RRSIG, DNSKEY, DS and NSEC records for DNSSEC-bis to new parser infrastructure. * Recursor now honours EDNS0 allowing it to send out larger answers. Bugs fixed: * Domain name validation has been made a lot stricter - it turns out PostgreSQL was interpreting some (corrupt) domain names as unicode. Tested and suggested by Register.com (commit 451). * LDAP backend did not compile (commits 452, 453) due to partially applied patch (Norbert Sendetzky) * Incoming zone transfers work reliably again. Fixed in commit 460 and beyond. And commit 523 - closing Debian bug 330184. * Recent g++ versions exposed a mistake in the PowerDNS recursor cache pruning code, causing random crashes. Fixed in commit 465. Reported by several Red Hat users. * PowerDNS recursor, and MTasker in general, did not work on Solaris. Patch by Juergen Ilse, commit 471. Also moved most of PowerDNS over to uint32_t style typedefs, which eases compilation problems on Solaris, commit 477. * Bindbackend2 did not properly search its include path for $INCLUDE statements. Noted by Mark Bergsma, commit 474. * Bindbackend did not notice changed zones, this problem has been fixed by the move to Bind2. * Pipebackend did not clean up, leading to an additional pipe backend per AXFR or pdns_control reload. Discovered by Marc Jauvin, fixed by commit 525. * Bindbackend (both old and current versions) did not honour 'include' statements in named.conf on pdns_control rediscover. Noted by Marc Jauvin, fixed by commit 526. * Zone transfers were sometimes shuffled, which wastes useless time, commit 478. * CNAMEs and Wildcards now work as in Bind, fixing many complaints, commit 487. * NAPTR records were compressed, which would work, but was in violation of the RFC, commit 493. * NAPTR records were not always parsed correctly from BIND zonefiles, fixed, commit 494. * Geobackend needed additional include statement to compile on more recent Linux distrbutions, commit 496. __________________________________________________________ 1.3.16. Version 2.9.18 Released on the 16th of July 2005. The '8 million domains' release, which also marks the battle readiness of the PowerDNS Recursor. The latest improvements have been made possible by financial support and contributions by Register.com and XS4ALL. Thanks! This release brings a number of new features (vastly improved recursor, Generic Oracle Support, DNS analysis and replay tools, and more) but also has a new build dependency, the Boost library (version 1.31 or higher). Currently several big ISPs are evaluating the PowerDNS recursor for their resolving needs, some of them have switched already. In the course of testing, over 350 million actual queries have been recorded and replayed, the answers turn out to be satisfactorily. This testing has verified that the pdns recursor, as shipped in this release, can stand up to heavy duty ISP loads (over 20000 queries/second) and in fact does so better than major other nameservers, giving more complete answers and being faster to boot. We invite ISPs who note recursor problems to record their problematic traffic and replay it using the tools described in Chapter 20 to discover if PowerDNS does a better job, and to let us know the results. Additionally, the bind2backend is almost ready to replace the stock bind backend. If you run with Bind zones, you are cordially invited to substitute 'launch=bind2' for 'launch=bind'. This will happen automatically in 2.9.19! In other news, the entire Wikipedia constellation now runs on PowerDNS using the Geo Backend! Thanks to Mark Bergsma for keeping us updated. There are two bugs with security implications, which only apply to installations running with the LDAP backend, or installations providing recursion to a limited range of IP addresses. If any of these apply to you, an upgrade is highly advised: * The LDAP backend did not properly escape all queries, allowing it to fail and not answer questions. We have not investigated further risks involved, but we advise LDAP users to update as quickly as possible (Norbert Sendetzky, Jan de Groot) * Questions from clients denied recursion could blank out answers to clients who are allowed recursion services, temporarily. Reported by Wilco Baan. This would've made it possible for outsiders to blank out a domain temporarily to your users. Luckily PowerDNS would send out SERVFAIL or Refused, and not a denial of a domain's existence. General bugs fixed: * TCP authoritative server would not relaunch a backend after failure (reported by Norbert Sendetzky) * Fix backend restarting logic (reported, and fix suggested by Norbert Sendetzky) * Launching identical backends multiple times, with different settings, did not work. Reported by Mario Manno. * Master/slave queries did not honour the query-local-address setting. Spotted by David Levy of Register.com. The fix also randomises the local port used, slightly improving security. Compilation fixes: * Fix compile on Solaris, they define 'PC' for some reason. Reported by Eric Yiu. * PowerDNS recursor would not compile on FreeBSD due to Linux specific defines, as reported in cvstrac ticket 26 (Ralf van der Enden) * Several 64 bits issues have been fixed, especially in the Logging subsystem. * SSQLite would fail to compile on recent Debian systems (Matthijs Mohlmann) * Generic MySQL would not compile on 64-bit platforms. Improvements: * PowerDNS now reports stray command line arguments, like when running '--local-port 5300' instead of '--local-port=5300'. Reported by Christian Welzel. * We now warn against erroneous logging-facility specification, ie specifying an unknown facility. * --version now outputs gcc version used, so we can tell people 2.95 is no longer supported. * Extended regression tests, moved them to the new 'sdig' tool (see below). * Bind2backend is now blazingly fast, and highly memory efficient to boot. As a special bonus it can read gzipped zones directly. The '.NET' zone is hosted using 401MB of memory, the same size as the zone on disk. * The Pipe Backend has been improved such that it can send out different answers based on the IP address the question was received ON. See Section A.1.1 for how this changed the Pipe Backend protocol. Note that you need to set pipebackend-abi-version to benefit from this change, existing clients are not affected. Change and documentation contributed by Marc Jauvin of Register4Less. * LDAP backend has been updated (Norbert Sendetzky). Recursor improvements and fixes. See Chapter 11 for details. The changes below mean that all of the caveats listed for the recursor have now been addressed. * After half an hour of uptime, the entire cache would be pruned for each packet, which is a tad slow. It now appears the pdns recursor is among the faststest around. * Under high loads, or when unlucky, some query mthreads would get 'stuck', and show up in the statistics as eternally running queries. * Lots of redundant gettimeofday() and time() calls were removed, which has resulted in a measurable speedup. * pdns_recursor can now listen on several addresses simultaneously. * Now supports setuid and setgid operation to allow running as a less privileged user (Bram Vandoren) * Return code of pdns_recursor binary did not make sense (Matthijs Mohlmann and Thomas Hood) * Timeouts and errors are now split out in statistics. * Many people reported broken statistics, it turned out that no statistics were being reported if there had been no questions to base them on. We now log a message to that effect. * Add query-local-address support, which allows the recursor to send questions from a specific IP address. Useful for anycast setups. * Add outgoing TCP query support and proper truncated answer support. Needed for Worldnic Denial of Service protection, which sends out truncated packets to force clients to connect over TCP, which prevents spoofing. * Properly truncate our own answers. * Improve our TCP answers by using writev, which is slightly friendlier to the network. * On FreeBSD, TCP errors could cause the recursor to exit suddenly due to a SIGPIPE signal. * Maximum number of simultaneous client TCP connections can now be limited with the max-tcp-clients setting. * Add agressive timeouts for TCP clients to make sure resources are not wasted. Defaults to two seconds, can be configured with the client-tcp-timeout setting. Backend fixes: * SQLite backend would not slave properly (Darron Broad) * Generic MySQL would not compile on 64-bit platforms. New technology: * Added the new DNS parser logic, called MOADNSParser. Completely modular, every memory access checked. * 'sdig', a simple dig workalike with 'canonical' output, which is used for the regression tests. Based on the new DNS parser logic. * dnswasher, dnsreplay and dnsscope, all DNS analysis tools. See Chapter 20 for more details. * Generic Oracle Backend, sponsored by Register.COM. See Section A.5.3. __________________________________________________________ 1.3.17. Version 2.9.17 See the new timeline for progress reports. The 'million domains' release - PowerDNS has now firmly established itself as a major player with the unofficial count (ie, guesswork) now at over two million PowerDNS domains! Also, the GeoBackend has been tested by a big website and may soon see wider deployment. Thanks to Mark Bergsma for spreading the word! It is also a release with lots of changes and fixes. Take care when deploying! Security issues: * PowerDNS could be temporarily DoSed using a random stream of bytes. Reported cause of this has been fixed. Enhancements: * Reported version can be changed, or removed - see the "version-string" setting. * Duplicate MX records are now no longer considered duplicate if their priorities differ. Some people need this feature for spam filtering. Bug fixes: * NAPTR records can now be slaved, patch by Lorens Kockum. * GMySQL now works on Solaris * PowerDNS could be confused by questions with a %-sign in them - fixing cvstrac ticket #16 (reported by dilinger at voxel.net) * An authentication bug in the webserver was possibly fixed, please report if you were suffering from this. Being unable to authenticate to the webserver was what you would've noticed. * Fix for cvstrac ticket #2, PowerDNS could lose sync when sending out a very large number of notifications. Excellent bug report by Martin Hoffman, who also improved our original bugfix. * Fix the oldest PowerDNS bug in existence - under some circumstances, PowerDNS would log to syslog one character at a time. This was cvstrac ticket #4 * HINFO records can now be slaved, fixing cvstrac ticket #8. * pdns_recursor could block under some circumstances, especially in case of corrupt UDP packets. Reported by Wichert Akkerman. Fix by Christopher Meer. This was cvstrac ticket #13. * Large SOA serial numbers would sometimes be logged as a signed integer, leading to negative numbers in the log. * PowerDNS now fully supports 32 bit SOA serial numbers (thanks to Mark Bergsma), closing cvstrac ticket #5. * pdns_recursor --local-address help text was wrong. * Very devious bug - PowerDNS did not clear its cache before sending out update notifications, leading slaves to conclude there was no update to AXFR. Excellent debugging by mkuchar at wproduction.cz. * Probably fixed cvstrac ticket #26, which caused pdns_recursor to fail on recent FreeBSD 5.3 systems. Please check, I have no such system to test on. * Geobackend did not get built for Debian. __________________________________________________________ 1.3.18. Version 2.9.16 The 'it must still be Friday somewhere' release. Massive number of fixes, portability improvements and the new Geobackend by Mark Bergsma & friends. New: * The Geobackend which makes it possible to send different answers to different IP ranges. Initial documentation can be found in pdns/modules/geobackend/README. * qgen query generation tool. Nearly completely undocumented and hard to build too, it requires Boost. But very spiffy. Use cd pdns; make qgen to build it. Bugfixes: * The most reported bug ever was fixed. Zone2sql required the inclusion of unistd.h, except on Debian unstable. * PowerDNS tried to listen on its control "pipe" which does not work. Probably harmless, but might have caused some oddities. * The Packet Cache did not always set its TTL immediately, causing some packets to be inserted, even when running with the cache disabled (Mark Bergsma). * Valgrind found some unitialized reads, causing bogus values in the priority field when it was not needed * Valgrind found a bug in MTasker where we used delete instead of delete[]. * SOA serials and other parameters are unsigned. This means that very large SOA serial numbers would be messed up (Michel Stol, Stefano Straus) * PowerDNS left its controlsocket around after exit and reported confusing errors if a socket was already in use. * The recursor proxy did not work on big endian systems like SPARC and some MIPS processors (Remco Post) * We no longer dump core on processing LOC records on UltraSPARC (Andrew Mulholland supplied a testing machine) Improvements: * MySQL can now connect to a specified port again (Chris Anderton) * When running chroot()ed and with master or slave support active, PowerDNS needs to resolve domain names to find slaves. This in turn may require access to certain libraries. Previously, these needed to be available in the chroot directory but by forcing an initial lookup, these libraries are now loaded before the chrooting. * pdns_recursor was very slow after having done a larger number of queries because of the checks to see if a query should be throttled. This is now done using a set which is a lot faster than the previous full sequential scan. * The throttling code may not have throttled as much as was configured. * Yet another big LDAP update. The LDAP backend now loadbalances connections over several hosts (Norbert Sendetzky) * Updated b.root-servers.net address in the recursor __________________________________________________________ 1.3.19. Version 2.9.15 This release fixes up some of the shortcomings in 2.9.14, and adds some new features too. Bugfixes: * allow-recursion-override was on by default, it was meant to be off. * Logging was still off in daemon mode, fixed. * debian/rules forgot to build an sqllite package * Recursor accidentally linked in MySQL - this was the result of an experiment with a persistent recursor cache. * The PowerDNS recursor had stability problems. It now sorts nameservers (roughly) by responsiveness. The 'roughly' part upset the sorting algorithm used, the speeds being sorted on changed during sorting. * The recursor now outputs the nameserver average response times in trace mode * LDAP compiles again. Improvements: * zone2sql can now accept - as a filename which causes it to read stdin. This allows the following to work: dig axfr ds9a.nl | zone2sql --gmysql --zone=- | mysql pdns, which is a nice way to import a zone. * zone2sql now ignores duplicate SOA records which are identical - which also makes the above possible. * Remove libpqpp dependencies - since we now use the native C API for PostgreSQL __________________________________________________________ 1.3.20. Version 2.9.14 Big release with the fix for the all important 2^30 seconds problem and a lot of other news. * errno problems would cause compilation problems when using LDAP (Norbert Sendetzky) * The Generic SQL backend could cause crashes on PostgreSQL when using pdns_control notify (Georg Bauer) * Debian compatible init.d script (Wichert Akkerman) * If using the master or slave features, pdns had the notion of eternity ending in 2038, except that due to a thinko, eternity ended out to be the 10th of January 2004. This caused a loop to timeout immediately. Many thanks to Jasper Spaans for spotting the bug within five minutes. * Parts of the SOA field were not cannonicalized * The loglevel could in fact cause nothing to be logged (Norbert Sendetzky) Improvements: * The recursor now chooses the fastest nameserver, which causes a big speedup! * LDAP now has different lookup models * Cleanups, better load distribution, better exception handling, zone2ldap improvements * The recursor was somewhat chatty about TCP connections * PostgreSQL now only depends on the C API and not on the deprecated C++ one * PowerDNS can now fully overrule external zones when doing recursion. See Chapter 11. __________________________________________________________ 1.3.21. Version 2.9.13 Big news! Windows is back! Our great friend Michel Stol found the time to update the PowerDNS code so it works again under windows. Furthermore, big thanks go out to Dell who quickly repaired my trusty laptop. His changes: * Generic SQLite support added * Removed the ODBC backend, replaced it by the Generic ODBC Backend, which has all the cool configurability of the Generic MySQL and PostgreSQL backends. * The PowerDNS Recursor now runs as a Service. It defaults to running on port 5300, PowerDNS itself is configured to expect the Recursor on port 5300 now. * The PowerDNS Service is now known as 'PowerDNS' to Windows. * The Installer was redone, this time with NSIS2. * General updates and fixes. Other news: Note There appears to be a problem with PowerDNS on Red Hat 7.3 with GCC 2.96 and self-compiled binaries. The symptoms are that PowerDNS works on the foreground but fails as a daemon. We're working on it. If you do note problems, let the list know, if you don't, please do so as well. Tell us if you use the RPM or compiled yourself. It is known that not compiling in MySQL support helps solve the problem, but then you don't have MySQL. There have been a number of reports on MySQL connections being dropped on FreeBSD 4.x, which sometimes causes PowerDNS to give up and reload itself. To combat this, MySQL error messages have been improved in some places in hopes of figuring out what is up. The initial indication is that MySQL itself sometimes terminates the connection and, amazingly, that switching to a Unix domain socket instead of TCP solves the problem. Bug fixes: * allow-axfr-ips did not work for individual IP addresses (bug & fix by Norbert Sendetzky) Improvements: * Opteron support! Thanks to Jeff Davey for providing a shell on an Opteron. The fixes should also help PowerDNS on other platforms with a 64 bit userspace. Btw, the PowerDNS team has a strong desire for an Opteron :-) * pdns_recursor jumbles answers now. This means that you can do poor man's roundrobin by supplying multiple A, MX or AAAA records for a service, and get a random one on top each time. Interestingly, this feature appeared out of nowhere, this change was made to the authoritative code but due to the wonders of code-reuse had an effect on pdns_recursor too. * Big LDAP cleanup. Support for TLS was added. Zone2LDAP also gained the ability to generate ldif files containing a tree or a list of entries. (Norbert Sendetzky) * Zone2sql is now somewhat clearer when reporting malformed line errors - it did not always include the name of the file causing a problem, especially for big installations. Problem noted by Thom May. * pdns_recursor now survives the expiration of all its root records, most often caused by prolonged disconnection from the net. __________________________________________________________ 1.3.22. Version 2.9.12 Release rich in features. Work on Verisign oddities, addition of SQLite backend, pdns_recursor maturity. New features: * --version command (requested by Mike Benoit) * delegation-only, a Verisign special. * Generic SQLite support, by Michel 'Who da man?' Stol. See Section A.7. * init.d script for pdns_recursor * Recursor now actually purges its cache, saving memory. * Slave configuration now no longer falls over when presented with a NULL master * Bindbackend2 now has supermaster support (Mark Bergsma, untested) * Answers are now shuffled! It turns out a few recursors don't do shuffling (pdns_recursor, djbdns), so we do it now. Requested by Jorn Ekkelenkamp of ISP-Services. This means that if you have multiple IP addresses for one host, they will be returned in differing order every once in a while. Bugs: * 0.0.0.0/0 didn't use to work (Norbert Sendetzky) * pdns_recursor would try to resolve IP address which to bind to, potentially causing chicken/egg problem * gpgsql no longer reports as gmysql (Sherwin Daganoto) * SRV would not be parsed right from disk (Christof Meerwald) * An AXFR from a zone hosted on the LDAP backend no longer transmits all the reverse entries too (Norbert Sendetzky) * PostgreSQL backend now does error checking. It would be a bit too trusting before. Improvements, cleanups: * PowerDNS now reports the numerical IP addresses it binds to instead of the, possibly, alphanumeric names the operator passed. * Removed only-soa hackery (noticed by Norbert Sendetzky) * Debian packaging fixes (Wichert Akkerman) * Some parameter descriptions were improved. * Cleanups by Norbert: getAuth moved to chopOff, arguments::contains massive cleanup, more. __________________________________________________________ 1.3.23. Version 2.9.11 Yet another iteration, hopefully this will be the last silly release. Warning There has been a change in behaviour whereby disable-axfr does what it means now! From now on, setting allow-axfr-ips automatically disables AXFR from unmentioned subnets. This release enables AXFR again, disable-axfr did the opposite of what it claimed. Furthermore, the pdns_recursor now cleans its cache, which should save some memory in the long run. Norbert contributed some small LDAP work which should come in useful in the future. __________________________________________________________ 1.3.24. Version 2.9.10 Small bugfixes, LDAP update. Released 3rd of July 2003. Apologies for the long delay, real life keeps interfering. Warning Do not use or try to use 2.9.9, it was a botched release! Warning There has been a change in behaviour whereby disable-axfr does what it means now! From now on, setting allow-axfr-ips automatically disables AXFR from unmentioned subnets. * 2.9.8 was prone to crash on adding additional records. Thanks to excellent debugging by PowerDNS users worldwide, the bug was found quickly and is in fact present in all earlier PowerDNS releases, but for some reason doesn't cause crashes there. * Notifications now jump in front of the queue of domains that need to be checked for changes, giving much greater perceived performance. This is needed if you have tens of thousands of slave domains and your master server is on a high latency link. Thanks to Mark Jeftovic of EasyDNS for suggesting this change and testing it on their platform. * Dean Mills reported that PowerDNS does confusing logging about changing GIDs and UIDs, fixed. Cosmetic only. * pdns_recursor may have logged empty lines for some users, fixed. Solution suggested by Norbert Sendetzky. * LDAP: DNS TTLs were random values (Norbert Sendetzky, Stefan Pfetzing). New ldap-default-ttl option. * LDAP: Now works with OpenLDAP 2.1 (Norbert Sendetzky) * LDAP: error handling for invalid MX records implemented (Norbert Sendetzky) * LDAP: better exception handling (Norbert Sendetzky) * LDAP: code cleanup of lookup() (Norbert Sendetzky) * LDAP: added support for scoped searches (Norbert Sendetzky) __________________________________________________________ 1.3.25. Version 2.9.8 Queen's day release! 30th of April 2003. Added support for AIX, fixed negative SOA caching. Some other cleanups. Not a major release but enough reasons to upgrade. Bugs fixed: * Recursor had problems expiring negatively cached entries, which wasted memory and also led to the continued non-existence of hosts that since had come into existence. * The Generic SQL backends did not lowercase the names of records, which led to new records not being found by case sensitive databases (notably PostgreSQL). Found by Volker Goetz. * NS queries for zones for which we did not carry authority, but only had delegation information, had their NS records in the wrong section. Minor detail, but a standards violation on etheless. Spotted by Stephane Bortzmeyer. Improvements: * Removed crypt.h dependency from powerldap.hh, which was a problem on some platforms (Richard Arends) * PowerDNS can't parse so called binary labels which we now detect and ignore, after printing a warning. * Specifying allow-axfr-ips now automatically disables AXFR for all non-mentioned addresses. * A Solaris ready init.d script is now part of the tar.gz (contributed, but I lost by whom). * Added some fixes to PowerDNS can work on AIX (spotted by Markus Heimhilcher). * Norbert Sendetzky contributed zone2ldap. * Everybody's favorite compiler warning from zone2sql.cc was removed! * Recursor now listens on TCP! __________________________________________________________ 1.3.26. Version 2.9.7 Released on 2003-03-20. This is a sweeping release in the sense of cleanup. There are some new features but mostly a lot of cleanup going on. Hiding inside is the bind2backend, the next generation of the bind backend. A work in progress. Those of you with overlapping zones, as mentioned in the changelog of 2.9.6, are invited to check it out by replacing launch=bind by launch=bind2 and renaming all bind- parameters to bind2-. Be aware that if you run with many small zones, this backend is faster, but if you run with a few large ones, it is slower. This will improve. Features: * Mark Bergsma contributed query-local-address which allows the operator to select which source address to use. This is useful on servers with multiple source addresses and the operating system selecting an unintended one, leading to remotes denying access. * PowerDNS can now perform AAAA additional processing optionally, turned on by setting do-ipv6-additional-processing. Thanks to Stephane Bortzmeyer for pointing out the need. * Bind2backend, which is almost in compliance with the new IETF AXFR-clarify (some would say 'redefinition') draft. This backend is not ready for primetime but you may want to try it if you currently have overlapping zones and note problems. An overlapping zone would be having "ipv6.powerdns.com" and "powerdns.com" zones on one server. Improvements: * Zone2sql would happily try to read from a directory and not give a useful error about this. * PowerDNS now reports the case where it can't figure out any IP address of slave nameservers for a zone * Removed receiver-threads setting which was experimental and in fact only made things worse. * LDAP backend updates from its author Norbert Sendetzky. Reverse lookups should work now too. * An error message about unparseable packets did not include the originating IP address (fixed by Mark Bergsma) * PowerDNS can now be started via path resolution while running with a guardian. Suggested by Maurice Nonnekes. * pdns_recursor moved to sbin (reported by Norbert Sendetzky) * Retuned some logger errorlevels, a lot of master/slave chatter was logged as 'Error'. Reported by Willem de Groot. Bugs fixed: * zone2sql did not remove trailing dots in SOA records. * ldapbackend did not include utility.hh which caused compilation problems on Solaris (reported by Remco Post) * pdns_control could leave behind remnants in case PowerDNS was not running (reported by dG) * Incoming AXFR did not work on Solaris and other big-endian systems (Willem de Groot helped debugging this long standing problem). * Recursor could crash on convoluted CNAME loops. Thanks to Dan Faerch for delivering coredumps. * Silly 'wuh' debugging output in zone2sql and bindbackend removed (spotted by Ivo van der Wijk) * Recursor neglected to differentiate between negative cache of NXDOMAIN and NOERROR, leading to problems with IPv6 enabled Windows clients. Thanks to Stuart Walsh for reporting this and testing the fix. * PowerDNS set the 'aa' bit on serving NS records in a zone for which it was authoritative. Most implementations drop the 'aa' bit in this case and Stephane Bortzmeyer informed us of this. PowerDNS now also drops the 'aa' bit in this case. * The webserver tended to fail after prolonged operation on FreeBSD, this was due to an uninitialised timeout, other platforms were lucky. Thanks to G.P. de Boer for helping debug this. * getAnswers() in dnspacket.cc could be forced to read bytes beyond the end of the packet, leading to crashes in the PowerDNS recursor. This is an ongoing project that needs more work. Reported by Dan Faerch, with a coredump proving the problem. __________________________________________________________ 1.3.27. Version 2.9.6 Two new backends - Generic ODBC (windows only) and LDAP. Furthermore, a few important bugs have been fixed which may have hampered sites seeing a lot of outgoing zonetransfers. Additionally, the pdns recursor now has 'query throttling' which is pretty cool. In short this makes sure that PowerDNS does not send out heaps of queries if a nameserver is unable to provide an answer. Many operators of authoritative setups are all too aware of recursing nameservers that hammer them for zones they don't have, PowerDNS won't do that anymore now, no matter what clients request of it. Warning There is an unresolved issue with the BIND backend and 'overlapping' slave zones. So if you have 'example.com' and also have a separate slave zone called 'external.example.com', things may go wrong badly. Thanks to Christian Laursen for working with us a lot in finding this issue. We hope to resolve it soon. * BIND Backend now honours notifies, code to support this was accidentally left out. Thanks to Christian Laursen for noticing this. * Massive speedup for those of you using the slightly deprecated MBOXFW records. Thanks to Jorn of ISP Services for helping and testing this improvement. * $GENERATE had an off-by-one bug where it would omit the last record to be generated (Christian Laursen) * Simultaneous AXFRs may have been problematic on some backends. Thanks to Jorn of ISP-Services again for helping us resolve this issue. * Added LDAP backend by Norbert Sendetzky, see Section A.12. * Added Generic ODBC backend for Windows by Michel Stol. * Simplified 'out of zone data' detection in incoming AXFR support, hopefully removing a case sensitivity bug there. Thanks again to Christian Laursen for reporting this issue. * $include in-zonefile was broken under some circumstances, losing the last character of a filename. Thanks to Joris Vandalon for noticing this. * The zoneparser was more case-sensitive than BIND, refusing to accept 'in' as well as 'IN'. Thanks to Joris Vandalon for noticing this. __________________________________________________________ 1.3.28. Version 2.9.5 Released on 2002-02-03. This version is almost entirely about recursion with major changes to both the pdns recursor, which is renamed to 'pdns_recursor' and to the main PowerDNS binary to make it interact better with the recursing component. Sadly, due to technical reasons, compiling the pdns recursor and pdns authoritative nameserver into one binary is not immediately possible. During the release of 2.9.4 we stated that the recursing nameserver would be integrated in the next release - this won't happen now. However, this turns out to not be that bad at all. The recursor can now be restarted without having to restart the rest of the nameserver, for example. Cooperation between the both halves of PDNS is also almost seamless. As a result, 'non-lazy recursion' has been dropped. See Chapter 11 for more details. Furthermore, the recursor only works on Linux, Windows and Solaris (not entirely). FreeBSD does not support the required functions. If you know any important FreeBSD people, plea with them to support set/get/swapcontext! Alternatively, FreeBSD coders could read the solution presented here in figure 5. The 'Contributor of the Month' award goes to Mark Bergsma who has responded to our plea for help with the label compressor and contributed a wonderfully simple and right fix that allows PDNS to compress just as well as Other namerervers out there. An honorary mention goes to Ueli Heuer who, despite having no C++ experience, submitted an excellent SRV record implementation. Excellent work was also performed by Michel Stol, the Windows guy, in fixing all our non-portable stuff again. Christof Meerwald has also done wonderful work in porting MTasker to Windows, which was then used by Michel to get the recursor functioning on Windows. Other changes: * dnspacket.cc was cleaned up by factoring out common operations * Heaps of work on the recursing nameserver. Has now achieved *days* of uptime! * Recursor renamed from syncres to pdns_recursor * PowerDNS can now serve records it does not know about. To benefit from this slightly undocumented feature, add 1024 to the numerical type of a record and include the record in binary form in your database. Used internally by the recursing nameserver but you can use it too. * PowerDNS now knows about SIG and KEY records *names*. It does not support them yet but can at least report so now. * HINFO records can now be transferred from a master to PowerDNS (thanks to Ueli Heuer for noticing it didn't work). * Yet more UltraSPARC alignment issues fixed (Chris Andrews). * Dropped non-lazy recursion, nobody was using it. Lazy recursion became even more lazy after Dan Bernstein pointed out that additional processing is not vital, so PowerDNS does its best to do additional processing on recursive queries, but does not scream murder if it does not succeed. Due to caching, the next identical query will be successfully additionally processed. * Label compression was improved so we can now fit all . records in 436 bytes, this used to be 460! (Code & formal proof of correctness by Mark Bergsma). * SRV support (incoming and outgoing), submitted by Ueli Heuer. * Generic backends do not support SOA serial autocalculation, it appears. Could lead to random SOA serials in case of a serial of 0 in the database. Fixed so that 0 stays zero in that case. Don't set the SOA serial to 0 when using Generic MySQL or Generic PostgreSQL! * J root-server address was updated to its new location. * SIGUSR1 now forces the recursor to print out statistics to the log. * Meaning of recursor logging was changed a bit - a cache hit is now a question that was answered with 0 outgoing packets needed. Used to be a weighted average of internal cache hits. * MySQL compilation did not include -lz which causes problems on some platforms. Thanks to James H. Cloos Jr for reporting this. * After a suggestion by Daniel Meyer and Florus Both, the built in webserver now reports the configuration name when multiple PowerDNS instances are active. * Brad Knowles noticed that zone2sql had problems with the root.zone, fixed. This also closes some other zone2sql annoyances with converting single zones. __________________________________________________________ 1.3.29. Version 2.9.4 Yet another grand release. Big news is the addition of a recursing nameserver which has sprung into existence over the past week. It is in use on several computers already but it is not ready for prime time. Complete integration with PowerDNS is expected around 2.9.5, for now the recursor is a separate program. In preliminary tests, the recursor appears to be four times faster than BIND 9 on a naive benchmark starting from a cold cache. BIND 9 managed to get through to some slower nameservers however, which were given up on by PowerDNS. We will continue to tune the recursor. See Chapter 12 for further details. The BIND Backend has also been tested (see the bind-domain-status item below) rather heavily by several parties. After some discussion online, one of the BIND authors ventured that the newsgroup comp.protocols.dns.bind may now in fact be an appropriate venue for discussing PowerDNS. Since this discussion, traffic to the PowerDNS pages has increased sixfold and shows no signs of slowing down. From this, it is apparent that far more people are interested in PowerDNS than yet know about it. So spread the word! In other news, we now have a security page at Section 1.4. Furthermore, Maurice Nonnekes contributed an OpenBSD port! See his page for more details! New features and improvements: * All SQL queries in the generic backends are now available for configuration. (Martin Klebermass, bert hubert). See Section A.5. * A recursing nameserver! See Chapter 12. * An incoming AXFR now only starts a backend zone replacement transaction after the first record arrived successfully, thus making sure no work is done when a remote nameserver is unable/unwilling to AXFR a zone to us. * Zoneparser error messages were improved slightly (thanks to Stef van Dessel for spotting this shortcoming) * XS4ALL's Erik Bos checked how PowerDNS reacted to a BIND installation with almost 60.000 domains, some of which with >100.000 records, and he discovered the pdns_control bind-domain-status command became very slow with larger numbers of domains. Fixed, 60.000 domains are now listed in under one second. * If a remote nameserver disconnects during an incoming AXFR, the update is now rolled back, unless the AXFR was properly terminated. * The migration chapter mentioned the use of deprecated backends. A tremendous number of bugs were discovered and fixed: * Zone parser would only accept $include and not $INCLUDE * Zone parser had problems with $lines with comments on the end * Wildcard ANY queries were broken (thanks Colemarcus for spotting this) * A connection failure with the Generic backends would lead to a powerdns reload (cast of many) * Generic backends had some semantic problems with slave support. Symptoms were oft-repeated notifications and transfers (thanks to Mark Bergsma for helping resolve this). * Solaris version compiles again. Thanks to Mohamed Lrhazi for reporting that it didn't. * Some UltraSPARC alignment fixes. Thanks to Mohamed Lrhazi for being helpful in spotting these. One problem is still outstanding, Mohamed sent a core dump that tells us where the problem is. Expect the fix to be in 2.9.5. Volunteers can grep the source for 'UltraSPARC' to find where the problem is. * Our support of IPv6 on FreeBSD had phase of moon dependent bugs, fixed by Peter van Dijk. * Some crashes of and by pdns_control were fixed, thanks to Mark Bergsma for helping resolve these. * Outgoing AXFR in pdns installations with multiple loaded backends was broken (thanks to Stuart Walsh for reporting this). * A failed BIND Backend incoming AXFR would block the zone until it succeeded again. * Generic PostgreSQL backend wouldn't compile with newer libpq++, fixed by Julien Lemoine/SpeedBlue. * Potential bug (not observed) when listening on multiple interfaces fixed. * Some typos in manpages fixed (reported by Marco Davids). __________________________________________________________ 1.3.30. Version 2.9.3a Note 2.9.3a is identical to 2.9.3 except that zone2sql does work Broad range of huge improvements. We now have an all-static .rpm and .deb for Linux users and a link to an OpenBSD port. Major news is that work on the Bind backend has progressed to the point that we've just retired our last Bind server and replaced it with PowerDNS in Bind mode! This server is operating a number of master and slave setups so it should stress the Bind backend somewhat. This version is rapidly approaching the point where it is a better-Bind-than-Bind and nearly a drop-in replacement for authoritative setups. PowerDNS is now equipped with a powerful master/slave apparatus that offers a lot of insight and control to the user, even when operating from Bind zonefiles and a Bind configuration. Observe. After the SOA of ds9a.nl was raised: pdns[17495]: All slave domains are fresh pdns[17495]: 1 domain for which we are master needs notifications pdns[17495]: Queued notification of domain 'ds9a.nl' to 195.193.163.3 pdns[17495]: Queued notification of domain 'ds9a.nl' to 213.156.2.1 pdns[17520]: AXFR of domain 'ds9a.nl' initiated by 195.193.163.3 pdns[17520]: AXFR of domain 'ds9a.nl' to 195.193.163.3 finished pdns[17521]: AXFR of domain 'ds9a.nl' initiated by 213.156.2.1 pdns[17521]: AXFR of domain 'ds9a.nl' to 213.156.2.1 finished pdns[17495]: Removed from notification list: 'ds9a.nl' to 195.193.163.3 (was acknowledged) pdns[17495]: Removed from notification list: 'ds9a.nl' to 213.156.2.1 (w as acknowledged) pdns[17495]: No master domains need notifications If however our slaves would ignore us, as some are prone to do, we can send some additional notifications: $ sudo pdns_control notify ds9a.nl Added to queue pdns[17492]: Notification request for domain 'ds9a.nl' received pdns[17492]: Queued notification of domain 'ds9a.nl' to 195.193.163.3 pdns[17492]: Queued notification of domain 'ds9a.nl' to 213.156.2.1 pdns[17495]: Removed from notification list: 'ds9a.nl' to 195.193.163.3 (was acknowledged) pdns[17495]: Removed from notification list: 'ds9a.nl' to 213.156.2.1 (w as acknowledged) Conversely, if PowerDNS needs to be reminded to retrieve a zone from a master, a command is provided: $ sudo pdns_control retrieve forfun.net Added retrieval request for 'forfun.net' from master 212.187.98.67 pdns[17495]: AXFR started for 'forfun.net', transaction started pdns[17495]: Zone 'forfun.net' (/var/cache/bind/forfun.net) reloaded pdns[17495]: AXFR done for 'forfun.net', zone committed Also, you can force PowerDNS to reload a zone from disk immediately with pdns_control bind-reload-now. All this happens 'live', per your instructions. Without instructions, the right things also happen, but the operator is in charge. For more about all this coolness, see Section B.1.1 and Section A.9.2. Warning Again some changes in compilation instructions. The hybrid pgmysql backend has been split up into 'gmysql' and 'gpgsql', sharing a common base within the PowerDNS server itself. This means that you can no longer compile --with-modules="pgmysql" --enable-mysql --enable-pgsql but that you should now use: --with-modules="gmysql gpgsql". The old launch-names remain available. If you launch the Generic PgSQL backend as gpgsql2, all parameters will have gpsql2 as a prefix, for example pgsql2-dbname. If launched as gpsql, the regular names are in effect. Warning The pdns_control protocol was changed which means that older pdns_controls cannot talk to 2.9.3. The other way around is broken too. This may lead to problems with automatic upgrade scripts, so pay attention if your daemon is truly restarted. Also make sure no old pdns_control command is around to confuse things. Improvements: * Bind backend can now deal with missing files and try to find them later. * Bind backend is now explicitly master capable and triggers the sending of notifications. * General robustness improvements in Bind backend - many errors are now non-fatal. * Accessability, Serviceability. New pdns_server commands like bind-list-rejects (lists zones that could not be loaded, and the reason why), bind-reload-now (reload a zone from disk NOW), rediscover (reread named.conf NOW). More is coming up. * Added support for retrieving RP (Responsible Person) records from remote masters. Serving them was already possible. * Added support for LOC records, which encode the geographical location of a host, both serving and retrieving (thanks to Marco Davids using them on our last Bind server, forcing us to implement this silly record). * Configuration file parser now strips leading spaces too, allowing "chroot= /tmp" to work, as well as "chroot=/tmp" (Thanks to Hub Dohmen for reporting this for months on end). * Added bind-domain-status command that shows the status of all domains (when/if they were parsed, any errors encountered while parsing them). * Added bind-reload-now command that tries to reload a zone from disk NOW, and reports back errors to the operator immediatly. * Added retrieve command that queues a request to retrieve a zone from its master. * Zones retrieved from masters are now stored way smaller on disk because the domain is stripped from records, which is derived from the configuration file. Retrieved zones are now prefixed with some information on where they came from. Changes: * gpgsql and gmysql backends split out of the hybrid pgmysqlbackend. This again changed compilation instructions! * pdns_control now uses the rarely seen SOCK_STREAM Unix Domain socket variety so it can transport large amounts of text, which is needed for the bind-domain-status command, for which see Section A.9.2. This breaks compatibility with older pdns_control and pdns_server binaries! * Bind backend now ignores 'hint' and 'forward' and other unsupported zone types. * AXFRs are now logged more heavily by default. An AXFR is a heavy operation anyhow, some more logging does not further increase the load materially. Does help in clearing up what slaves are doing. * A lot of master/slave chatter has been silenced, making output more relevant. No more repetitive 'No master domains need notifications' etc, only changes are reported now. Bugfixes: * Windows version did not compile without minor changes. * Confusing error reporting on Windows 98 (which does not support PowerDNS) fixed * Potential crashes with shortened packets addressed. An upgrade is advised! * notify (which was already there, just badly documented) no longer prints out debugging garbage. * pgmysql backend had problems launching when not compiled in but available as a module. Workaround for 2.9.2 is 'load-modules=pgmysql', but even then gpgsql would not work! gmysql would then, however. These modules are now split out, removing such issues. __________________________________________________________ 1.3.31. Version 2.9.2 Bugfixes galore. Solaris porting created some issues on all platforms. Great news is that PowerDNS is now in Debian 'sid' (unstable). The 2.9.1 packages in there currently aren't very good but the 2.9.2 ones will be. Many thanks to Wichert Akkerman, our 'downstream' for making this possible. Warning The Generic MySQL backend, part of the Generic MySQL & PostgreSQL backend, is now the DEFAULT! The previous default, the 'mysql' backend (note the lack of 'g') is now DEPRECATED. This was the source of much confusion. The 'mysql' backend does not support MASTER or SLAVE operation. The Generic backends do. To get back the mysql backend, add --with-modules="mysql" or --with-dynmodules="mysql" if you prefer to load your modules at runtime. Bugs fixed: * Silly debugging output removed from the webserver (found by Paul Wouters) * SEVERE: due to Solaris portability fixes, qtypes<127 were broken. These include NAPTR, ANY and AXFR. The upshot is that powerdns wasn't performing outgoing AXFRs nor ANY queries. These were the 'question for type -1' warnings in the log * incoming AXFR could theoretically miss some trailing records (not observed, but could happen) * incoming AXFR did not support TXT records (spotted by Paul Wouters) * with some remotes, an incoming AXFR would not terminate until a timeout occured (observed by Paul Wouters) * Documentation bug, pgmysql != mypgsql Documentation: * Documented the 'random backend', see Section A.3. * Wichert Akkerman contributed three manpages. * Building PowerDNS on Unix is now documented somewhat more, see Section D.1. Features: * pdns init.d script is now +x by default * OpenBSD is on its way of becoming a supported platform! As of 2.9.2, PowerDNS compiles on OpenBSD but swiftly crashes. Help is welcome. * ODBC backend (for Windows only) was missing from the distribution, now added. * xdb backend added - see Section A.11. Designed for use by root-server operators. * Dynamic modules are back which is good news for distributors who want to make a pdns packages that does not depend one every database under the sun. __________________________________________________________ 1.3.32. Version 2.9.1 Thanks to the great enthusiasm from around the world, powerdns is now available for Solaris and FreeBSD users again! Furthermore, the Windows build is back. We are very grateful for the help of: * Michel Stol * Wichert Akkerman * Edvard Tuinder * Koos van den Hout * Niels Bakker * Erik Bos * Alex Bleker * steven stillaway * Roel van der Made * Steven Van Steen We are happy to have been able to work with the open source community to improve PowerDNS! Changes: * The monitor command set no longer allows the changing of non-existant variables. * IBM Universal Database DB2 backend now included in source distribution (untested!) * Oracle backend now included in source distribution (sligthly tested!) * configure script now searches for postgresql and mysql includes * Bind parser now no longer dies on records with a ' in them (Erik Bos) * The pipebackend was accidentally left out of 2.9 * FreeBSD fixes (with help from Erik Bos, Alex Bleeker, Niels Bakker) * Heap of Solaris work (with help from Edvard Tuinder, Stefan Van Steen, Koos van den Hout, Roel van der Made and especially Mark Bakker). Now compiles in 2.7 and 2.8, haven't tried 2.9. May be a bit dysfunctional on 2.7 though - it won't do IPv6 and it won't serve AAAA. Patches welcome! * Windows 32 build is back! Michel Stol updated his earlier work to the current version. * S/Linux (Linux on Sparc) build works now (with help from steven stillaway). * Silly debugging message ('sd.ttl from cache') removed * .debs are back, hopefully in 'sid' soon! (Wichert Akkerman) * Removal of bzero and other less portable constructs. Discovered that recent Linux glibc's need -D_GNU_SOURCE (Wichert Akkerman). __________________________________________________________ 1.3.33. Version 2.9 Open source release. Do not deploy unless you know what you are doing. Stability is expected to return with 2.9.1, as are the binary builds. * License changed to the GNU General Public License version 2. * Cleanups by Erik Bos @ xs4all. * Build improvements by Wichert Akkerman * Lots of work on the build system, entirely revamped. By PowerDNS. __________________________________________________________ 1.3.34. Version 2.8 From this release onwards, we'll concentrate on stabilising for the 3.0 release. So if you have any must-have features, let us know soonest. The 2.8 release fixes a bunch of small stability issues and add two new features. In the spirit of the move to stability, this release has already been running 24 hours on our servers before release. * pipe backend gains the ability to restricts its invocation to a limited number of requests. This allows a very busy nameserver to still serve packets from a slow perl backend. * pipe backend now honors query-logging, which also documents which queries were blocked by the regex. * pipe backend now has its own backend chapter. * An incoming AXFR timeout at the wrong moment had the ability to crash the binary, forcing a reload. Thanks to our bug spotting champions Mike Benoit and Simon Kirby of NetNation for reporting this. __________________________________________________________ 1.3.35. Version 2.7 and 2.7.1 This version fixes some very long standing issues and adds a few new features. If you are still running 2.6, upgrade yesterday. If you were running 2.6.1, an upgrade is still strongly advised. Features: * The controlsocket is now readable and writable by the 'setgid' user. This allows for non-root access to PDNS which is nice for mrtg or cricket graphs. * MySQL backend (the non-generic one) gains the ability to read from a different table using the mysql-table setting. * pipe backend now has a configurable timeout using the pipe-timeout setting. Thanks fo Steve Bromwich for pointing out the need for this. * Experimental backtraces. If PowerDNS crashes, it will log a lot of numbers and sometimes more to the syslog. If you see these, please report them to us. Only available under Linux. Bugs: * 2.7 briefly broke the mysql backend, so don't use it if you use that. 2.7.1 fixes this. * SOA records could sometimes have the wrong TTL. Thanks to Jonas Daugaard for reporting this. * An ANY query might lead to duplicate SOA records being returned under exceptional circumstances. Thanks to Jonas Daugaard for reporting this. * Underlying the above bug, packet compression could sometimes suddenly be turned off, leading to overly large responses and non-removal of duplicate records. * The allow-axfr-ips setting did not accept IP ranges (1.2.3.0/24) which the documentation claimed it did (thanks to Florus Both of Ascio technologies for being sufficiently persistent in reporting this). * Killed backends were not being respawned, leading to suboptimal behaviour on intermittent database errors. Thanks to Steve Bromwich for reporting this. * Corrupt packets during an incoming AXFR when acting as a slave would cause a PowerDNS reload instead of just failing that AXFR. Thanks to Mike Benoit and Simon Kirby of NetNation for reporting this. * Label compression in incoming AXFR had problems with large offsets, causing the above mentioned errors. Thanks to Mike Benoit and Simon Kirby of NetNation for reporting this. __________________________________________________________ 1.3.36. Version 2.6.1 Quick fix release for a big cache problem. __________________________________________________________ 1.3.37. Version 2.6 Performance release. A lot of work has been done to raise PDNS performance to staggering levels in order to take part in benchmarketing efforts. Together with our as yet unnamed partner, PDNS has been benchmarked at 60.000 mostly cached queries/second on off the shelf PC hardware. Uncached performance was 17.000 uncached DNS queries/second on the .ORG domain. Performance has been increased by both making PDNS itself quicker but also by lowering the number of backend queries typically needed. Operators will typically see PDNS taking less CPU and the backend seeing less load. Furthermore, some real bugs were fixed. A couple of undocumented performance switches may appear in --help output but you are advised to stay away from these. Developers: this version needs the pdns-2.5.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. Performance: * A big error in latency calculations - cached packets were weighed 50 times less, leading to inflated latency reporting. Latency calculations are now correct and way lower - often in the microseconds range. * It is now possible to run with 0 second cache TTLs. This used to cause very frequent cache cleanups, leading to performance degradation. * Many tiny performance improvements, removing duplicate cache key calculations, etc. The cache itself has also been reworked to be more efficient. * First 'CNAME' backend query replaced by an 'ANY' query, which most of the time returns the actual record, preventing the need for a separate CNAME lookup, halving query load. * Much of the same for same-level-NS records on queries needing delegation. Bugs fixed: * Incidentally, the cache count would show 'unknown' packets, which was harmless but confusing. Thanks to Mike and Simon of NetNation for reporting this. * SOA hostmaster with a . in the local-part would be cached wrongly, leading to a stray backslash in case of multiple successively SOA queries. Thanks to Ascio Techologies for spotting this bug. * zone2sql did not parse Verisign zonefiles correctly as these contained a $TTL statement in mid-record. * Sometimes packets would not be accounted, leading to 'udp-queries' and 'udp-answers' divergence. Features: * 'cricket' command added to init.d scripts that provides unadorned output for parsing by 'Cricket'. __________________________________________________________ 1.3.38. Version 2.5.1 Brown paper bag release fixing a huge memory leak in the new Query Cache. Developers: this version needs the new pdns-2.5.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. And some small changes: * Added support for RF2038 compliant negative-answer caching. This allows remotes to cache the fact that a domain does not exist and will not exist for a while. Thanks to Chris Thompson for pointing out how tiny our minds are. This feature may cause a noticeable reduction in query load. * Small speedup to non-packet-cached queries, incidentally fixing the huge memory leak. * pdns_control ccounts command outputs statistics on what is in the cache, which is useful to help optimize your caching strategy. __________________________________________________________ 1.3.39. Version 2.5 An important release which has seen quite a lot of trial and error testing. As a result, PDNS can now run with a huge cache and concurrent invalidations. This is useful when running of a slower database or under high traffic load with a fast database. Furthermore, the gpgsql2 backend has been validated for use and will soon supplant the gpgsql backend entirely. This also bodes well for the gmysql backend which is the same code. Also, a large amount of issues biting large scale slave operators were addressed. Most of these issues would only show up after prolonged uptime. New features: * Query cache. The old Packet Cache only cached entire questions and their answers. This is very CPU efficient but does not lead to maximum hitrate. Two packets both needing to resolve smtp.you.com internally would not benefit from any caching. Furthermore, many different DNS queries lead to the same backend queries, like 'SOA for .COM?'. PDNS now also caches backend queries, but only those having no answer (the majority) and those having one answer (almost the rest). In tests, these additional caches appear to halve the database backend load numerically and perhaps even more in terms of CPU load. Often, queries with no answer are more expensive than those having one. The default ttls for the query-cache and negquery-cache are set to safe values (20 and 60 seconds respectively), you should be seeing an improvement in behaviour without sacrificing a lot in terms of quick updates. The webserver also displays the efficiency of the new Query Cache. The old Packet Cache is still there (and useful) but see Chapter 9 for more details. * There is now the ability to shut off some logging at a very early stage. High performance sites doing thousands of queries/second may in fact spend most of their CPU time on attempting to write out logging, even though it is ignored by syslog. The new flag log-dns-details, on by default, allows the operator to kill most informative-only logging before it takes any cpu. * Flags which can be switched 'on' and 'off' can now also be set to 'off' instead of only to 'no' to turn them off. Enhancements: * Packet Cache is now case insensitive, leading to a higher hitrate because identical queries only differing in case now both match. Care is taken to restore the proper case in the answer sent out. * Packet Cache stores packets more efficiently now, savings are estimated at 50%. * The Packet Cache is now asynchronous which means that PDNS continues to answer questions while the cache is busy being purged or queried. Incidentally this will mean a cache miss where previously the question would wait until the cache became available again. The upshot of this is that operators can call pdns_control purge as often as desired without fearing performance loss. Especially the full, non-specific, purge was speeded up tremendously. This optimization is of little merit for small sites but is very important when running with a large packetcache, such as when using recursion under high load. * AXFR log messages now all contain the word 'AXFR' to ease grepping. * Linux static version now compiled with gcc 3.2 which is known to output better and faster code than the previously used 3.0.4. Bugs fixed: * Packetcache would sometimes send packets back with slightly modified flags if these differed from the flags of the cached copy. * Resolver code did bad things with filedescriptors leading to fd exhaustion after prolonged uptimes and many slave SOA currency checks. * Resolver code failed to properly log some errors, leading to operator uncertainty regarding to AXFR problems with remote masters. * After prolonged uptime, slave code would try to use privileged ports for originating queries, leading to bad replication efficiency. * Masters sending back answers in differing case from questions would lead to bogus 'Master tried to sneak in out-of-zone data' errors and failing AXFRs. __________________________________________________________ 1.3.40. Version 2.4 Developers: this version is compatible with the pdns-2.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. This version fixes some stability issues with malformed or malcrafted packets. An upgrade is advised. Furthermore, there are interesting new features. New features: * Recursive queries are now also cached, but in a separate namespace so non-recursive queries don't get recursed answers and vice versa. This should mean way lower database load for sites running with the current default lazy-recursion. Up to now, each and every recursive query would lead to a large amount of SQL queries. To prevent the packetcache from becoming huge, a separate recursive-cache-ttl can be specified. * The ability to change parameters at runtime was added. Currently, only the new query-logging flag can be changed. * Added query-logging flag which hints a backend that it should output a textual representation of queries it receives. Currently only gmysql and gpgsql2 honor this flag. * Gmysql backend can now also talk to PgSQL, leading to less code. Currently, the old postgresql driver ('gpgsql') is still the default, the new driver is available as 'gpgsql2' and has the benefit that it does query logging. In the future, gpgsql2 will become the default gpgsql driver. * DNS recursing proxy is now more verbose in logging odd events which may be caused by buggy recursing backends. * Webserver now displays peak queries/second 1 minute average. Bugs fixed: * Failure to connect to database in master/slave communicator thread could lead to an unclean reload, fixed. Documentation: added details for strict-rfc-axfrs. This feature can be used if very old clients need to be able to do zone transfers with PDNS. Very slow. __________________________________________________________ 1.3.41. Version 2.3 Developers: this version is compatible with the pdns-2.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. This release adds the Generic MySQL backend which allows full master/slave semantics with MySQL and InnoDB tables (or other tables that support transactions). See Section A.5. Other new features: * Improved error messages in master/slave communicator will help down track problems. * slave-cycle-interval setting added. Very large sites with thousands of slave domains may need to raise this value above the default of 60. Every cycle, domains in undeterminate state are checked for their condition. Depending on the health of the masters, this may entail many SOA queries or attempted AXFRs. Bugs fixed: * 'pdns_control purge domain' and 'pdns_control purge domain$' were broken in version 2.2 and did not in fact purge the cache. There is a slight risk that domain-specific purge commands could force a reload in previous version. Thanks to Mike Benoit of NetNation for discovering this. * Master/slave communicator thread got confused in case of delayed answers from slow masters. While not causing harm, this caused inefficient behaviour when testing large amounts of slave domains because additional 'cycles' had to pass before all domains would have their status ascertained. * Backends implementing special SOA semantics (currently only the undocumented 'pdns express backend', or homegrown backends) would under some circumstances not answer the SOA record in case of an ANY query. This should put an end to the last DENIC problems. Thanks to DENIC for helping us find the problem. __________________________________________________________ 1.3.42. Version 2.2 Developers: this version is compatible with the pdns-2.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. Again a big release. PowerDNS is seeing some larger deployments in more demanding environments and these are helping shake out remaining issues, especially with recursing backends. The big news is that wildcard CNAMEs are now supported, an oft requested feature and nearly the only part in which PDNS differed from BIND in authoritative capabilities. If you were seeing signal 6 errors in PDNS causing reloads and intermittent service disruptions, please upgrade to this version. For operators of PowerDNS Express trying to host .DE domains, the very special soa-serial-offset feature has been added to placate the new DENIC requirement that the SOA serial be at least six digits. PowerDNS Express uses the SOA serial as an actual serial and not to insert dates and hence often has single digit soa serial numbers, causing big problems with .DE redelegations. Bugs fixed: * Malformed or shortened TCP recursion queries would cause a signal 6 and a reload. Same for EOF from the TCP recursing backend. Thanks to Simon Kirby and Mike Benoit of NetNation for helping debug this. * Timeouts on the TCP recursing backend were far too long, leading to possible exhaustion of TCP resolving threads. * pdns_control purge domain accidentally cleaned all packets with that name as a prefix. Thanks to Simon Kirby for spotting this. * Improved exception error logging - in some circumstances PDNS would not properly log the cause of an exception, which hampered problem resolution. New features: * Wildcard CNAMEs now work as expected! * pdns_control purge can now also purge based on suffix, allowing operators to purge an entire domain from the packet cache instead of only specific records. See also Section B.1.1 Thanks to Mike Benoit for this suggestion. * soa-serial-offset for installations with small SOA serial numbers wishing to register .DE domains with DENIC which demands six-figure SOA serial numbers. See also Chapter 15. __________________________________________________________ 1.3.43. Version 2.1 This is a somewhat bigger release due to pressing demands from customers. An upgrade is advised for installations using Recursion. If you are using recursion, it is vital that you are aware of changes in semantics. Basically, local data will now override data in your recursing backend under most circumstances. Old behaviour can be restored by turning lazy-recursion off. Developers: this version has a new pdns-2.1 development kit, available on http://downloads.powerdns.com/releases/dev. See also Appendix C. Warning Most users will run a static version of PDNS which has no dependencies on external libraries. However, some may need to run the dynamic version. This warning applies to these users. To run the dynamic version of PDNS, which is needed for backend drivers which are only available in source form, gcc 3.0 is required. RedHat 7.2 comes with gcc 3.0 as an optional component, RedHat 7.3 does not. However, the RedHat 7.2 Update gcc rpms install just fine on RedHat 7.3. For Debian, we suggest running 'woody' and installing the g++-3.0 package. We expect to release a FreeBSD dynamic version shortly. Bugs fixed: * RPM releases sometimes overwrote previous configuration files. Thanks to Jorn Ekkelenkamp of Hubris/ISP Services for reporting this. * TCP recursion sent out overly large responses due to a byteorder mistake, confusing some clients. Thanks to the capable engineers of NetNation for bringing this to our attention. * TCP recursion in combination with a recursing backend on a non-standard port did not work, leading to a non-functioning TCP listener. Thanks to the capable engineers of NetNation for bringing this to our attention. Unexpected behaviour: * Wildcard URL records where not implemented because they are a performance penalty. To turn these on, enable wildcard-url in the configuration. * Unlike other nameservers, local data did not override the internet for recursing queries. This has mostly been brought into conformance with user expectations. If a recursive question can be answered entirely from local data, it is. To restore old behaviour, disable lazy-recursion. Also see Chapter 11. Features: * Oracle support has been tuned, leading to the first public release of the Oracle backend. Zone2sql now outputs better SQL and the backend is now fully documented. Furthermore, the queries are compatible with the PowerDNS XML-RPC product, allowing PowerDNS express to run off Oracle. See Section A.6. * Zone2sql now accepts --transactions to wrap zones in a transaction for PostgreSQL and Oracle output. This is a major speedup and also makes for better isolation of inserts. See Section 10.1. * pdns_control now has the ability to purge the PowerDNS cache or parts of it. This enables operators to raise the TTL of the Packet Cache to huge values and only to invalidate the cache when changes are made. See also Chapter 9 and Section B.1.1. __________________________________________________________ 1.3.44. Version 2.0.1 Maintenance release, fixing three small issues. Developers: this version is compatible with 1.99.11 backends. * PowerDNS ignored the logging-facility setting unless it was specified on the commandline. Thanks to Karl Obermayer from WebMachine Technologies for noticing this. * Zone2sql neglected to preserve 'slaveness' of domains when converting to the slave capable PostgreSQL backend. Thanks to Mike Benoit of NetNation for reporting this. Zone2sql now has a --slave option. * SOA Hostmaster addresses with dots in them before the @-sign were mis-encoded on the wire. __________________________________________________________ 1.3.45. Version 2.0 Two bugfixes, one stability/security related. No new features. Developers: this version is compatible with 1.99.11 backends. Bugfixes: * zone2sql refused to work under some circumstances, taking 100% cpu and not functioning. Thanks to Andrew Clark and Mike Benoit for reporting this. * Fixed a stability issue where malformed packets could force PDNS to reload. Present in all earlier 2.0 versions. __________________________________________________________ 1.3.46. Version 2.0 Release Candidate 2 Mostly bugfixes, no really new features. Developers: this version is compatible with 1.99.11 backends. Bugs fixed: * chroot() works again - 2.0rc1 silently refused to chroot. Thanks to Hub Dohmen for noticing this. * setuid() and setgid() security features were silently not being performed in 2.0rc1. Thanks to Hub Dohmen for noticing this. * MX preferences over 255 now work as intended. Thanks to Jeff Crowe for noticing this. * IPv6 clients can now also benefit from the recursing backend feature. Thanks to Andy Furnell for proving beyond any doubt that this did not work. * Extremely bogus code removed from DNS notification reception code - please test! Thanks to Jakub Jermar for working with us in figuring out just how broken this was. * AXFR code improved to handle more of the myriad different zonetransfer dialects available. Specifically, interoperability with Bind 4 was improved, as well as Bind 8 in 'strict rfc conformance' mode. Thanks again for Jakub Jermar for running many tests for us. If your transfers failed with 'Unknown type 14!!' or words to that effect, this was it. Features: * Win32 version now has a zone2sql tool. * Win32 version now has support for specifying how urgent messages should be before they go to the NT event log. Remaining issues: * One persistent report of the default 'chroot=./' configuration not working. * One report of disable-axfr and allow-axfr-ips not working as intended. * Support for relative paths in zones and in Bind configuration is not bug-for-bug compatible with bind yet. __________________________________________________________ 1.3.47. Version 2.0 Release Candidate 1 The MacOS X release! A very experimental OS X 10.2 build has been added. Furthermore, the Windows version is now in line with Unix with respect to capabilities. The ODBC backend now has the code to function as both a master and a slave. Developers: this version is compatible with 1.99.11 backends. * Implemented native packet response parsing code, allowing Windows to perform AXFR and NS and SOA queries. * This is the first version for which we have added support for Darwin 6.0, which is part of the forthcoming Mac OS X 10.2. Please note that although this version is marked RC1, that we have not done extensive testing yet. Consider this a technology preview. + The Darwin version has been developed on Mac OS X 10.2 (635). Other versions may or may not work. + Currently only the random, bind, mysql and pdns backends are included. + The menu based installer script does not work, you will have to edit pathconfig by hand as outlined in chapter 2. + On Mac OS X Client, PDNS will fail to start because a system service is already bound to port 53. This version is distributed as a compressed tar file. You should follow the generic UNIX installation instructions. Bugs fixed: * Zone2sql PostgreSQL mode neglected to lowercase $ORIGIN. Thanks to Maikel Verheijen of Ladot for spotting this. * Zone2sql PostgreSQL mode neglected to remove a trailing dot from $ORIGIN if present. Thanks to Thanks to Maikel Verheijen of Ladot for spotting this. * Zonefile parser was not compatible with bind when $INCLUDING non-absolute filenames. Thanks to Jeff Miller for working out how this should work. * Bind configuration parser was not compatible with bind when including non-absolute filenames. Thanks to Jeff Miller for working out how this should work. * Documentation incorrectly listed the Bind backend as 'slave capable'. This is not yet true, now labeled 'experimental'. Windows changes. We are indebted to Dimitry Andric who educated us in the ways of distributing Windows software. * pdns.conf is now read if available. * Console version responds to ^c now. * Default pdns.conf added to distribution * Uninstaller missed several files, leaving remnants behind * DLLs are now installed locally, with the pdns executable. * pdns_control is now also available on Windows * ODBC backend can now act as master and slave. Experimental. * The example zone missed indexes and had other faults. * A runtime DLL that is present on most windows systems (but not all!) was missing. __________________________________________________________ 1.3.48. Version 1.99.12 Prerelease The Windows release! See Chapter 3. Beware, windows support is still very fresh and untested. Feedback is very welcome. Developers: this version is compatible with 1.99.11 backends. * Windows 2000 codebase merge completed. This resulted in quite some changes on the Unix end of things, so this may impact reliability * ODBC backend added for Windows. See Section A.10. * IBM DB2 Universal Database backend available for Linux. See Section A.8. * Zone2sql now understands $INCLUDE. Thanks to Amaze Internet for nagging about this * The SOA Mininum TTL now has a configurable default (soa-minimum-ttl)value to placate the DENIC requirements. * Added a limit on the simultaneous numbers of TCP connections to accept (max-tcp-connections). Defaults to 10. Bugs fixed: * When operating in virtual hosting mode (See Chapter 8), the additional init.d scripts would not function correctly and interface with other pdns instances. * PDNS neglected to conserve case on answers. So a query for WwW.PoWeRdNs.CoM would get an answer listing the address of www.powerdns.com. While this did not confuse resolvers, it is better to conserve case. This has semantical concequences for all backends, which the documentation now spells out. * PostgreSQL backend was case sensitive and returned only answers in case an exact match was found. The Generic PostgreSQL backend is now officially all lower case and zone2sql in PostgreSQL mode enforces this. Documentation has been been updated to reflect the case change. Thanks to Maikel Verheijen of Ladot for spotting this! * Documentation bug - postgresql create/index statements created a duplicate index. If you've previously copy pasted the commands and not noticed the error, execute CREATE INDEX rec_name_index ON records(name) to remedy. Thanks to Jeff Miller for reporting this. This also lead to depressingly slow 'ANY' lookups for those of you doing benchmarks. Features: * pdns_control (see Section B.1.1) now opens the local end of its socket in /tmp instead of next to the remote socket (by default /var/run). This eases the way for allowing non-root access to pdns_control. When running chrooted (see Chapter 7), the local socket again moves back to /var/run. * pdns_control now has a 'version' command. See Section B.1.1. __________________________________________________________ 1.3.49. Version 1.99.11 Prerelease This release is important because it is the first release which is accompanied by an Open Source Backend Development Kit, allowing external developers to write backends for PDNS. Furthermore, a few bugs have been fixed: * Lines with only whitespace in zone files confused PDNS (thanks Henk Wevers) * PDNS did not properly parse TTLs with symbolic sufixes in zone files, ie 2H instead of 7200 (thanks Henk Wevers) __________________________________________________________ 1.3.50. Version 1.99.10 Prerelease IMPORTANT: there has been a tiny license change involving free public webbased dns hosting, check out the changes before deploying! PDNS is now feature complete, or very nearly so. Besides adding features, a lot of 'fleshing out' work is done now. There is an important performance bug fix which may have lead to disappointing benchmarks - so if you saw any of that, please try either this version or 1.99.8 which also does not have the bug. This version has been very stable for us on multiple hosts, as was 1.99.9. PostgreSQL users should be aware that while 1.99.10 works with the schema as presented in earlier versions, advanced features such as master or slave support will not work unless you create the new 'domains' table as well. Bugs fixed: * Wildcard AAAA queries sometimes received an NXDOMAIN error where they should have gotten an empty NO ERROR. Thanks to Jeroen Massar for spotting this on the .TK TLD! * Do not disable the packetcache for 'recursion desired' packets unless a recursor was configured. Thanks to Greg Schueler for noticing this. * A failing backend would not be reinstated. Thanks to 'Webspider' for discovering this problem with PostgreSQL connections that die after prolonged inactivity. * Fixed loads of IPv6 transport problems. Thanks to Marco Davids and others for testing. Considered ready for production now. * Zone2sql printed a debugging statement on range $GENERATE commands. Thanks to Rene van Valkenburg for spotting this. Features: * PDNS can now act as a master, sending out notifications in case of changes and allowing slaves to AXFR. Big rewording of replication support, domains are now either 'native', 'master' or 'slave'. See Chapter 13 for lots of details. * Zone2sql in PostgreSQL mode now populates the 'domains' table for easy master, slave or native replication support. * Ability to disable those annoying Windows DNS Dynamic Update messages from appearing in the log. See log-failed-updates in Chapter 15. * Ability to run on IPv6 transport only * Logging can now happen under a 'facility' so all PDNS messages appear in their own file. See Section 6.3. * Different OS releases of PDNS now get different install path defaults. Thanks to Mark Lastdrager for nagging about this and to Nero Imhard and Frederique Rijsdijk for suggesting saner defaults. * Infrastructure for 'also-notify' statements added. __________________________________________________________ 1.3.51. Version 1.99.9 Early Access Prerelease This is again a feature and an infrastructure release. We are nearly feature complete and will soon start work on the backends to make sure that they are all master, slave and 'superslave' capable. Bugs fixed: * PDNS sometimes sent out duplicate replies for packets passed to the recursing backend. Mostly a problem on SMP systems. Thanks to Mike Benoit for noticing this. * Out-of-bailiwick CNAMES (ie, a CNAME to a domain not in PDNS) caused a 'ServFail' packet in 1.99.8, indicating failure, leading to hosts not resolving. Thanks to Martin Gillstrom for noticing this. * Zone2sql balked at zones editted under operating sytems terminating files with ^Z (Windows). Thanks Brian Willcott for reporting this. * PostgreSQL backend logged the password used to connect. Now only does so in case of failure to connect. Thanks to 'Webspider' for noticing this. * Debian unstable distribution wrongly depended on home compiled PostgreSQL libraries. Thanks to Konrad Wojas for noticing this. Features: * When operating as a slave, AAAA records are now supported in the zone. They were already supported in master zones. * IPv6 transport support - PDNS can now listen on an IPv6 socket using the local-ipv6 setting. * Very silly randombackend added which appears in the documentation as a sample backend. See Appendix C. * When transferring a slave zone from a master, out of zone data is now rejected. Malicious operators might try to insert bad records otherwise. * 'Supermaster' support for automatic provisioning from masters. See Section 13.2.1. * Recursing backend can now live on a non-standard (!=53) port. See Chapter 11. * Slave zone retrieval is now queued instead of immediate, which scales better and is more resilient to temporary failures. * max-queue-length parameter. If this many packets are queued for database attention, consider the situation hopeless and respawn. Internal: * SOA records are now 'special' and each backend can optionally generate them in special ways. PostgreSQL backend does so when operating as a slave. * Writing backends is now a lot easier. See Appendix C. * Added Bindbackend to internal regression tests, confirming that it is compliant. __________________________________________________________ 1.3.52. Version 1.99.8 Early Access Prerelease A lot of infrastructure work gearing up to 2.0. Some stability bugs fixed and a lot of new features. Bugs fixed: * Bindbackend was overly complex and crashed on some systems on startup. Simplified launch code. * SOA fields were not always properly filled in, causing default values to go out on the wire * Obscure bug triggered by malicious packets (we know who you are) in SOA finding code fixed. * Magic serial number calculation contained a double free leading to instability. * Standards violation, questions for domains for which PDNS was unauthoritative now get a SERVFAIL answer. Thanks to the IETF Namedroppers list for helping out with this. * Slowly launching backends were being relaunched at a great rate when queries were coming in while launching backends. * MySQL-on-unix-domain-socket on SMP systems was overwhelmed by the quick connection rate on launch, inserted a small 50ms delay. * Some SMP problems appear to be compiler related. Shifted to GCC 3.0.4 for Linux. * Ran ispell on documentation. Feature enhancements: * Recursing backend. See Chapter 11. Allows recursive and authoritative DNS on the same IP address. * NAPTR support, which is especially useful for the ENUM/E.164 community. * Zone transfers can now be allowed per netmask instead of only per IP address. * Preliminary support for slave operation included. Only for the adventurous right now! See Section 13.2 * All record types now documented, see Chapter 17. __________________________________________________________ 1.3.52.1. Known bugs Wildcard CNAMES do not work as they do with bind. Recursion sometimes sends out duplicate packets (fixed in 1.99.9 snapshots) Some stability issues which are caught by the guardian __________________________________________________________ 1.3.52.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend __________________________________________________________ 1.3.53. Version 1.99.7 Early Access Prerelease Named.conf parsing got a lot of work and many more bind configurations can now be parsed. Furthermore, error reporting was improved. Stability is looking good. Bugs fixed: * Bind parser got confused by filenames with underscores and colons. * Bind parser got confused by spaces in quoted names * FreeBSD version now stops and starts when instructed to do so. * Wildcards were off by default, which violates standards. Now on by default. * --oracle was broken in zone2sql Feature enhancements: * Line number counting goes on as it should when including files in named.conf * Added --no-config to enable users to start the pdns daemon without parsing the configuration file. * zone2sql now has --bare for unformatted output which can be used to generate insert statements for different database layouts * zone2sql now has --gpgsql, which is an alias for --mysql, to output in a format useful for the default Generic PgSQL backend * zone2sql is now documented. __________________________________________________________ 1.3.53.1. Known bugs Wildcard CNAMES do not work as they do with bind. __________________________________________________________ 1.3.53.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend Some of these features will be present in newer releases. __________________________________________________________ 1.3.54. Version 1.99.6 Early Access Prerelease This version is now running on dns-eu1.powerdns.net and working very well for us. But please remain cautious before deploying! Bugs fixed: * Webserver neglected to show log messages * TCP question/answer miscounted multiple questions over one socket. Fixed misnaming of counter * Packetcache now detects clock skew and times out entries * named.conf parser now reports errors with line number and offending token * Filenames in named.conf can now contain : Feature enhancements: * The webserver now by default does not print out configuration statements, which might contain database backends. Use webserver-print-arguments to restore the old behaviour. * Generic PostgreSQL backend is now included. Still rather beta. __________________________________________________________ 1.3.54.1. Known bugs FreeBSD version does not stop when requested to do so. Wildcard CNAMES do not work as they do with bind. __________________________________________________________ 1.3.54.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend Some of these features will be present in newer releases. __________________________________________________________ 1.3.55. Version 1.99.5 Early Access Prerelease The main focus of this release is stability and TCP improvements. This is the first release PowerDNS-the-company actually considers for running on its production servers! Major bugs fixed: * Zone2sql received a floating point division by zero error on named.confs with less than 100 domains. * Huffman encoder failed without specific error on illegal characters in a domain * Fixed huge memory leaks in TCP code. * Removed further file descriptor leaks in guardian respawning code * Pipebackend was too chatty. * pdns_server neglected to close fds 0, 1 & 2 when daemonizing Feature enhancements: * bindbackend can be instructed not to check the ctime of a zone by specifying bind-check-interval=0, which is also the new default. * pdns_server --list-modules lists all available modules. Performance enhancements: * TCP code now only creates a new database connection for AXFR. * TCP connections timeout rather quickly now, leading to less load on the server. __________________________________________________________ 1.3.55.1. Known bugs FreeBSD version does not stop when requested to do so. Wildcard CNAMES do not work as they do with bind. __________________________________________________________ 1.3.55.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend, gpgsqlbackend Some of these features will be present in newer releases. __________________________________________________________ 1.3.56. Version 1.99.4 Early Access Prerelease A lot of new named.confs can now be parsed, zone2sql & bindbackend have gained features and stability. Major bugs fixed: * Label compression was not always enabled, leading to large reply packets sometimes. * Database errors on TCP server lead to a nameserver reload by the guardian. * MySQL backend neglected to close its connection properly. * BindParser miss parsed some IP addresses and netmasks. * Truncated answers were also truncated on the packetcache, leading to truncated TCP answers. Feature enhancements: * Zone2sql and the bindbackend now understand the Bind $GENERATE{} syntax. * Zone2sql can optionally gloss over non-existing zones with --on-error-resume-next. * Zone2sql and the bindbackend now properly expand @ also on the right hand side of records. * Zone2sql now sets a default TTL. * DNS UPDATEs and NOTIFYs are now logged properly and sent the right responses. Performance enhancements: * 'Fancy records' are no longer queried for on ANY queries - this is a big speedup. __________________________________________________________ 1.3.56.1. Known bugs FreeBSD version does not stop when requested to do so. Zone2sql refuses named.confs with less than 100 domains. Wildcard CNAMES do not work as they do with bind. __________________________________________________________ 1.3.56.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend, gpgsqlbackend Some of these features will be present in newer releases. __________________________________________________________ 1.3.57. Version 1.99.3 Early Access Prerelease The big news in this release is the BindBackend which is now capable of parsing many more named.conf Bind configurations. Furthermore, PDNS has successfully parsed very large named.confs with large numbers of small domains, as well as small numbers of large domains (TLD). Zone transfers are now also much improved. Major bugs fixed: * zone2sql leaked file descriptors on each domain, used wrong Bison recursion leading to parser stack overflows. This limited the amount of domains that could be parsed to 1024. * zone2sql can now read all known zonefiles, with the exception of those containing $GENERATE * Guardian relaunching a child lost two file descriptors * Don't die on a connection reset by peer during zone transfer. * Webserver does not crash anymore on ringbuffer resize Feature enhancements: * AXFR can now be disabled, and re-enabled per IP address * --help accepts a parameter, will then show only help items with that prefix. * zone2sql now accepts a --zone-name parameter * BindBackend maturing - 9500 zones parsed in 3.5 seconds. No longer case sensitive. Performance enhancements: * Implemented RFC-breaking AXFR format (which is the industry standard). Zone transfers now zoom along at wirespeed (many megabits/s). __________________________________________________________ 1.3.57.1. Known bugs FreeBSD version does not stop when requested to do so. BindBackend cannot parse zones with $GENERATE statements. __________________________________________________________ 1.3.57.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend, gpgsqlbackend Some of these features will be present in newer releases. __________________________________________________________ 1.3.58. Version 1.99.2 Early Access Prerelease Major bugs fixed: * Database backend reload does not hang the daemon anymore * Buffer overrun in local socket address initialisation may have caused binding problems * setuid changed the uid to the gid of the selected user * zone2sql doesn't coredump on invocation anymore. Fixed lots of small issues. * Don't parse configuration file when creating configuration file. This was a problem with reinstalling. Performance improvements: * removed a lot of unnecessary gettimeofday calls * removed needless select(2) call in case of listening on only one address * removed 3 useless syscalls in the fast path Having said that, more work may need to be done. Testing on a 486 saw packet rates in a simple setup (question/wait/answer/question..) improve from 200 queries/second to over 400. Usability improvements: * Fixed error checking in init.d script (show, mrtg) * Added 'uptime' to the mrtg output * removed further GNUisms from installer and init.d scripts for use on FreeBSD * Debian package and apt repository, thanks to Wichert Akkerman. * FreeBSD /usr/ports, thanks to Peter van Dijk (in progress). Stability may be an issue as well as performance. This version has a tendency to log a bit too much which slows the nameserver down a lot. __________________________________________________________ 1.3.58.1. Known bugs Decreasing a ringbuffer on the website is a sure way to crash the daemon. Zone2sql, while improved, still has problems with a zone in the following format: name IN A 1.2.3.4 IN A 1.2.3.5 To fix, add 'name' to the second line. Zone2sql does not close filedescriptors. FreeBSD version does not stop when requested via the init.d script. __________________________________________________________ 1.3.58.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend, gpgsqlbackend * fully functioning bindbackend - will try to parse named.conf, but probably fail Some of these features will be present in newer releases. __________________________________________________________ 1.3.59. Version 1.99.1 Early Access Prerelease This is the first public release of what is going to become PDNS 2.0. As such, it is not of production quality. Even PowerDNS-the-company does not run this yet. Stability may be an issue as well as performance. This version has a tendency to log a bit too much which slows the nameserver down a lot. __________________________________________________________ 1.3.59.1. Known bugs Decreasing a ringbuffer on the website is a sure way to crash the daemon. Zone2sql is very buggy. __________________________________________________________ 1.3.59.2. Missing features Features present in this document, but disabled or withheld from the current release: * gmysqlbackend, oraclebackend, gpgsqlbackend * fully functioning bindbackend - will not parse configuration files Some of these features will be present in newer releases. __________________________________________________________ 1.4. Security If you have a security problem to report, please email us at both and at . We adhere to the Rain Forest Puppy Full Disclosure Policy (RFPolicy) v2.0 and we ask you to do the same. In particular, please do not mail security issues to public lists unless we do not get back to you in a timely manner. We remind PowerDNS users that under the terms of the GNU General Public License, PowerDNS comes with ABSOLUTELY NO WARRANTY. This license is included in the distribution and in this documentation, see Appendix E. As of the 6th of August 2008, no actual security problems with PowerDNS 2.9.21.1, Recursor 3.1.5, or later are known about. This page will be updated with all bugs which are deemed to be security problems, or could conceivably lead to those. Any such notifications will also be sent to all PowerDNS mailinglists. Version 3.1.4 and earlier of the PowerDNS recursor were vulnerable to a spoofing attack. For more detail, see Section 1.7. Version 3.1.3 and earlier of the PowerDNS recursor contain two security issues, both of which can lead to a denial of service, both of which can be triggered by remote users. One of the issues might lead be exploited and lead to a system compromise. For more detail, see Section 1.5 and Section 1.6. Version 3.0 of the PowerDNS recursor contains a denial of service bug which can be exploited remotely. This bug, which we believe to only lead to a crash, has been fixed in 3.0.1. There are no guarantees however, so an upgrade from 3.0 is highly recommended. All versions of PowerDNS before 2.9.21.1 do not respond to certain queries. This in itself is not a problem, but since the discovery by Dan Kaminsky of a new spoofing technique, this silence for queries PowerDNS considers invalid, within a valid domain, allows attackers more chances to feed *other* resolvers bad data. All versions of PowerDNS before 2.9.18 contain the following two bugs, which only apply to installations running with the LDAP backend, or installations providing recursion to a limited range of IP addresses. If any of these apply to you, an upgrade is highly advised: * The LDAP backend did not properly escape all queries, allowing it to fail and not answer questions. We have not investigated further risks involved, but we advise LDAP users to update as quickly as possible (Norbert Sendetzky, Jan de Groot) * Questions from clients denied recursion could blank out answers to clients who are allowed recursion services, temporarily. Reported by Wilco Baan. This would've made it possible for outsiders to blank out a domain temporarily to your users. Luckily PowerDNS would send out SERVFAIL or Refused, and not a denial of a domain's existence. All versions of PowerDNS before 2.9.17 are known to suffer from remote denial of service problems which can disrupt operation. Please upgrade to 2.9.17 as this page will only contain detailed security information from 2.9.17 onwards. __________________________________________________________ 1.5. PowerDNS Security Advisory 2006-01: Malformed TCP queries can lead to a buffer overflow which might be exploitable Table 1-1. PowerDNS Security Advisory CVE CVE-2006-4251 Date 13th of November 2006 Affects PowerDNS Recursor versions 3.1.3 and earlier, on all operating systems. Not affected No versions of the PowerDNS Authoritative Server ('pdns_server') are affected. Severity Critical Impact Potential remote system compromise. Exploit As far as we know, no exploit is available as of 11th of November 2006. Solution Upgrade to PowerDNS Recursor 3.1.4, or apply the patches referred below and recompile Workaround Disable TCP access to the Recursor. This will have slight operational impact, but it is likely that this will not lead to meaningful degradation of service. Disabling access is best performed at packet level, either by configuring a firewall, or instructing the host operating system to drop TCP connections to port 53. Additionally, exposure can be limited by configuring the allow-from setting so only trusted users can query your nameserver. PowerDNS Recursor 3.1.3 and previous miscalculate the length of incoming TCP DNS queries, and will attempt to read up to 4 gigabytes of query into a 65535 byte buffer. We have not verified if this problem might actually lead to a system compromise, but are acting on the assumption that it might. For distributors, a minimal patch is available on the PowerDNS wiki. Additionally, those shipping very old versions of the PowerDNS Recursor might benefit from this patch. The impact of these and other security problems can be lessened by considering the advice in Chapter 7. __________________________________________________________ 1.6. PowerDNS Security Advisory 2006-02: Zero second CNAME TTLs can make PowerDNS exhaust allocated stack space, and crash Table 1-2. PowerDNS Security Advisory CVE CVE-2006-4252 Date 13th of November 2006 Affects PowerDNS Recursor versions 3.1.3 and earlier, on all operating systems. Not affected No versions of the PowerDNS Authoritative Server ('pdns_server') are affected. Severity Moderate Impact Denial of service Exploit This problem can be triggered by sending queries for specifically configured domains Solution Upgrade to PowerDNS Recursor 3.1.4, or apply commit 919. Workaround None known. Exposure can be limited by configuring the allow-from setting so only trusted users can query your nameserver. PowerDNS would recurse endlessly on encountering a CNAME loop consisting entirely of zero second CNAME records, eventually exceeding resources and crashing. __________________________________________________________ 1.7. PowerDNS Security Advisory 2008-01: System random generator can be predicted, leading to the potential to 'spoof' PowerDNS Recursor Table 1-3. PowerDNS Security Advisory CVE Not yet assigned Date 31st of March 2008 Affects PowerDNS Recursor versions 3.1.4 and earlier, on most operating systems Not affected No versions of the PowerDNS Authoritative Server ('pdns_server') are affected. Severity Moderate Impact Data manipulation; client redirection Exploit This problem can be triggered by sending queries for specifically configured domains, sending spoofed answer packets immediately afterwards. Solution Upgrade to PowerDNS Recursor 3.1.5, or apply changesets 1159, 1160 and 1164. Workaround None known. Exposure can be limited by configuring the allow-from setting so only trusted users can query your nameserver. We would like to thank Amit Klein of Trusteer for bringing a serious vulnerability to our attention which would enable a smart attacker to 'spoof' previous versions of the PowerDNS Recursor into accepting possibly mallicious data. Details can be found on this Trusteer page. This security problem was announced in this email message. It is recommended that all users of the PowerDNS Recursor upgrade to 3.1.5 as soon as practicable, while we simultaneously note that busy servers are less susceptible to the attack, but not immune. The vulnerability is present on all operating systems where the behaviour of the libc random() function can be predicted based on its past output. This includes at least all known versions of Linux, as well as Microsoft Windows, and probably FreeBSD and Solaris. The magnitude of this vulnerability depends on internal details of the system random() generator. For Linux, the mathematics of the random generator are complex, but well understood and Amit Klein has written and published a proof of concept that can succesfully predict its output after uninterrupted observation of 40-50 DNS queries. Because the observation needs to be uninterrupted, busy PowerDNS Recursor instances are harder to subvert - other data is highly likely to be interleaved with traffic generated by an attacker. Nevertheless, operators are urged to update at their earliest convenience. __________________________________________________________ 1.8. PowerDNS Security Advisory 2008-02: By not responding to certain queries, domains become easier to spoof Table 1-4. PowerDNS Security Advisory CVE CVE-2008-3337 Date 6th of August 2008 Affects PowerDNS Authoritative Server 2.9.21 and earlier Not affected No versions of the PowerDNS Recursor ('pdns_recursor') are affected. Severity Moderate Impact Data manipulation; client redirection Exploit Domains with servers that drop certain queries can be spoofed using simpler measures than would usually be required Solution Upgrade to PowerDNS Authoritative Server 2.9.21.1, or apply commit 1239. Workaround None known. Brian J. Dowling of Simplicity Communications has discovered a security implication of the previous PowerDNS behaviour to drop queries it considers malformed. We are grateful that Brian notified us quickly about this problem. The implication is that while the PowerDNS Authoritative server itself does not face a security risk because of dropping these malformed queries, other resolving nameservers run a higher risk of accepting spoofed answers for domains being hosted by PowerDNS Authoritative Servers before 2.9.21.1. While the dropping of queries does not aid sophisticated spoofing attempts, it does facilitate simpler attacks. __________________________________________________________ 1.9. PowerDNS Security Advisory 2008-02: Some PowerDNS Configurations can be forced to restart remotely Table 1-5. PowerDNS Security Advisory CVE Not yet assigned Date 18th of November 2008 Affects PowerDNS Authoritative Server 2.9.21.1 and earlier Not affected No versions of the PowerDNS Recursor ('pdns_recursor') are affected. Versions not running in single threaded mode ('distributor-threads=1') are probably not affected. Severity Moderate Impact Denial of Service Exploit Send PowerDNS an CH HINFO query. Solution Upgrade to PowerDNS Authoritative Server 2.9.21.2, or wait for 2.9.22. Workaround Remove 'distributor-threads=1' if this is set. Daniel Drown discovered that his PowerDNS 2.9.21.1 installation crashed on receiving a HINFO CH query. In his enthousiasm, he shared his discovery with the world, forcing a rapid over the weekend release cycle. While we thank Daniel for his discovery, please study our security policy as outlined in Section 1.4 before making vulnerabilities public. It is believed that this issue only impacts PowerDNS Authoritative Servers operating with 'distributor-threads=1', but even on other configurations a database reconnect occurs on receiving a CH HINFO query. __________________________________________________________ 1.10. Acknowledgements PowerDNS is grateful for the help of the following people or institutions: * Dave Aaldering * Wichert Akkerman * Antony Antony * Mike Benoit (NetNation Communication Inc) * Peter van Dijk * Koos van den Hout * Andre Koopal * Eric Veldhuyzen * Paul Wouters * Thomas Wouters * IETF Namedroppers mailinglist Thanks! (these people don't share the blame for any errors or mistakes in powerdns - those are all ours) __________________________________________________________ Chapter 2. Installing on Unix You will typically install PDNS > 2.9 via source or via a package. Earlier versions used a clumsy binary installer. __________________________________________________________ 2.1. Possible problems at this point At this point some things may have gone wrong. Typical errors include: error while loading shared libraries: libstdc++.so.x: cannot open shared object file: No such file or directory Errors looking like this indicate a mismatch between your PDNS distribution and your Unix operating system. Download the static PDNS distribution for your operating system and try again. Please contact if this is impractical. __________________________________________________________ 2.2. Testing your install After installing, it is a good idea to test the basic functionality of the software before configuring database backends. For this purpose, PowerDNS contains the 'bindbackend' which has a domain built in example.com, which is officially reserved for testing. To test, edit pdns.conf and add the following if not already present: launch=bind bind-example-zones Warning As of 2.9.21, the BIND backend no longer features the 'bind-example-zones' command. These will return in 2.9.22. This configures powerdns to 'launch' the bindbackend, and enable the example zones. To fire up PDNS in testing mode, execute: /etc/init.d/pdns monitor, where you may have to substitute the location of your SysV init.d location you specified earlier. In monitor mode, the pdns process runs in the foreground and is very verbose, which is perfect for testing your install. If everything went all right, you can query the example.com domain like this: host www.example.com 127.0.0.1 www.example.com should now have IP address 1.2.3.4. The host command can usually be found in the dnsutils package of your operating system. Alternate command is: dig www.example.com A @127.0.0.1 or even nslookup www.example.com 127.0.0.1, although nslookup is not advised for DNS diagnostics. * example.com SOA record * example.com NS record pointing to ns1.example.com * example.com NS record pointing to ns2.example.com * example.com MX record pointing to mail.example.com * example.com MX record pointing to mail1.example.com * mail.example.com A record pointing to 4.3.2.1 * mail1.example.com A record pointing to 5.4.3.2 * ns1.example.com A record pointing to 4.3.2.1 * ns2.example.com A record pointing to 5.4.3.2 * host-0 to host-9999.example.com A record pointing to 2.3.4.5 When satisfied that basic functionality is there, type QUIT to exit the monitor mode. The adventurous may also type SHOW * to see some internal statistics. In case of problems, you will want to read the following section. __________________________________________________________ 2.2.1. Typical errors At this point some things may have gone wrong. Typical errors include: binding to UDP socket: Address already in use This means that another nameserver is listening on port 53 already. You can resolve this problem by determining if it is safe to shutdown the nameserver already present, and doing so. If uncertain, it is also possible to run PDNS on another port. To do so, add local-port=5300 to pdns.conf, and try again. This however implies that you can only test your nameserver as clients expect the nameserver to live on port 53. binding to UDP socket: Permission denied You must be superuser in order to be able to bind to port 53. If this is not a possibility, it is also possible to run PDNS on another port. To do so, add local-port=5300 to pdns.conf, and try again. This however implies that you can only test your nameserver as clients expect the nameserver to live on port 53. Unable to launch, no backends configured for querying PDNS did not find the launch=bind instruction in pdns.conf. Multiple IP addresses on your server, PDNS sending out answers on the wrong one, Massive amounts of 'recvfrom gave error, ignoring: Connection refused' If you have multiple IP addresses on the internet on one machine, UNIX often sends out answers over another interface than which the packet came in on. In such cases, use local-address to bind to specific IP addresses, which can be comma separated. The second error comes from remotes disregarding answers to questions it didn't ask to that IP address and sending back ICMP errors. __________________________________________________________ 2.3. Running PDNS on unix PDNS is normally controlled via a SysV-style init.d script, often located in /etc/init.d or /etc/rc.d/init.d. This script accepts the following commands: monitor Monitor is a special way to view the daemon. It executes PDNS in the foreground with a lot of logging turned on, which helps in determining startup problems. Besides running in the foreground, the raw PDNS control socket is made available. All external communication with the daemon is normally sent over this socket. While useful, the control console is not an officially supported feature. Commands which work are: QUIT, SHOW *, SHOW varname, RPING. start Start PDNS in the background. Launches the daemon but makes no special effort to determine success, as making database connections may take a while. Use status to query success. You can safely run start many times, it will not start additional PDNS instances. restart Restarts PDNS if it was running, starts it otherwise. status Query PDNS for status. This can be used to figure out if a launch was successful. The status found is prefixed by the PID of the main PDNS process. stop Requests that PDNS stop. Again, does not confirm success. Success can be ascertained with the status command. dump Dumps a lot of statistics of a running PDNS daemon. It is also possible to single out specific variable by using the show command. show variable Show a single statistic, as present in the output of the dump. mrtg See the performance monitoring Chapter 6. __________________________________________________________ Chapter 3. Installing on Microsoft Windows Note PowerDNS support for Windows is, as of 1.99.12, very recent and therefore quite 'beta'. For reliability, we currently advise the use of the Unix versions. Furthermore there is no support for master or slave operation in the ODBC backend, which is the only one provided currently. As of 1.99.12, PowerDNS supports Windows natively. PDNS can act as an NT service and works with any ODBC drivers you may have. To install PowerDNS for Windows you should check if your PC meets the following requirements: * A PC running Microsoft NT (with a recent servicepack and at least mdac 2.5), 2000 or XP. * An ODBC source containing valid zone information (an example MS Access database is supplied in the form of powerdns.mdb). After installing the software you should create a valid ODBC source. To do this you have open the ODBC sources dialog: Start->Settings->Control Panel->Administrative Tools->Data Sources (ODBC). We'll use the example zone database that is included in the installation to explain how to create a source. When you are in the ODBC sources dialog you activate the System DSN tab. Note It is important to create a System DSN instead of an User DNS, otherwise the ODBC backend cannot function. Press Add..., then you have to select a driver. Select Microsoft Access Driver (*.mdb). Use PowerDNS as the DSN name, you can leave the description empty. Then press Select... to select the database (ie. C:\Program Files\PowerDNS\powerdns.mdb). Press Ok and you should be done. For more information, see Section A.10. __________________________________________________________ 3.1. Configuring PDNS on Microsoft Windows You can specify program parameters in the pdns.conf file which should be located in pdns directory (ie. C:\Program Files\PowerDNS\). To see a list of available parameters you can run pdns.exe --help. Note A default configuration file has been supplied with the installation. __________________________________________________________ 3.2. Running PDNS on Microsoft Windows If you installed pdns on Windows NT, 2000 or XP you can run pdns as a service. This is how to do it: Go to services (Start->Settings->Control Panel->Administrative Tools->Services) and locate PDNS (you should have registered the program as a NT service during the installation). Double-click on PDNS and push the start button. You should now see a progress bar that gets to the end and see the status change to 'Started'. This is the same as starting pdns like this: pdns.exe --ntservice If you haven't registered pdns as a service during the installation you can do so from the commandline by starting pdns like this: pdns.exe --register-service You can run pdns as a standard console program by using a command prompt or Start->Run... This way you can specify command-line parameters (see the documentation for commandline options). If you chose to add a PowerDNS menu to the start menu during the installation you can start pdns using the pdns shortcut in that menu. __________________________________________________________ Chapter 4. Basic setup: configuring database connectivity This chapter shows you how to configure the Generic MySQL backend, which we like a lot. But feel free to use any of the myriad other backends. This backend is called 'gmysql', and needs to be configured in pdns.conf. Add the following lines, adjusted for your local setup: launch=gmysql gmysql-host=127.0.0.1 gmysql-user=root gmysql-dbname=pdnstest Remove any earlier launch statements. Also remove the bind-example-zones statement as the bind module is no longer launched. Warning Make sure that you can actually resolve the hostname of your database without accessing the database! It is advised to supply an IP address here to prevent chicken/egg problems! Warning Be very very sure that you configure the *g*mysql backend and not the mysql backend. See Section A.5. If you use the 'mysql' backend things will only appear to work. Now start PDNS using the monitor command: # /etc/init.d/pdns monitor (...) 15:31:30 About to create 3 backend threads 15:31:30 [gMySQLbackend] Failed to connect to database: Error: U nknown database 'pdnstest' 15:31:30 [gMySQLbackend] Failed to connect to database: Error: U nknown database 'pdnstest' 15:31:30 [gMySQLbackend] Failed to connect to database: Error: U nknown database 'pdnstest' This is as to be expected - we did not yet add anything to MySQL for PDNS to read from. At this point you may also see other errors which indicate that PDNS either could not find your MySQL server or was unable to connect to it. Fix these before proceeding. General MySQL knowledge is assumed in this chapter, please do not interpret these commands as DBA advice! __________________________________________________________ 4.1. Example: configuring MySQL Connect to MySQL as a user with sufficient privileges and issue the following commands: create table domains ( id INT auto_increment, name VARCHAR(255) NOT NULL, master VARCHAR(128) DEFAULT NULL, last_check INT DEFAULT NULL, type VARCHAR(6) NOT NULL, notified_serial INT DEFAULT NULL, account VARCHAR(40) DEFAULT NULL, primary key (id) )type=InnoDB; CREATE UNIQUE INDEX name_index ON domains(name); CREATE TABLE records ( id INT auto_increment, domain_id INT DEFAULT NULL, name VARCHAR(255) DEFAULT NULL, type VARCHAR(6) DEFAULT NULL, content VARCHAR(255) DEFAULT NULL, ttl INT DEFAULT NULL, prio INT DEFAULT NULL, change_date INT DEFAULT NULL, primary key(id) )type=InnoDB; CREATE INDEX rec_name_index ON records(name); CREATE INDEX nametype_index ON records(name,type); CREATE INDEX domain_id ON records(domain_id); create table supermasters ( ip VARCHAR(25) NOT NULL, nameserver VARCHAR(255) NOT NULL, account VARCHAR(40) DEFAULT NULL ); GRANT SELECT ON supermasters TO pdns; GRANT ALL ON domains TO pdns; GRANT ALL ON records TO pdns; Now we have a database and an empty table. PDNS should now be able to launch in monitor mode and display no errors: # /etc/init.d/pdns monitor (...) 15:31:30 PowerDNS 1.99.0 (Mar 12 2002, 15:00:28) starting up 15:31:30 About to create 3 backend threads 15:39:55 [gMySQLbackend] MySQL connection succeeded 15:39:55 [gMySQLbackend] MySQL connection succeeded 15:39:55 [gMySQLbackend] MySQL connection succeeded A sample query sent to the database should now return quickly without data: $ host www.test.com 127.0.0.1 www.test.com A record currently not present at localhost And indeed, the control console now shows: Mar 12 15:41:12 We're not authoritative for 'www.test.com', se nding unauth normal response Now we need to add some records to our database: # mysql pdnstest mysql> INSERT INTO domains (name, type) values ('test.com', 'N ATIVE'); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'test.com','localhost ahu@ds9a.nl 1','SOA',86400,NUL L); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'test.com','dns-us1.powerdns.net','NS',86400,NULL); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'test.com','dns-eu1.powerdns.net','NS',86400,NULL); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'www.test.com','199.198.197.196','A',120,NULL); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'mail.test.com','195.194.193.192','A',120,NULL); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'localhost.test.com','127.0.0.1','A',120,NULL); INSERT INTO records (domain_id, name, content, type,ttl,prio) VALUES (1,'test.com','mail.test.com','MX',120,25); If we now requery our database, www.test.com should be present: $ host www.test.com 127.0.0.1 www.test.com A 199.198.197.196 $ host -v -t mx test.com 127.0.0.1 Address: 127.0.0.1 Aliases: localhost Query about test.com for record types MX Trying test.com ... Query done, 1 answer, authoritative status: no error test.com 120 IN MX 25 mail.test.com Additional information: mail.test.com 120 IN A 195.194.193.192 To confirm what happened, issue the command SHOW * to the control console: % show * corrupt-packets=0,latency=0,packetcache-hit=2,packetcache-miss =5,packetcache-size=0, qsize-a=0,qsize-q=0,servfail-packets=0,tcp-answers=0,tcp-queri es=0, timedout-packets=0,udp-answers=7,udp-queries=7, % The actual numbers will vary somewhat. Now enter QUIT and start PDNS as a regular daemon, and check launch status: # /etc/init.d/pdns start pdns: started # /etc/init.d/pdns status pdns: 8239: Child running # /etc/init.d/pdns dump pdns: corrupt-packets=0,latency=0,packetcache-hit=0,packetcach e-miss=0, packetcache-size=0,qsize-a=0,qsize-q=0,servfail-packets=0,tcp- answers=0, tcp-queries=0,timedout-packets=0,udp-answers=0,udp-queries=0, You now have a working database driven nameserver! To convert other zones already present, use the zone2sql described in Appendix A. __________________________________________________________ 4.1.1. Common problems Most problems involve PDNS not being able to connect to the database. Can't connect to local MySQL server through socket '/tmp/mysql.sock' (2) Your MySQL installation is probably defaulting to another location for its socket. Can be resolved by figuring out this location (often /var/run/mysqld.sock), and specifying it in the configuration file with the gmysql-socket parameter. Another solution is to not connect to the socket, but to 127.0.0.1, which can be achieved by specifying gmysql-host=127.0.0.1. Host 'x.y.z.w' is not allowed to connect to this MySQL server These errors are generic MySQL errors. Solve them by trying to connect to your MySQL database with the MySQL console utility mysql with the parameters specified to PDNS. Consult the MySQL documentation. __________________________________________________________ Chapter 5. Dynamic resolution using the PipeBackend Also included in the PDNS distribution is the PipeBackend. The PipeBackend is primarily meant for allowing rapid development of new backends without tight integration with PowerDNS. It allows end-users to write PDNS backends in any language. A perl sample is provided. The PipeBackend is also very well suited for dynamic resolution of queries. Example applications include DNS based loadbalancing, geo-direction, DNS based failover with low TTLs. The Pipe Backend also has a separate chapter in the backends appendix, see Section A.1. Note The Pipe Backend currently does not function under FreeBSD 4.x and 5.x, probably due to unfavorable interactions between its threading implementation and the fork system call. Interestingly, the Linux PowerDNS binary running under the Linuxulator on FreeBSD does work. __________________________________________________________ 5.1. Deploying the PipeBackend with the BindBackend Included with the PDNS distribution is the example.pl backend which has knowledge of the example.com zone, just like the BindBackend. To install both, add the following to your pdns.conf: launch=pipe,bind bind-example-zones pipe-command=location/of/backend.pl Please adjust the pipe-command statement to the location of the unpacked PDNS distribution. If your backend is slow, raise pipe-timeout from its default of 2000ms. Now launch PDNS in monitor mode, and perform some queries. Note the difference with the earlier experiment where only the BindBackend was loaded. The PipeBackend is launched first and thus gets queried first. The sample backend.pl script knows about: * webserver.example.com A records pointing to 1.2.3.4, 1.2.3.5, 1.2.3.6 * www.example.com CNAME pointing to webserver.example.com * MBOXFW (mailbox forward) records pointing to powerdns@example.com. See the smtpredir documentation for information about MBOXFW. For more information about how to write exciting backends with the PipeBackend, see Appendix A. __________________________________________________________ Chapter 6. Logging & Monitoring Authoritative Server performance In a production environment, you will want to be able to monitor PDNS performance. For this purpose, currently two methods are available, the webserver and the init.d dump, show and mrtg, commands. Furthermore, PDNS can perform a configurable amount of operational logging. This chapter also explains how to configure syslog for best results. __________________________________________________________ 6.1. Webserver To launch the internal webserver, add a webserver statement to the pdns.conf. This will instruct the PDNS daemon to start a webserver on localhost at port 8081, without password protection. Only local users (on the same host) will be able to access the webserver by default. The webserver lists a lot of information about the PDNS process, including frequent queries, frequently failing queries, lists of remote hosts sending queries, hosts sending corrupt queries etc. The webserver does not allow remote management of the daemon. The following nameserver related configuration items are available: webserver If set to anything but 'no', a webserver is launched. webserver-address Address to bind the webserver to. Defaults to 127.0.0.1, which implies that only the local computer is able to connect to the nameserver! To allow remote hosts to connect, change to 0.0.0.0 or the physical IP address of your nameserver. webserver-password If set, viewers will have to enter this plaintext password in order to gain access to the statistics. webserver-port Port to bind the webserver to. Defaults to 8081. __________________________________________________________ 6.2. Via init.d commands As mentioned before, the init.d commands dump, show and mrtg fetch data from a running PDNS process. Especially mrtg is powerful - it outputs data in a format that is ready for processing by the MRTG graphing tool. MRTG can make insightful graphics on the performance of your nameserver, enabling the operator to easily spot trends. MRTG can be found on http://people.ee.ethz.ch/~oetiker/webtools/mrtg/mrtg.html A sample mrtg.conf: Interval: 5 WorkDir: /var/www/mrtg WriteExpires: yes Options[_]: growright,nopercent XSize[_]: 600 #--------------------------------------------------------------- Target[udp-queries]: `/etc/init.d/pdns mrtg udp-queries udp-answers` Options[udp-queries]: growright,nopercent,perminute MaxBytes[udp-queries]: 600000 AbsMax[udp-queries]: 600000 Title[udp-queries]: Queries per minute PageTop[udp-queries]:

Queries per minute

WithPeak[udp-queries]: ymwd YLegend[udp-queries]: queries/minute ShortLegend[udp-queries]: q/m LegendI[udp-queries]: udp-questions LegendO[udp-queries]: udp-answers Target[perc-failed]: `/etc/init.d/pdns mrtg udp-queries udp-answers` Options[perc-failed]: growright,dorelpercent,perminute MaxBytes[perc-failed]: 600000 AbsMax[perc-failed]: 600000 Title[perc-failed]: Queries per minute, with percentage success PageTop[perc-failed]:

Queries per minute, with percentage success WithPeak[perc-failed]: ymwd YLegend[perc-failed]: queries/minute ShortLegend[perc-failed]: q/m LegendI[perc-failed]: udp-questions LegendO[perc-failed]: udp-answers Target[packetcache-rate]: `/etc/init.d/pdns mrtg packetcache-hit udp-que ries` Options[packetcache-rate]: growright,dorelpercent,perminute Title[packetcache-rate]: packetcache hitrate MaxBytes[packetcache-rate]: 600000 AbsMax[packetcache-rate]: 600000 PageTop[packetcache-rate]:

packetcache hitrate

WithPeak[packetcache-rate]: ymwd YLegend[packetcache-rate]: queries/minute ShortLegend[packetcache-rate]: q/m LegendO[packetcache-rate]: total LegendI[packetcache-rate]: hit Target[packetcache-missrate]: `/etc/init.d/pdns mrtg packetcache-miss ud p-queries` Options[packetcache-missrate]: growright,dorelpercent,perminute Title[packetcache-missrate]: packetcache MISSrate MaxBytes[packetcache-missrate]: 600000 AbsMax[packetcache-missrate]: 600000 PageTop[packetcache-missrate]:

packetcache MISSrate

WithPeak[packetcache-missrate]: ymwd YLegend[packetcache-missrate]: queries/minute ShortLegend[packetcache-missrate]: q/m LegendO[packetcache-missrate]: total LegendI[packetcache-missrate]: MISS Target[latency]: `/etc/init.d/pdns mrtg latency` Options[latency]: growright,nopercent,gauge MaxBytes[latency]: 600000 AbsMax[latency]: 600000 Title[latency]: Query/answer latency PageTop[latency]:

Query/answer latency

WithPeak[latency]: ymwd YLegend[latency]: usec ShortLegend[latency]: usec LegendO[latency]: latency LegendI[latency]: latency Target[recursing]: `/etc/init.d/pdns mrtg recursing-questions recursing- answers` Options[recursing]: growright,nopercent,gauge MaxBytes[recursing]: 600000 AbsMax[recursing]: 600000 Title[recursing]: Recursive questions/answers PageTop[recursing]:

Recursing questions/answers

WithPeak[recursing]: ymwd YLegend[recursing]: queries/minute ShortLegend[recursing]: q/m LegendO[recursing]: recursing-questions LegendI[recursing]: recursing-answers __________________________________________________________ 6.3. Operational logging using syslog (logging-facility is available from 1.99.10 and onwards) This chapter assumes familiarity with syslog, the unix logging device. PDNS logs messages with different levels. The more urgent the message, the lower the 'priority'. By default, PDNS will only log messages with an urgency of 3 or lower, but this can be changed using the loglevel setting in the configuration file. Setting it to 0 will eliminate all logging, 9 will log everything. By default, logging is performed under the 'DAEMON' facility which is shared with lots of other programs. If you regard nameserving as important, you may want to have it under a dedicated facility so PDNS can log to its own files, and not clutter generic files. For this purpose, syslog knows about 'local' facilities, numbered from LOCAL0 to LOCAL7. To move PDNS logging to LOCAL0, add logging-facility=0 to your configuration. Furthermore, you may want to have separate files for the differing prioties - preventing lower priority messages from obscuring important ones. A sample syslog.conf might be: local0.info -/var/log/pdns.info local0.warn -/var/log/pdns.warn local0.err /var/log/pdns.err Where local0.err would store the really important messages. For performance and diskspace reasons, it is advised to audit your syslog.conf for statements also logging PDNS activities. Many syslog.confs have a '*.*' statement to /var/log/syslog, which you may want to remove. For performance reasons, be especially certain that no large amounts of synchronous logging take place. Under Linux, this is indicated by filenames not starting with a '-' - indicating a synchronous log, which hurts performance. Be aware that syslog by default logs messages at the configured priority and higher! To log only info messages, use local0.=info. __________________________________________________________ Chapter 7. Security settings & considerations 7.1. Settings PDNS has several options to easily allow it to run more securely. Most notable are the chroot, setuid and setgid options which can be specified. For additional information on PowerDNS security, PowerDNS security incidents and PowerDNS security policy, see Section 1.4. __________________________________________________________ 7.1.1. Running as a less privileged identity By specifying setuid and setgid, PDNS changes to this identity shortly after binding to the privileged DNS ports. These options are highly recommended. It is suggested that a separate identity is created for PDNS as the user 'nobody' is in fact quite powerful on most systems. Both these parameters can be specified either numerically or as real names. You should set these parameters immediately if they are not set! __________________________________________________________ 7.1.2. Jailing the process in a chroot The chroot option secures PDNS to its own directory so that even if it should become compromised and under control of external influences, it will have a hard time affecting the rest of the system. Even though this will hamper hackers a lot, chroot jails have been known to be broken. Warning When chrooting PDNS, take care that backends will be able to get to their files. Many databases need access to a UNIX domain socket which should live within the chroot. It is often possible to hardlink such a socket into the chroot dir. When running with master or slave support, be aware that many operating systems need access to specific libraries (ofen /lib/libnss*) in order to support resolution of domain names! You can also hardlink these. The default PDNS configuration is best chrooted to ./, which boils down to the configured location of the controlsocket. This is achieved by adding the following to pdns.conf: chroot=./, and restarting PDNS. __________________________________________________________ 7.2. Considerations In general, make sure that the PDNS process is unable to execute commands on your backend database. Most database backends will only need SELECT privilege. Take care to not connect to your database as the 'root' or 'sa' user, and configure the chosen user to have very slight privileges. Databases empathic-ally do not need to run on the same machine that runs PDNS! In fact, in benchmarks it has been discovered that having a separate database machine actually improves performance. Separation will enhance your database security highly. Recommended. __________________________________________________________ Chapter 8. Virtual hosting It may be advantageous to run multiple separate PDNS installations on a single host, for example to make sure that different customers cannot affect each others zones. PDNS fully supports running multiple instances on one host. To generate additional PDNS instances, copy the init.d script pdns to pdns-name, where name is the name of your virtual configuration. Must not contain a - as this will confuse the script. When you launch PDNS via this renamed script, it will seek configuration instructions not in pdns.conf but in pdns-name.conf, allowing for separate specification of parameters. Be aware however that the init.d force-stop will kill all PDNS instances! __________________________________________________________ Chapter 9. Performance 9.1. General advice In general, best performance is achieved on recent Linux 2.6 kernels and using MySQL, although many of the largest PowerDNS installations are based on PostgreSQL. FreeBSD appears to achieve lower packet rates both for the PowerDNS recursor as for the authoritative nameserver, this is still being investigated. No comparative measurements have been done for Solaris installations. On Linux, make sure to read Section 9.2. Database servers can require configuration to achieve decent performance. It is especially worth noting that several vendors ship PostgreSQL with a slow default configuration. __________________________________________________________ 9.2. Native Posix Thread Library vs LinuxThreads To get the best performance under Linux, especially on SMP systems, the use of NPTL is advised. The difference in performance can be over a factor of ten in some circumstances. NPTL is the default library on modern Linux distributions, so there is generally not a problem, except if you use a statically compiled version that, for portability reasons, defaults to LinuxThreads. This includes all .deb's and .rpm's provided by us up to and including 2.9.18. When running a PowerDNS-provided static binary of 2.9.18 or lower, it may make sense to recompile, or to upgrade to a newer version, if available. When recompiling, be sure to use a supported compiler, like g++ >3.2. You might also consider moving to a distribution supplied version. A good indication that your installation might benefit from such an upgrade is to watch the 'cs' count in the output of vmstat 1. If this is very high (> 10000), you are suffering from a LinuxThreads performance problem called 'overspin'. Thanks are due to L. Bunt Jackson who noted the static compilation problem in an article in Dr. Dobb's Journal. __________________________________________________________ 9.3. Performance related settings Different backends will have different characteristics - some will want to have more parallel instances than others. In general, if your backend is latency bound, like most relational databases are, it pays to open more backends. This is done with the distributor-threads setting. Of special importance is the choice between 1 or more backends. In case of only 1 thread, PDNS reverts to unthreaded operation which may be a lot faster, depending on your operating system and architecture. Another very important setting cache-ttl. PDNS caches entire packets it sends out so as to save the time to query backends to assemble all data. The default setting of 10 seconds may be low for high traffic sites, a value of 60 seconds rarely leads to problems. Some PDNS operators set cache-ttl to many hours or even days, and use pdns_control purge to selectively or globally notify PDNS of changes made in the backend. Also look at the Query Cache described in this chapter. It may materially improve your performance. To determine if PDNS is unable to keep up with packets, determine the value of the qsize-q variable. This represents the number of packets waiting for database attention. During normal operations the queue should be small. If it is known that backends will not contain CNAME records, the skip-cname setting can be used to prevent the normally mandatory CNAME lookup that is needed at least once for each DNS query. Much the same holds for the wildcards setting. On by default, each non-existent query will lead to a number of additional wildcard queries. If it is known that the backends do not contain wildcard records, performance can be improved by adding wildcards=no to pdns.conf. Logging truly kills performance as answering a question from the cache is an order of magnitude less work than logging a line about it. Busy sites will prefer to turn log-dns-details and log-failed-updates off. __________________________________________________________ 9.3.1. Packet Cache PDNS by default uses the 'Packet Cache' to recognise identical questions and supply them with identical answers, without any further processing. The default time to live is 10 seconds. It has been observed that the utility of the packet cache increases with the load on your nameserver. Not all backends may benefit from the packetcache. If your backend is memory based and does not lead to context switches, the packetcache may actually hurt performance. The size of the packetcache can be observed with /etc/init.d/pdns show packetcache-size __________________________________________________________ 9.3.2. Query Cache Besides entire packets, PDNS can also cache individual backend queries. Each DNS query leads to a number of backend queries, the most obvious additional backend query is the check for a possible CNAME. So, when a query comes in for the 'A' record for 'www.powerdns.com', PDNS must first check for a CNAME for 'www.powerdns.com'. The Query Cache caches these backend queries, many of which are quite repetitive. PDNS only caches queries with no answer, or with exactly one. In the future this may be expanded but this lightweight solution is very simple and therefore fast. Most gain is made from caching negative entries, ie, queries that have no answer. As these take little memory to store and are typically not a real problem in terms of speed-of-propagation, the default TTL for negative queries is a rather high 60 seconds. This only is a problem when first doing a query for a record, adding it, and immediately doing a query for that record again. It may then take up to 60 seconds to appear. Changes to existing records however do not fall under the negative query ttl ( negquery-cache-ttl), but under the generic query-ttl which defaults to 20 seconds. The default values should work fine for many sites. When tuning, keep in mind that the Query Cache mostly saves database access but that the Packet Cache also saves a lot of CPU because 0 internal processing is done when answering a question from the Packet Cache. __________________________________________________________ Chapter 10. Migrating to PDNS Before migrating to PDNS a few things should be considered. PDNS does not operate as a 'slave' or 'master' server with all backends Only the Generic SQL, OpenDBX and BIND backends have the ability to act as master or slave. To migrate, the zone2sql tool is provided. Additionally, the PowerDNS source comes with a number of diagnostic tools, which can be helpful in verifying proper PowerDNS operation, versus incumbent nameservers. See Chapter 20 for more details. __________________________________________________________ 10.1. Zone2sql Zone2sql parses Bind named.conf files and zonefiles and outputs SQL on standard out, which can then be fed to your database. Zone2sql understands the Bind master file extension '$GENERATE' and will also honour '$ORIGIN' and '$TTL'. For backends supporting slave operation (currently only the Generic PostgreSQL, Generic MySQL and BIND backend), there is also an option to keep slave zones as slaves, and not convert them to native operation. By default, zone2sql outputs code suitable for the mysqlbackend, but it can also generate SQL for the Generic PostgreSQL, Generic MySQL and Oracle backends. The following commands are available: --bare Output in a bare format, suitable for further parsing. The output is formatted as follows: domain_id'qname''qtype''content'pr iottl --gmysql Output in format suitable for the default configuration of the Generic MySQL backend. --gpgsql Output in format suitable for the default configuration of the Generic PostgreSQL backend. --help List options. --mysql Output in format suitable for the default configuration of the MySQL backend. Default. --named-conf=... Parse this named.conf to find locations of zones. --on-error-resume-next Ignore missing files during parsing. Dangerous. --oracle Output in format suitable for the default configuration of the Generic Oracle backend. --slave Maintain slave status of zones listed in named.conf as being slaves. The default behaviour is to convert all zones to native operation. --startid Supply a value for the first domain_id generated. Defaults at 0. --transactions For Oracle and PostgreSQL output, wrap each domain in a transaction for higher speed and integrity. --verbose Be verbose during conversion. --zone=... Parse only this zone file. Conflicts with --named-conf parameter. --zone-name=... When parsing a single zone without $ORIGIN statement, set this as the zone name. __________________________________________________________ Chapter 11. Recursion (only available from 1.99.8 and onwards, recursing component available since 2.9.5) From 2.9.5 onwards, PowerDNS offers both authoritative nameserving capabilities and a recursive nameserver component. These two halves are normally separate but many users insist on combining both recursion and authoritative service on one IP address. This can be likened to running Apache and Squid both on port 80. However, many sites want to do this anyhow and some with good reason. For example, a setup like this allows the creation of fake domains which only exist for local users. Such domains often don't end on ".com" or ".org" but on ".intern" or ".name-of-isp". PowerDNS can cooperate with either its own recursor or any other you have available to deliver recursive service on its port. By specifying the recursor option in the configuration file, questions requiring recursive treatment will be handed over to the IP address specified. An example configuration might be recursor=130.161.180.1, which designates 130.161.180.1 as the nameserver to handle recursive queries. As of 2.9.5, the recursing component of PowerDNS is a bit young and relatively untested but we hope people will want to use it anyhow. As an alternative, we highly advise the use of the DJBDNS dnscache (http://cr.yp.to/djbdns/dnscache.html). Take care not to point recursor to PDNS, which leads to a very tight packet loop! By specifying allow-recursion, recursion can be restricted to netmasks specified. The default is to allow recursion from everywhere. Example: allow-recursion=192.168.0.0/24, 10.0.0.0/8, 1.2.3.4. __________________________________________________________ 11.1. Details Questions carry a number of flags. One of these is called 'Recursion Desired'. If PDNS is configured to allow recursion, AND such a flag is seen, AND the IP address of the client is allowed to recurse via PDNS, then the packet may be handed to the recursing backend. If a Recursion Desired packet arrives and PDNS is configured to allow recursion, but not to the IP address of the client, resolution will proceed as if the RD flag were unset and the answer will indicate that recursion was not available. It is also possible to use a resolver living on a different port. To do so, specify a recursor like this: recursor=130.161.180.1:5300. If the backend does not answer a question within a large amount of time, this is logged as 'Recursive query for remote 10.96.0.2 with internal id 0 was not answered by backend within timeout, reusing id'. This may happen when using 'BIND' as a recursor as it is prone to drop queries which it can't answer immediately. To make sure that the local authoritative database overrides recursive information, PowerDNS first tries to answer a question from its own database. If that succeeds, the answer packet is sent back immediately without involving the recursor in any way. This means that for questions for which there is no answer, PowerDNS will consult the recursor for an recursive query, even if PowerDNS is authoritative for a domain! This will only cause problems if you 'fake' domains which don't really exist. If you want to create such fake domains or override existing domains, please set the allow-recursion-override feature (available as of 2.9.14). Some packets, like those asking for MX records which are needed for SMTP transport of email, can be subject to 'additional processing'. This means that a recursing nameserver is obliged to try to add A records (IP addresses) for any of the mailservers mentioned in the packet, should it have these addresses available. If PowerDNS encounters records needing such processing and finds that it does not have the data in its authoritative database, it will send an opportunistic quick query to the recursing component to see if it perhaps has such data. This question is worded such that the recursing nameserver should return immediately such as not to block the authoritative nameserver. This marks a change from pre-2.9.5 behaviour where a packet was handed wholesale to the recursor in case it needed additional processing which could not proceed from the authoritative database. __________________________________________________________ Chapter 12. PowerDNS resolver/recursing nameserver The PowerDNS recursor is part of the source tarball of the main PowerDNS distribution, but it is released separately. Starting from the version 3.0 pre-releases, there are zero known bugs or issues with the recursor. It is known to power the resolving needs of over 2 million internet connections. The documentation below is only for the 3.0 series, users of older versions are urged to upgrade! Notable features: * Uses MTasker (homepage) * Can handle thousands of concurrent questions. A dual Xeon 3GHz has been measured functioning very well at 9000 real life replayed packets per second, with 40% cpu idle. More testing equipment is needed to max out the recursor. * Powered by a highly modern DNS packet parser that should be resistant against many forms of buffer overflows. * Best spoofing protection that we know about, involving both source port randomisation and spoofing detection. * Uses 'connected' UDP sockets which allow the recursor to react quickly to unreachable hosts or hosts for which the server is running, but the nameserver is down. This makes the recursor faster to respond in case of misconfigured domains, which are sadly very frequent. * Special support for FreeBSD, Linux and Solaris stateful multiplexing (kqueue, epoll, completion ports). * Very fast, and contains innovative query-throttling code to save time talking to obsolete or broken nameservers. * Code is written linearly, sequentially, which means that there are no problems with 'query restart' or anything. * Relies heavily on Standard C++ Library infrastructure, which makes for little code (406 core lines). * Is very verbose in showing how recursion actually works, when enabled to do so with --verbose. * The algorithm is simple and quite nifty. The PowerDNS recursor is controlled and queried using the rec_control tool. __________________________________________________________ 12.1. pdns_recursor settings At startup, the recursing nameserver reads the file recursor.conf from the configuration directory, often /etc/powerdns or /usr/local/etc. Each setting below can appear on the command line, prefixed by '--', or in the configuration file. The command line overrides the configuration file. A switch can be set to on simply by passing it, like '--daemon', and turned off explicitly by '--daemon=off' or '--daemon=no'. The following settings can be configured: aaaa-additional-processing If turned on, the recursor will attempt to add AAAA IPv6 records to questions for MX records and NS records. Can be quite slow as absence of these records in earlier answers does not guarantee their non-existance. Can double the amount of queries needed. Off by default. allow-from Comma separated netmasks (both IPv4 and IPv6) that are allowed to use the server. The default allows access only from RFC 1918 private IP addresses, like 10.0.0.0/8. Due to the agressive nature of the internet these days, it is highly recommended to not open up the recursor for the entire internet. Questions from IP addresses not listed here are ignored and do not get an answer. allow-from-file Like allow-from, except reading from file. Overrides the 'allow-from' setting. To use this feature, supply one netmask per line, with optional comments preceeded by a #. Available since 3.1.5. auth-can-lower-ttl Authoritative zones can transmit a TTL value that is lower than that specified in the parent zone. This is called a 'delegation inconsistency'. To follow RFC 2181 paragraphs 5.2 and 5.4 to the letter, enable this feature. This will mean a slight deterioration of performance, and it will not solve any problems, but does make the recursor more standards compliant. Not recommended unless you have to tick an 'RFC 2181 compliant' box. Off by default. auth-zones Comma separated list of 'zonename=filename' pairs. Zones read from these files are served authoritatively. Example: auth-zones= ds9a.nl=/var/zones/ds9a.nl, powerdns.com=/var/zones/powerdns.com. Available since 3.1. chroot If set, chroot to this directory for more security. See Chapter 7. client-tcp-timeout Time to wait for data from TCP clients. Defaults to 2 seconds. config-dir Directory where the configuration file can be found. daemon Operate in the background, which is the default. delegation-only A Verisign special. dont-query The DNS is a public database, but sometimes contains delegations to private IP addresses, like for example 127.0.0.1. This can have odd effects, depending on your network, and may even be a security risk. Therefore, since version 3.1.5, the PowerDNS recursor by default does not query private space IP addresses. This setting can be used to expand or reduce the limitations. entropy-source From version 3.1.5 onwards, PowerDNS can read entropy from a (hardware) source. This is used for generating random numbers which are very hard to predict. Generally on UNIX platforms, this source will be /dev/urandom, which will always supply random numbers, even if entropy is lacking. Change to /dev/random if PowerDNS should block waiting for enough entropy to arrive. export-etc-hosts If set, this flag will export the host names and IP addresses mentioned in /etc/hosts. Available since 3.1. fork If running on an SMP system with enough memory, this feature forks PowerDNS so it benefits from two processors. Experimental. Renames controlsockets, so care is needed to connect to the right one using rec_control, using --socket-pid. forward-zones Comma separated list of 'zonename=IP' pairs. Queries for zones listed here will be forwarded to the IP address listed. forward-zones= ds9a.nl=213.244.168.210, powerdns.com=127.0.0.1. Available since 3.1. Since version 3.1.5, multiple IP addresses can be specified. Additionally, port numbers other than 53 can be configured. Sample syntax: forward-zones=ds9a.nl=213.244.168.210:5300;127.0.0.1, powerdns.com=127.0.0.1;9.8.7.6:530, or on the command line: --forward-zones="ds9a.nl=213.244.168.210:5300;127.0.0.1, powerdns.com=127.0.0.1;9.8.7.6:530", forward-zones-file Same as forward-zones, parsed from a file. Only 1 zone is allowed per line, specified as follows: ds9a.nl=213.244.168.210, 1.2.3.4:5300. No comments are allowed. Available since 3.1.5. hint-file If set, the root-hints are read from this file. If unset, default root hints are used. Available since 2.9.19. local-address Local IPv4 or IPv6 addresses to bind to, comma separated. Defaults to only loopback. Addresses can also contain port numbers, for IPv4 specify like this: 1.2.3.4:5300, for IPv6: [::1]:5300. Port specifications are available since 3.1.2. local-port Local port (singular) to bind to. Defaults to 53. log-common-errors Some DNS errors occur rather frequently and are no cause for alarm. Logging these is on by default. logging-facility If set to a digit, logging is performed under this LOCAL facility. See Section 6.3>. Available from 3.1.3 and onwards. Do not pass names like 'local0'! max-cache-entries Maximum number of cache entries. 1 million will generally suffice for most installations. max-negative-ttl A query for which there is authoritatively no answer is cached to quickly deny a record's existence later on, without putting a heavy load on the remote server. In practice, caches can become saturated with hundreds of thousands of hosts which are tried only once. This setting, which defaults to 3600 seconds, puts a maximum on the amount of time negative entries are cached. max-tcp-clients Maximum number of simultaneous incoming TCP connections allowed. Defaults to 128. Available since 2.9.18. max-tcp-per-client Maximum number of simultaneous incoming TCP connections allowed per client (remote IP address). Defaults to 0, which means unlimited. query-local-address Send out local queries from this address. Useful for anycast. query-local-address6 Send out local IPv6 queries from this address. Disabled by default, which also disables outgoing IPv6 support. A useful setting is ::0. quiet Don't log queries. On by default. remotes-ringbuffer-entries Number of entries in the remotes ringbuffer, which keeps statistics on who is querying your server. Can be read out using rec_control top-remotes. Defaults to 0. serve-rfc1918 On by default, this makes the server authoritatively aware of: 10.in-addr.arpa, 168.192.in-addr.arpa, 16-31.172.in-addr.arpa, which saves load on the AS112 servers. Individual parts of these zones can still be loaded or forwarded. server-id The PowerDNS recursor by replies to a query for 'id.server' with its hostname, useful for in clusters. Use this setting to override the answer it gives. setgid, setuid PowerDNS can change its user and group id after binding to its socket. Can be used for better security. socket-dir Where to store the control socket. This option also works with the controller, rec_control. spoof-nearmiss-max If set to non-zero, PowerDNS will assume it is being spoofed after seeing this many answers with the wrong id. Defaults to 20. trace If turned on, output impressive heaps of logging. May destroy performance under load. version Print version of this binary. Useful for checking which version of the PowerDNS recursor is installed on a system. Available since 3.1.5. version-string By default, PowerDNS replies to the 'version.bind' query with its version number. Security concious users may wish to override the reply PowerDNS issues. __________________________________________________________ 12.2. Controlling and querying the recursor To control and query the PowerDNS recursor, the tool rec_control is provided. This program talks to the recursor over the 'controlsocket', often stored in /var/run. As a sample command, try: # rec_control ping pong When not running as root, --socket-dir=/tmp might be appropriate. All rec_control commands are documented below: dump-cache filename Dumps the entire cache to the filename mentioned. This file should not exist already, PowerDNS will refuse to overwrite it. While dumping, the recursor will not answer questions. get statistic Retrieve a statistic. For items that can be queried, see below. get-parameter parameter1 parameter2 ../term> Retrieve a configuration parameter. All parameters from the configuration and command line can be queried. ping Check if server is alive. quit Request shutdown of the recursor. reload-zones Reload data about all authoritative and forward zones. The configuration file is also scanned to see if the auth-domain, forward-domain and export-etc-hosts statements have changed, and if so, these changes are incorporated. top-remotes Shows the top-20 most active remote hosts. Statistics are over the last 'remotes-ringbuffer-entries' queries, which defaults to 0. wipe-cache domain0. [domain1. domain2.] Wipe entries from the cache. This is useful if, for example, an important server has a new IP address, but the TTL has not yet expired. Multiple domain names can be passed. For versions before 3.1, you must terminate a domain with a .! So to wipe powerdns.org, issue 'rec_control wipe-cache powerdns.org.'. For later versions, the dot is optional. Note that deletion is exact, wiping 'com.' will leave 'www.powerdns.com.' untouched! Warning As of 3.1.7, this command also wipes the negative query cache for the specified domain. Warning Don't just wipe 'www.somedomain.com', its NS records or CNAME target may still be undesired, so wipe 'somedomain.com' as well. The command 'get' can query a large number of statistics, which are detailed in Section 12.5. More details on what 'throttled' queries and the like are can be found below in Section 12.4. __________________________________________________________ 12.3. PowerDNS Recursor performance To get the best out of the PowerDNS recursor, which is important if you are doing thousands of queries per second, please consider the following. * Limit the size of the cache to a sensible value. Cache hit rate does not improve meaningfully beyond 4 million max-cache-entries, reducing the memory footprint reduces CPU cache misses. * Compile using g++ 4.1 or later. This compiler really does a good job on PowerDNS, much better than 3.4 or 4.0. * Consider performing a 'profiled build' as described in the README. This is good for a 20% performance boost in some cases. * When running with >3000 queries per second, and running Linux versions prior to 2.6.17 on some motherboards, your computer may spend an inordinate amount of time working around an ACPI bug for each call to gettimeofday. This is solved by rebooting with 'clock=tsc' or upgrading to a 2.6.17 kernel. The above is relevant if dmesg shows Using pmtmr for high-res timesource * A busy server may need hundreds of file descriptors on startup, and deals with spikes better if it has that many available later on. Linux by default restricts processes to 1024 file descriptors, which should suffice most of the time, but Solaris has a default limit of 256. This can be raised using the ulimit command. FreeBSD has a default limit that is high enough for even very heavy duty use. * If you need it, try --fork, this will fork the daemon into two halves, allowing it to benefit from a second CPU. This feature almost doubles performance, but is a bit of a hack. Following the instructions above, you should be able to attain very high query rates. __________________________________________________________ 12.4. Details 12.4.1. Anti-spoofing The PowerDNS recursor 3.0 uses a fresh UDP source port for each outgoing query, making spoofing around 64000 times harder. This raises the bar from 'easily doable given some time' to 'very hard'. Under some circimstances, 'some time' has been measured at 2 seconds. This technique was first used by dnscache by Dan J. Bernstein. In addition, PowerDNS detects when it is being sent too many unexpected answers, and mistrusts a proper answer if found within a clutch of unexpected ones. This behaviour can be tuned using the spoof-nearmiss-max. __________________________________________________________ 12.4.2. Throttling PowerDNS implements a very simple but effective nameserver. Care has been taken not to overload remote servers in case of overly active clients. This is implemented using the 'throttle'. This accounts all recent traffic and prevents queries that have been sent out recently from going out again. There are three levels of throttling. * If a remote server indicates that it is lame for a zone, the exact question won't be repeated in the next 60 seconds. * After 4 ServFail responses in 60 seconds, the query gets throttled too. * 5 timeouts in 20 seconds also lead to query suppression. __________________________________________________________ 12.5. Statistics The rec_control get command can be used to query the following keys, either single keys or multiple keys at once: all-outqueries counts the number of outgoing UDP queries since star ting answers0-1 counts the number of queries answered within 1 milis econd answers100-1000 counts the number of queries answered within 1 secon d answers10-100 counts the number of queries answered within 100 mil iseconds answers1-10 counts the number of queries answered within 10 mili seconds answers-slow counts the number of queries answered after 1 second cache-entries shows the number of entries in the cache cache-hits counts the number of cache hits since starting cache-misses counts the number of cache misses since starting chain-resends number of queries chained to existing outstanding qu ery client-parse-errors counts number of client packets that could not be pa rsed concurrent-queries shows the number of MThreads currently running dlg-only-drops number of records dropped because of delegation only setting negcache-entries shows the number of entries in the Negative answer c ache noerror-answers counts the number of times it answered NOERROR since starting nsspeeds-entries shows the number of entries in the NS speeds map nsset-invalidations number of times an nsset was dropped because it no l onger worked nxdomain-answers counts the number of times it answered NXDOMAIN sinc e starting outgoing-timeouts counts the number of timeouts on outgoing UDP querie s since starting qa-latency shows the current latency average questions counts all End-user initiated queries with the RD bi t set resource-limits counts number of queries that could not be performed because of resource limits server-parse-errors counts number of server replied packets that could n ot be parsed servfail-answers counts the number of times it answered SERVFAIL sinc e starting spoof-prevents number of times PowerDNS considered itself spoofed, and dropped the data sys-msec number of CPU milliseconds spent in 'system' mode tcp-client-overflow number of times an IP address was denied TCP access because it already had too many connections tcp-outqueries counts the number of outgoing TCP queries since star ting tcp-questions counts all incoming TCP queries (since starting) throttled-out counts the number of throttled outgoing UDP queries since starting throttle-entries shows the number of entries in the throttle map unauthorized-tcp number of TCP questions denied because of allow-from restrictions unauthorized-udp number of UDP questions denied because of allow-from restrictions unexpected-packets number of answers from remote servers that were unex pected (might point to spoofing) uptime number of seconds process has been running (since 3. 1.5) user-msec number of CPU milliseconds spent in 'user' mode In the rrd/ subdirectory a number of rrdtool scripts is provided to make nice graphs of all these numbers. Every half our or so, the recursor outputs a line with statistics. More infrastructure is planned so as to allow for Cricket or MRTG graphs. To force the output of statistics, send the process a SIGUSR1. A line of statistics looks like this: Feb 10 14:16:03 stats: 125784 questions, 13971 cache entries, 309 negati ve entries, 84% cache hits, outpacket/query ratio 37%, 12% throttled This means that there are 13791 different names cached, which each may have multiple records attached to them. There are 309 items in the negative cache, items of which it is known that don't exist and won't do so for the near future. 84% of incoming questions could be answered without any additional queries going out to the net. The outpacket/query ratio means that on average, 0.37 packets were needed to answer a question. Initially this ratio may be well over 100% as additional queries may be needed to actually recurse the DNS and figure out the addresses of nameservers. Finally, 12% of queries were not performed because identical queries had gone out previously, saving load servers worldwide. __________________________________________________________ 12.6. Scripting As of version 3.1.7 of the PowerDNS Recursor, it is possible to modify resolving behaviour using simple scripts written in the Lua programming language. Warning This functionality is expected to change from version to version as additional scripting needs become apparant! These scripts can be used to quickly override dangerous domains, for load balancing or for legal or commercial purposes. As of 3.1.7, queries can be intercepted in two places: before the resolving logic starts to work, plus after the resolving process failed to find a correct answer for a domain. __________________________________________________________ 12.6.1. Configuring Lua scripts In order to load scripts, the PowerDNS Recursor must have Lua support built in. The packages distributed from the PowerDNS website have this language enabled, other distributions may differ. To compile with Lua support, use: LUA=1 make or LUA=1 gmake as the case may be. Paths to the Lua include files and binaries may be found near the top of the Makefile. If lua support is available, a script can be configured either via the configuration file, or at runtime via the rec_control tool. Scripts can be reloaded or unloaded at runtime with no interruption in operations. If a new script contains syntax errors, the old script remains in force. On the commandline, or in the configuration file, the setting lua-dns-script can be used to supply a full path to a 'lua' script. At runtime, rec_control reload-lua-script can be used to either reload the script from its current location, or, when passed a new filename, load one from a new location. A failure to parse the new script will leave the old script in working order. Finally, rec_control unload-lua-script can be used to remove the currently installed script, and revert to unmodified behaviour. __________________________________________________________ 12.6.2. Writing Lua PowerDNS Recursor scripts Once a script is loaded, PowerDNS looks for two functions: preresolve and nxdomain. Either or both of these can be absent, in which case the corresponding functionality is disabled. preresolve is called before any DNS resolution is attempted, and if this function indicates it, it can supply a direct answer to the DNS query, overriding the internet. This is useful to combat botnets, or to disable domains unacceptable to an organization for whatever reason. nxdomain is called after the DNS resolution process has run its course, but ended in an 'NXDOMAIN' situation, indicating that the domain or the specific record does not exist. This can be used for various purposes. Both functions are passed the IP address of the requestor, plus the name and type being requested. In return, these functions indicate if they have taken over the request, or want to let normal proceedings take their course. Beyond version 3.1.7, the IP address on which the question was received is inserted immediately after the IP address of the requestor - so in that case there are 4 parameters. If a function has taken over a request, it should return an rcode (usually 0), and specify a table with records to be put in the answer section of a packet. An interesting rcode is NXDOMAIN (3, or pdns.NXDOMAIN), which specifies the non-existence of a domain. Returning -1 and an empty table signifies that the function chose not to intervene. A minimal sample script: function nxdomain ( ip, domain, qtype ) print ("nxhandler called for: ", ip, domain, qtype) ret={} if qtype ~= pdns.A then return -1, ret end -- only A records if not string.find(domain, "^www%.") then return -1, ret end -- only things that start with www. if not matchnetmask(ip, "10.0.0.0/8", "192.168.0.0/16") then return - 1, ret end -- only interfere with local queries ret[1]={qtype=pdns.A, content="127.1.2.3"} -- add IN A 127.1.2.3 ret[2]={qtype=pdns.A, content="127.3.2.1"} -- add IN A 127.3.2.1 return 0, ret -- return no error, plus records end Warning Please do NOT use the above sample script in production! Responsible NXDomain redirection requires more attention to detail. Note that the domain is passed to the Lua function terminated by a '.'. A more complete sample script is provided as powerdns-example-script.lua in the PowerDNS Recursor distribution. The answer content format is (nearly) identical to the storage in the PowerDNS Authoritative Server database, or as in zone files. The exception is that, unlike in the database, there is no 'prio' field, which means that an MX record with priority 25 pointing to 'smtp.mailserver.com' would be encoded as '25 smtp.mailserver.com.'. Useful return 'rcodes' include 0 for "no error" and pdns.NXDOMAIN for "NXDOMAIN". Fields that can be set in the return table include: content Content of the record, as specified above in 'zone file format'. No default, mandatory field. place Place of this record. Defaults to 1, indicating 'Answer' section. Can also be 2, for Authority of 3 for Additional. When using this rare feature, always emit records with 'Place' in ascending order. This field is usually not needed. qname qname of the answer, the 'name' of the record. Defaults to the name of the query, which is almost always correct except when specifying additional records or rolling out a CNAME chain. qtype Currently the numerical qtype of the answer, defaulting to '1' which is an A record. Can be also be specified as pdns.A, or pdns.CNAME etc. ttl Time to live of a record. Defaults to 3600. Be sure not to specify differing TTLs within answers with an identical qname. While this will be encoded in DNS, actual results may be undesired. Warning The result table must have indexes that start at 1! Otherwise the first or confusingly the last entry of the table will be ignored. A useful technique is to return data using: return 0, {{qtype=1, content="1.2.3.4"}, {qtype=1, content="4.3.2.1"}} as this will get the numbering right automatically. The function matchnetmask(ip, netmask1, netmask2..) (or matchnetmask(ip, {netmask1, netmask2})) is available to match incoming queries against a number of netmasks. If any of these matches, the function returns true. To log messages with the main PowerDNS Recursor process, use pdnslog(message). Available since 3.1.8. __________________________________________________________ 12.7. Design and Engineering of the PowerDNS Recursor Warning This section is aimed at programmers wanting to contibute to the recursor, or to help fix bugs. It is not required reading for a PowerDNS operator, although it might prove interesting. The PowerDNS Recursor consists of very little code, the core DNS logic is less than a thousand lines. This smallness is achieved through the use of some fine infrastructure: MTasker, MOADNSParser, MPlexer and the C++ Standard Library/Boost. This page will explain the conceptual relation between these components, and the route of a packet through the program. __________________________________________________________ 12.7.1. The PowerDNS Recursor The Recursor started out as a tiny project, mostly a technology demonstration. These days it consists of the core plus 9000 lines of features. This combined with a need for very high performance has made the recursor code less accessible than it was. The page you are reading hopes to rectify this situation. __________________________________________________________ 12.7.2. Synchronous code using MTasker The original name of the program was syncres, which is still reflected in the filename syncres.cc, and the class SyncRes. This means that PowerDNS is written naively, with one thread of execution per query, synchronously waiting for packets, Normally this would lead to very bad performance (unless running on a computer with very fast threading, like possibly the Sun CoolThreads family), so PowerDNS employs MTasker for very fast userspace threading. MTasker, which was developed separately from PowerDNS, does not provide a full multithreading system but restricts itself to those features a nameserver needs. It offers cooperative multitasking, which means there is no forced preemption of threads. This in turn means that no two MThreads ever really run at the same time. This is both good and bad, but mostly good. It means PowerDNS does not have to think about locking. No two threads will ever be talking to the DNS cache at the same time, for example. It also means that the recursor could block if any operation takes too long. The core interaction with MTasker are the waitEvent() and sendEvent() functions. These pass around PacketID objects. Everything PowerDNS needs to wait for is described by a PacketID event, so the name is a bit misleading. Waiting for a TCP socket to have data available is also passed via a PacketID, for example. The version of MTasker in PowerDNS is newer than that described at the MTasker site, with a vital difference being that thet waitEvent() structure passes along a copy of the exact PacketID sendEvent() transmitted. Furthermore, threads can trawl through the list of events being waited for and modify the respective PacketIDs. This is used for example with near miss packets: packets that appear to answer questions we asked, but differ in the DNS id. On seeing such a packet, the recursor trawls through all PacketIDs and if it finds any nearmisses, it updates the PacketID::nearMisses counter. The actual PacketID thus lives inside MTasker while any thread is waiting for it. __________________________________________________________ 12.7.3. MPlexer The Recursor uses a separate socket per outgoing query. This has the important benefit of making spoofing 64000 times harder, and additionally means that ICMP errors are reported back to the program. In measurements this appears to happen to one in ten queries, which would otherwise take a two-second timeout before PowerDNS moves on to another nameserver. However, this means that the program routinely needs to wait on hundreds or even thousands of sockets. Different operating systems offer various ways to monitor the state of sockets or more generally, filedescriptors. To abstract out the differing strategies (select, epoll, kqueue, completion ports), PowerDNS contains MPlexer classes, all of which descend from the FDMultiplexer class. This class is very simple and offers only five important methods: addReadFD(), addWriteFD(), removeReadFD(), removeWriteFD() and run. The arguments to the add functions consist of an fd, a callback, and a boost::any variable that is passed as a reference to the callback. This might remind you of the MTasker above, and it is indeed the same trick: state is stored within the MPlexer. As long as a filedescriptor remains within either the Read or Write active list, its state will remain stored. On arrival of a packet (or more generally, when an FD becomes readable or writable, which for example might mean a new TCP connection), the callback is called with the aforementioned reference to its parameter. The callback is free to call removeReadFD() or removeWriteFD() to remove itself from the active list. PowerDNS defines such callbacks as newUDPQuestion(), newTCPConnection(), handleRunningTCPConnection(). Finally, the run() method needs to be called whenever the program is ready for new data. This happens in the main loop in pdns_recursor.cc. This loop is what MTasker refers to as the kernel. In this loop, any packets or other MPlexer events get translated either into new MThreads within MTasker, or into calls to sendEvent(), which in turn wakes up other MThreads. __________________________________________________________ 12.7.4. MOADNSParser Yes, this does stand for the Mother of All DNS Parsers. And even that name does not do it justice! The MOADNSParser is the third attempt I've made at writing DNS packet parser and after two miserable failures, I think I've finally gotten it right. Writing and parsing DNS packets, and the DNS records it contains, consists of four things: 1. Parsing a DNS record (from packet) into memory 2. Generating a DNS record from memory (to packet) 3. Writing out memory to user-readable zone format 4. Reading said zone format into memory This gets tedious very quickly, as one needs to implement all four operations for each new record type, and there are dozens of them. While writing the MOADNSParser, it was discovered there is a remarkable symmetry between these four transitions. DNS Records are nearly always laid out in the same order in memory as in their zone format representation. And reading is nothing but inverse writing. So, the MOADNSParser is built around the notion of a Conversion, and we write all Conversion types once. So we have a Conversion from IP address in memory to an IP address in a DNS packet, and vice versa. And we have a Conversion from an IP address in zone format to memory, and vice versa. This in turn means that the entire implementation of the ARecordContent is as follows (wait for it!) conv.xfrIP(d_ip); Through the use of the magic called c++ Templates, this one line does everything needed to perform the four operations mentioned above. At one point, I got really obsessed with PowerDNS memory use. So, how do we store DNS data in the PowerDNS recorsor? I mentioned memory above a lot - this means we could just store the DNSRecordContent objects. However, this would be wasteful. For example, storing the following: www.ds9a.nl 3600 IN CNAME outpost.ds9a.nl. Would duplicate a lot of data. So, what is actually stored is a partial DNS packet. To store the CNAMEDNSRecordContent that corresponds to the above, we generate a DNS packet that has www.ds9a.nl IN CNAME as its question. Then we add 3600 IN CNAME outpost.ds9a.nl. as its answer. Then we chop off the question part, and store the rest in the www.ds9a.nl IN CNAME key in our cache. When we need to retrieve www.ds9a.nl IN CNAME, the inverse happens. We find the proper partial packet, prefix it with a question for www.ds9a.nl IN CNAME, and expand the resulting packet into the answer 3600 IN CNAME outpost.ds9a.nl.. Why do we go through all these motions? Because of DNS compression, which allows us to omit the whole .ds9a.nl. part, saving us 9 bytes. This is amplified when storing multiple MX records which all look more or less alike. This optimization is not performed yet though. Even without compression, it makes sense as all records are automatically stored very compactly. The PowerDNS recursor only parses a number of well known record types and passes all other information across verbatim - it doesn't have to know about the content it is serving. __________________________________________________________ 12.7.5. The C++ Standard Library / Boost C++ is a powerful language. Perhaps a bit too powerful at times, you can turn a program into a real freakshow if you so desire. PowerDNS generally tries not to go overboard in this respect, but we do build upon a very advanced part of the Boost C++ library: boost::multi index container. This container provides the equivalent of SQL indexes on multiple keys. It also implements compound keys, which PowerDNS uses as well. The main DNS cache is implemented as a multi index container object, with a compound key on the name and type of a record. Furthermore, the cache is sequenced, each time a record is accessed it is moved to the end of the list. When cleanup is performed, we start at the beginning. New records also get inserted at the end. For DNS correctness, the sort order of the cache is case insensitive. The multi index container appears in other parts of PowerDNS, and MTasker as well. __________________________________________________________ 12.7.6. Actual DNS Algorithm The DNS rfcs do define the DNS algorithm, but you can't actually implement it exactly that way, it was written in 1987. Also, like what happened to HTML, it is expected that even non-standards conforming domains work, and a sizeable fraction of them is misconfigured these days. Everything begins with SyncRes::beginResolve(), which knows nothing about sockets, and needs to be passed a domain name, dns type and dns class which we are interested in. It returns a vector of DNSResourceRecord objects, ready for writing either into an answer packet, or for internal use. After checking if the query is for any of the hardcoded domains (localhost, version.bind, id.server), the query is passed to SyncRes::doResolve, together with two vital parameters: the depth and beenthere set. As the word recursor implies, we will need to recurse for answers. The depth parameter documents how deep we've recursed already. The beenthere set prevents loops. At each step, when a nameserver is queried, it is added to the beenthere set. No nameserver in the set will ever be queried again for the same question in the recursion process - we know for a fact it won't help us further. This prevents the process from getting stuck in loops. SyncRes::doResolve first checks if there is a CNAME in cache, using SyncRes::doCNAMECacheCheck, for the domain name and type queried and if so, changes the query (which is passed by reference) to the domain the CNAME points to. This is the cause of many DNS problems, a CNAME record really means start over with this query. This is followed by a call do SyncRes::doCacheCheck, which consults the cache for a straight answer to the question (as possibly rerouted by a CNAME). This function also consults the so called negative cache, but we won't go into that just yet. If this function finds the correct answer, and the answer hasn't expired yet, it gets returned and we are (almost) done. This happens in 80 to 90% of all queries. Which is good, as what follows is a lot of work. To recap: 1. beginResolve() - entry point, does checks for hardcoded domains 2. doResolve() - start of recursion process, gets passed depth of 0 and empty beenthere set 3. doCNAMECacheCheck() - check if there is a CNAME in cache which would reroute the query 4. doCacheCheck() - see if cache contains straight answer to possibly rerouted query. If the data we were queried for was in the cache, we are almost done. One final step, which might as well be optional as nobody benefits from it, is SyncRes::addCruft. This function does additional processing, which means that if the query was for the MX record of a domain, we also add the IP address of the mail exchanger. __________________________________________________________ 12.7.7. The non-cached case This is where things get interesting, because we start out with a nearly empty cache and have to go out to the net to get answers to fill it. The way DNS works, if you don't know the answer to a question, you find somebody who does. Initially you have no other place to go than the root servers. This is embodied in the SyncRes::getBestNSNamesFromCache method, which gets passed the domain we are interested in, as well as the depth and beenthere parameters mentioned earlier. From now on, assume our query will be for www.powerdns.com.. SyncRes::getBestNSNamesFromCache will first check if there are NS records in cache for www.powerdns.com., but there won't be. It then checks powerdns.com. NS, and while these records do exist on the internet, the recursor doesn't know about them yet. So, we go on to check the cache for com. NS, for which the same holds. Finally we end up checking for . NS, and these we do know about: they are the root servers and were loaded into PowerDNS on startup. So, SyncRes::getBestNSNamesFromCache fills out a set with the names of nameservers it knows about for the . zone. This set, together with the original query www.powerdns.com gets passed to SyncRes::doResolveAt. This function can't yet go to work immediately though, it only knows the names of nameservers it can try. This is like asking for directions and instead of hearing take the third right you are told go to 123 Fifth Avenue, and take a right - the answer doesn't help you further unless you know where 123 Fifth Avenue is. SyncRes::doResolveAt first shuffles the nameservers both randomly and on performance order. If it knows a nameserver was fast in the past, it will get queried first. More about this later. Ok, here is the part where things get a bit scary. How does SyncRes::doResolveAt find the IP address of a nameserver? Well, by calling SyncRes::getAs (get A records), which in turn calls.. SyncRes::doResolve. Hang on! That's where we came from! Massive potential for loops here. Well, it turns out that for any domain which can be resolved, this loop terminates. We do pass the beenthere set again, which makes sure we don't keep on asking the same questions to the same nameservers. Ok, SyncRes::getAs will give us the IP addresses of the chosen root-server, because these IP addresses were loaded on startup. We then ask these IP addresses (nameservers can have several) for its best answer for www.powerdns.com.. This is done using the LWRes class and specifically LWRes::asyncresolve, which gets passed domain name, type and IP address. This function interacts with MTasker and MPlexer above in ways which needn't concern us now. When it returns, the LWRes object contains the best answers the queried server had for our domain, which in this case means it tells us about the nameservers of com., and their IP addresses. All the relevant answers it gives are stored in the cache (or actually, merged), after which SyncRes::doResolveAt (which we are still in) evaluates what to do now. There are 6 options: 1. The final answer is in, we are done, return to SyncRes::doResolve and SyncRes::beginResolve 2. The nameserver we queried tells us the domain we asked for authoritatively does not exist. In case of the root-servers, this happens when we query for www.powerdns.kom. for example, there is no kom.. Return to SyncRes::beginResolve, we are done. 3. A lesser form - it tells us it is authoritative for the query we asked about, but there is no record matching our type. This happens when querying for the IPv6 address of a host which only has an IPv4 address. Return to SyncRes::beginResolve, we are done. 4. The nameserver passed us a CNAME to another domain, and we need to reroute. Go to SyncRes::doResolve for the new domain. 5. The namserver did not know about the domain, but does know who does, a referral. Stay within doResolveAt and loop to these new nameservers. 6. The nameserver replied saying no idea. This is called a lame delegation. Stay within SyncRes::doResolveAt and try the other nameservers we have for this domain. When not redirected using a CNAME, this function will loop until it has exhausted all nameservers and all their IP addresses. DNS is surprisingly resilient that there is often only a single non-broken nameserver left to answer queries, and we need to be prepared for that. This is the whole DNS algorithm in PowerDNS, all in less than 700 lines of code. It contains a lot of tricky bits though, related to the cache. __________________________________________________________ 12.7.8. Some of the things we glossed over Whenever a packet is sent to a remote nameserver, the response time is stored in the SyncRes::s_nsSpeeds map, using an exponentially weighted moving average. This EWMA averages out different response times, and also makes them decrease over time. This means that a nameserver that hasn't been queried recently gradually becomes faster in the eyes of PowerDNS, giving it a chance again. A timeout is accounted as a 1s response time, which should take that server out of the running for a while. Furthermore, queries are throttled. This means that each query to a nameserver that has failed is accounted in the s_throttle object. Before performing a new query, the query and the nameserver are looked up via shouldThrottle. If so, the query is assumed to have failed without even being performed. This saves a lot of network traffic and makes PowerDNS quick to respond to lame servers. It also offers a modicum of protection against birthday attack powered spoofing attempts, as PowerDNS will not innundate a broken server with queries. The negative query cache we mentioned earlier caches the cases 2 and 3 in the enumeration above. This data needs to be stored separately, as it represents non-data. Each negcache query entry is the name of the SOA record that was presented with the evidence of non-existance. This SOA record is then retrieved from the regular cache, but with the TTL that originally came with the NXDOMAIN (case 2) or NXRRSET (case 3). __________________________________________________________ 12.7.9. The Recursor Cache As mentioned before, the cache stores partial packets. It also stores not the Time To Live of records, but in fact the Time To Die. If the cache contains data, but it is expired, that data should not be deemed present. This bit of PowerDNS has proven tricky, leading to deadlocks in the past. There are some other very tricky things to deal with. For example, through a process called more details, a domain might have more nameservers than listed in its parent zone. So, there might only be two nameservers for powerdns.com. in the com. zone, but the powerdns.com zone might list more. This means that the cache should not, when talking to the com. servers later on, overwrite these four nameservers with only the two copies the com. servers pass us. However, in other cases (like for example for SOA and CNAME records), new data should overwrite old data. Note that PowerDNS deviates from RFC 2181 (section 5.4.1) in this respect. __________________________________________________________ 12.7.10. Some small things The server-side part of PowerDNS (pdns_recursor.cc), which listens to queries by end-users, is fully IPv6 capable using the ComboAddress class. This class is in fact a union of a struct sockaddr_in and a struct sockaddr_in6. As long as the sin_family (or sin6_family) and sin_port members are in the same place, this works just fine, allowing us to pass a ComboAddress*, cast to a sockaddr* to the socket functions. For convenience, the ComboAddress also offers a length() method which can be used to indicate the length - either sizeof(sockaddr_in) or sizeof(sockaddr_in6). Access to the recursor is governed through the NetmaskGroup class, which internally contains Netmaks, which in turn contain a ComboAddress. __________________________________________________________ Chapter 13. Master/Slave operation & replication PDNS offers full master and slave semantics for replicating domain information. Furthermore, PDNS can benefit from native database replication. __________________________________________________________ 13.1. Native replication Native replication is the default, unless other operation is specifically configured. Native replication basically means that PDNS will not send out DNS update notifications, nor will react to them. PDNS assumes that the backend is taking care of replication unaided. MySQL replication has proven to be very robust and well suited, even over transatlantic connections between badly peering ISPs. Other PDNS users employ Oracle replication which also works very well. To use native replication, configure your backend storage to do the replication and do not configure PDNS to do so. __________________________________________________________ 13.2. Slave operation On launch, PDNS requests from all backends a list of domains which have not been checked recently for changes. This should happen every 'refresh' seconds, as specified in the SOA record. All domains that are unfresh are then checked for changes over at their master. If the SOA serial number there is higher, the domain is retrieved and inserted into the database. In any case, after the check the domain is declared 'fresh', and will only be checked again after 'refresh' seconds have passed. Warning Slave support is OFF by default, turn it on by adding slave to the configuration. The same holds for master operation. Both can be on simultaneously. PDNS also reacts to notifies by immediately checking if the zone has updated and if so, retransfering it. All backends which implement this feature must make sure that they can handle transactions so as to not leave the zone in a half updated state. MySQL configured with either BerkeleyDB or InnoDB meets this requirement, as do PostgreSQL and Oracle. The Bindbackend implements transaction semantics by renaming files if and only if they have been retrieved completely and parsed correctly. Slave operation can also be programmed using several pdns_control commands, see Section B.1.1. The 'retrieve' command is especially useful as it triggers an immediate retrieval of the zone from the configured master. Since 2.9.21, PowerDNS supports multiple masters. For the BIND backend, the native BIND configuration language suffices to specify multiple masters, for SQL based backends, list all master servers separated by commas in the 'master' field of the domains table. __________________________________________________________ 13.2.1. Supermaster automatic provisioning of slaves PDNS can recognize so called 'supermasters'. A supermaster is a host which is master for domains and for which we are to be a slave. When a master (re)loads a domain, it sends out a notification to its slaves. Normally, such a notification is only accepted if PDNS already knows that it is a slave for a domain. However, a notification from a supermaster carries more persuasion. When PDNS determines that a notification comes from a supermaster and it is bonafide, PDNS can provision the domain automatically, and configure itself as a slave for that zone. Before a supermaster notification succeeds, the following conditions must be met: * The supermaster must carry a SOA record for the notified domain * The supermaster IP must be present in the 'supermaster' table * The set of NS records for the domain, as retrieved by the slave from the supermaster, must include the name that goes with the IP address in the supermaster table So, to benefit from this feature, a backend needs to know about the IP address of the supermaster, and how PDNS will be listed in the set of NS records remotely, and the 'account' name of your supermaster. There is no need to fill the account name out but it does help keep track of where a domain comes from. __________________________________________________________ 13.3. Master operation When operating as a master, PDNS sends out notifications of changes to slaves, which react to these notifications by querying PDNS to see if the zone changed, and transferring its contents if it has. Notifications are a way to promptly propagate zone changes to slaves, as described in RFC 1996. Warning Master support is OFF by default, turn it on by adding master to the configuration. The same holds for slave operation. Both can be on simultaneously. Left open by RFC 1996 is who is to be notified - which is harder to figure out than it sounds. All slaves for this domain must receive a notification but the nameserver only knows the names of the slaves - not the IP addresses, which is where the problem lies. The nameserver itself might be authoritative for the name of its secondary, but not have the data available. To resolve this issue, PDNS tries multiple tactics to figure out the IP addresses of the slaves, and notifies everybody. In contrived configurations this may lead to duplicate notifications being sent out, which shouldn't hurt. Some backends may be able to detect zone changes, others may chose to let the operator indicate which zones have changed and which haven't. Consult the documentation for your backend to see how it processes changes in zones. To help deal with slaves that may have missed notifications, or have failed to respond to them, several override commands are available via the pdns_control tool (Section B.1.1): pdns_control notify domain This instructs PDNS to notify all IP addresses it considers to be slaves of this domain. pdns_control notify-host domain ip-address This is truly an override and sends a notification to an arbitrary IP address. Can be used in 'also-notify' situations or when PDNS has trouble figuring out who to notify - which may happen in contrived configurations. __________________________________________________________ Chapter 14. Fancy records for seamless email and URL integration PDNS also supports so called 'fancy' records. A Fancy Record is actually not a DNS record, but it is translated into one. Currently, two fancy records are implemented, but not very useful without additional unreleased software. For completeness, they are listed here. The software will become available later on and is part of the Express and PowerMail suite of programs. These records imply extra database lookups which has a performance impact. Therefore fancy records are only queried for if they are enabled with the fancy-records command in pdns.conf. MBOXFW This record denotes an email forward. A typical entry looks like this: support@yourdomain.com MBOXFW you@yourcompany. com When PDNS encounters a request for an MX record for yourdomain.com it will, if fancy records are enabled, also check for the existence of an MBOXFW record ending on '@yourdomain.com', in which case it will hand out a record containing the configured smtpredirector. This server should then also be able to access the PDNS database to figure out where mail to support@yourdomain.com should go to. URL URL records work in much the same way, but for HTTP. A sample record: yourdomain.com URL http://somewhere.else.com/y ourdomain A URL record is converted into an A record containing the IP address configured with the urlredirector setting. On that IP address a webserver should live that knows how to redirect yourdomain.com to http://somewhere.else.com/yourdomain. __________________________________________________________ Chapter 15. Index of all Authoritative Server settings All PDNS Authoritative Server settings are listed here, excluding those that originate from backends, which are documented in the relevant chapters. allow-axfr-ips=... Behaviour pre 2.9.10: When not allowing AXFR (disable-axfr), DO allow from these IP addresses or netmasks. Behaviour post 2.9.10: If set, only these IP addresses or netmasks will be able to perform AXFR. allow-recursion=... By specifying allow-recursion, recursion can be restricted to netmasks specified. The default is to allow recursion from everywhere. Example: allow-recursion=192.168.0.0/24, 10.0.0.0/8, 1.2.3.4. allow-recursion-override=on|off By specifying allow-recursion-override, local data even about hosts that don't exist will override the internet. This allows you to generate zones that don't really exist on the internet. Does increase the number of SQL queries for hosts that truly don't exist, also not in your database. cache-ttl=... Seconds to store packets in the PacketCache. See Section 9.3.1. chroot=... If set, chroot to this directory for more security. See Chapter 7. config-dir=... Location of configuration directory (pdns.conf) config-name=... Name of this virtual configuration - will rename the binary image. See Chapter 8. control-console=... Debugging switch - don't use. daemon=... Operate as a daemon default-soa-name=... name to insert in the SOA record if none set in the backend disable-axfr=... Do not allow zone transfers. Before 2.9.10, this could be overridden by allow-axfr-ips. disable-tcp=... Do not listen to TCP queries. Breaks RFC compliance. distributor-threads=... Default number of Distributor (backend) threads to start. See Chapter 9. do-ipv6-additional-processing=... Perform AAAA additional processing. fancy-records=... Process URL and MBOXFW records. See Chapter 14. guardian | --guardian=yes | --guardian=no Run within a guardian process. See Section B.2. help Provide a helpful message launch=... Which backends to launch and order to query them in. See Section B.3. lazy-recursion=... On by default as of 2.1. Checks local data first before recursing. See Chapter 11. load-modules=... Load this module - supply absolute or relative path. See Section B.3. local-address=... Local IP address to which we bind. You can specify multiple addresses separated by commas or whitespace. It is highly advised to bind to specific interfaces and not use the default 'bind to any'. This causes big problems if you have multiple IP addresses. Unix does not provide a way of figuring out what IP address a packet was sent to when binding to any. local-ipv6=... Local IPv6 address to which we bind. You can specify multiple addresses separated by commas or whitespace. local-port=... The port on which we listen. Only one port possible. log-failed-updates=... If set to 'no', failed Windows Dynamic Updates will not be logged. log-dns-details=... If set to 'no', informative-only DNS details will not even be sent to syslog, improving performance. Available from 2.5 and onwards. logging-facility=... If set to a digit, logging is performed under this LOCAL facility. See Section 6.3. Available from 1.99.9 and onwards. Do not pass names like 'local0'! loglevel=... Amount of logging. Higher is more. Do not set below 3 master [,=on]. Turn on master support. Boolean. max-cache-entries Maximum number of cache entries. 1 million will generally suffice for most installations. Available since 2.9.22. max-queue-length=... If this many packets are waiting for database attention, consider the situation hopeless and respawn. max-tcp-connections=... Allow this many incoming TCP DNS connections simultaneously. module-dir=... Default directory for modules. See Section B.3. negquery-cache-ttl=... Seconds to store queries with no answer in the Query Cache. See Section 9.3.2. no-config Do not attempt to read the configuration file. no-shuffle Do not attempt to shuffle query results. server-id This is the server ID that will be returned on an EDNS NSID query. Defaults to the host name. out-of-zone-additional-processing | --out-of-zone-additional-processing=yes | --out-of-zone-additional-processing=no Do out of zone additional processing. This means that if a malicious user adds a '.com' zone to your server, it is not used for other domains and will not contaminate answers. Do not enable this setting if you run a public DNS service with untrusted users. Off by default. query-cache-ttl=... Seconds to store queries with an answer in the Query Cache. See Section 9.3.2. query-local-address=... The IP address to use as a source address for sending queries. Useful if you have multiple IPs and pdns is not bound to the IP address your operating system uses by default for outgoing packets. query-logging | query-logging=yes | query-logging=no Hints to a backend that it should log a textual representation of queries it performs. Can be set at runtime. queue-limit=... Maximum number of miliseconds to queue a query. See Chapter 9. recursive-cache-ttl=... Seconds to store recursive packets in the PacketCache. See Section 9.3.1. recursor=... If set, recursive queries will be handed to the recursor specified here. See Chapter 11. send-root-referral | --send-root-referral=yes | --send-root-referral=no | --send-root-referral=lean If set, PowerDNS will send out old-fashioned root-referrals when queried for domains for which it is not authoritative. Wastes some bandwidth but may solve incoming query floods if domains are delegated to you for which you are not authoritative, but which are queried by broken recursors. Available since 2.9.19. Since 2.9.21, it is possible to specify 'lean' root referrals, which waste less bandwidth. setgid=... If set, change group id to this gid for more security. See Chapter 7. setuid=... If set, change user id to this uid for more security. See Chapter 7. skip-cname | --skip-cname=yes | --skip-cname=no Do not perform CNAME indirection for each query. Has performance implications. See Chapter 7. slave-cycle-interval=60 Schedule slave up-to-date checks of domains whose status is unknown every .. seconds. See Chapter 14. smtpredirector=... Our smtpredir MX host. See Chapter 14. soa-expire-default=604800 Default SOA expire. soa-minimum-ttl=3600 Default SOA minimum ttl. soa-refresh-default=10800 Default SOA refresh. soa-retry-default=3600 Default SOA retry. soa-serial-offset=... If your database contains single-digit SOA serials and you need to host .DE domains, this setting can help placate their 6-digit SOA serial requirements. Suggested value is to set this to 1000000 which adds 1000000 to all SOA Serials under that offset. socket-dir=... Where the controlsocket will live. See Section B.1. strict-rfc-axfrs | --strict-rfc-axfrs=yes | --strict-rfc-axfrs=no Perform strictly RFC conformant AXFRs, which are slow, but needed to placate some old client tools. urlredirector=... Where we send hosts to that need to be url redirected. See Chapter 14. version-string=anonymous|powerdns|full|custom When queried for its version over DNS (dig chaos txt version.bind @pdns.ip.address), PowerDNS normally resonds truthfully. With this setting you can overrule what will be returned. Set the version-string to 'full' to get the default behaviour, to 'powerdns' to just make it state 'served by PowerDNS - http://www.powerdns.com'. The 'anonymous' setting will return a ServFail, much like Microsoft nameservers do. You can set this response to a custom value as well. webserver | --webserver=yes | --webserver=no Start a webserver for monitoring. See Chapter 6. webserver-address=... IP Address of webserver to listen on. See Chapter 6. webserver-password=... Password required for accessing the webserver. See Chapter 6. webserver-port=... Port of webserver to listen on. See Chapter 6. wildcard-url=... Check for wildcard URL records. wildcards=... Honor wildcards in the database. On by default. Turning this off has performance implications, see Chapter 9. __________________________________________________________ Chapter 16. Index of all Authoritative Server metrics 16.1. Counters & variables A number of counters and variables are set during PDNS Authoritative Server operation. These can be queried with the init.d dump, show and mrtg commands, or viewed with the webserver. __________________________________________________________ 16.1.1. Counters corrupt-packets Number of corrupt packets received latency Average number of microseconds a packet spends within PDNS packetcache-hit Number of packets which were answered out of the cache packetcache-miss Number of times a packet could not be answered out of the cache packetcache-size Amount of packets in the packetcache qsize-a Size of the queue before the transmitting socket. qsize-q Number of packets waiting for database attention servfail-packets Amount of packets that could not be answered due to database problems tcp-answers Number of answers sent out over TCP tcp-questions Number of questions received over TCP timedout-questions Amount of packets that were dropped because they had to wait too long internally udp-answers Number of answers sent out over UDP udp-questions Number of questions received over UDP __________________________________________________________ 16.1.2. Ring buffers Besides counters, PDNS also maintains the ringbuffers. A ringbuffer records events, each new event gets a place in the buffer until it is full. When full, earlier entries get overwritten, hence the name 'ring'. By counting the entries in the buffer, statistics can be generated. These statistics can currently only be viewed using the webserver and are in fact not even collected without the webserver running. The following ringbuffers are available: Log messages (logmessages) All messages logged Queries for existing records but for a type we don't have (noerror-queries) Queries for, say, the AAAA record of a domain, when only an A is available. Queries are listed in the following format: name/type. So an AAA query for pdns.powerdns.com looks like pdns.powerdns.com/AAAA. Queries for non-existing records within existing domains(nxdomain-queries) If PDNS knows it is authoritative over a domain, and it sees a question for a record in that domain that does not exist, it is able to send out an authoritative 'no such domain' message. Indicates that hosts are trying to connect to services really not in your zone. UDP queries received (udp-queries) All UDP queries seen. Remote server IP addresses (remotes) Hosts querying PDNS. Be aware that UDP is anonymous - person A can send queries that appear to be coming from person B. Remotes sending corrupt packets (remote-corrupts) Hosts sending PDNS broken packets, possibly meant to disrupt service. Be aware that UDP is anonymous - person A can send queries that appear to be coming from person B. Remotes querying domains for which we are not auth (remote-unauth) It may happen that there are misconfigured hosts on the internet which are configured to think that a PDNS installation is in fact a resolving nameserver. These hosts will not get useful answers from PDNS. This buffer lists hosts sending queries for domains which PDNS does not know about. Queries that could not be answered due to backend errors (servfail-queries) For one reason or another, a backend may be unable to extract answers for a certain domain from its storage. This may be due to a corrupt database or to inconsistent data. When this happens, PDNS sends out a 'servfail' packet indicating that it was unable to answer the question. This buffer shows which queries have been causing servfails. Queries for domains that we are not authoritative for (unauth-queries) If a domain is delegated to a PDNS instance, but the backend is not made aware of this fact, questions come in for which no answer is available, nor is the authority. Use this ringbuffer to spot such queries. __________________________________________________________ Chapter 17. Supported record types and their storage This chapter lists all record types PDNS supports, and how they are stored in backends. The list is mostly alphabetical but some types are grouped. The PowerDNS Recursor can serve and store all record types, regardless of wether these are explicitly supported. A The A record contains an IP address. It is stored as a decimal dotted quad string, for example: '213.244.168.210'. AAAA The AAAA record contains an IPv6 address. An example: '3ffe:8114:2000:bf0::1'. AFSDB (since 2.9.21) Specialised record type for the 'Andrew Filesystem'. Stored as: '#subtype hostname', where subtype is a number. CERT (since 2.9.21) Specialised record type for storing certificates, defined in RFC 2538. CNAME The CNAME record specifies the canonical name of a record. It is stored plainly. Like all other records, it is not terminated by a dot. A sample might be 'webserver-01.yourcompany.com'. DNSKEY (since 2.9.21) The DNSKEY DNSSEC record type is fully supported, as described in RFC 3757. Note that while PowerDNS can store, retrieve and serve DNSSEC records, no further DNSSEC processing is performed. DS (since 2.9.21) The DS DNSSEC record type is fully supported, as described in RFC 3757. Note that while PowerDNS can store, retrieve and serve DNSSEC records, no further DNSSEC processing is performed. HINFO Hardware Info record, used to specify CPU and operating system. Stored with a single space separating these two, example: 'i386 Linux'. KEY (since 2.9.21) The KEY record is fully supported. For its syntax, see RFC 2535. LOC The LOC record is fully supported. For its syntax, see RFC 1876. A sample content would be: '51 56 0.123 N 5 54 0.000 E 4.00m 1.00m 10000.00m 10.00m' MX The MX record specifies a mail exchanger host for a domain. Each mail exchanger also has a priority or preference. This should be specified in the separate field dedicated for that purpose, often called 'prio'. NAPTR Naming Authority Pointer, RFC 2915. Stored as follows: '100 50 "s" "z3950+I2L+I2C" "" _z3950._tcp.gatech. edu'. The fields are: order, preference, flags, service, regex, replacement. Note that the replacement is not enclosed in quotes, and should not be. The replacement may be omitted, in which case it is empty. See also RFC 2916 for how to use NAPTR for ENUM (E.164) purposes. NS Nameserver record. Specifies nameservers for a domain. Stored plainly: 'ns1.powerdns.com', as always without a terminating dot. NSEC (since 2.9.21) The NSEC DNSSEC record type is fully supported, as described in RFC 3757. Note that while PowerDNS can store, retrieve and serve DNSSEC records, no further DNSSEC processing is performed. PTR Reverse pointer, used to specify the host name belonging to an IP or IPv6 address. Name is stored plainly: 'www.powerdns.com'. As always, no terminating dot. RP Responsible Person record, as described in RFC 1183. Stored with a single space between the mailbox name and the more-information pointer. Example 'peter.powerdns.com peter.people.powerdns.com', to indicate that peter@powerdns.com is responsible and that more information about peter is available by querying the TXT record of peter.people.powerdns.com. RRSIG (since 2.9.21) The RRSIG DNSSEC record type is fully supported, as described in RFC 3757. Note that while PowerDNS can store, retrieve and serve DNSSEC records, no further DNSSEC processing is performed. SOA The Start of Authority record is one of the most complex available. It specifies a lot about a domain: the name of the master nameserver ('the primary'), the hostmaster and a set of numbers indicating how the data in this domain expires and how often it needs to be checked. Further more, it contains a serial number which should rise on each change of the domain. The stored format is: primary hostmaster serial refresh retry expire default_t tl Besides the primary and the hostmaster, all fields are numerical. PDNS has a set of default values: Table 17-1. SOA fields primary default-soa-name configuration option hostmaster hostmaster@domain-name serial 0 refresh 10800 (3 hours) retry 3600 (1 hour) expire 604800 (1 week) default_ttl 3600 (1 hour) The fields have complicated and sometimes controversial meanings. The 'serial' field is special. If left at 0, the default, PDNS will perform an internal list of the domain to determine highest change_date field of all records within the zone, and use that as the zone serial number. This means that the serial number is always raised when changes are made to the zone, as long as the change_date field is being set. SPF (since 2.9.21) SPF records can be used to store Sender Permitted From details. SSHFP (since 2.9.21) The SSHFP record type, used for storing Secure Shell (SSH) fingerprints, is fully supported. A sample from RFC 4255 is: '2 1 123456789abcdef67890123456789abcdef67890'. SRV SRV records can be used to encode the location and port of services on a domain name. When encoding, the priority field is used to encode the priority. For example, '_ldap._tcp.dc._msdcs.conaxis.ch SRV 0 100 389 mars.conaxis.ch' would be encoded with 0 in the priority field and '100 389 mars.conaxis.ch' in the content field. TXT The TXT field can be used to attach textual data to a domain. Text is stored plainly. __________________________________________________________ Chapter 18. HOWTO & Frequently Asked Questions This chapter contains a number of FAQs and HOWTOs. __________________________________________________________ 18.1. Getting support, free and paid FAQ PowerDNS is an open source program so you may get help from the PowerDNS users' community or from its authors. You may also help others (please do). The PowerDNS company provides free support on the public mailing lists, and can help or support you in private as well. For first class and rapid support, please contact pdns.bd@trilab.com. More information about the PowerDNS community, and its mailing lists, can be found on its Wiki. Below, please find a list of common questions asked on our public mailing lists. Q: Help! A: Please try harder :-) Specifically, before people will be able to help you, they need to know a lot about your system. If you list more details, chances are you'll get better answers. Q: I have a question, what details should I supply? A: Start out with stating what you think should be happening. Quite often, wrong expectations are the actual proble. Furthermore, which database backend you use, your operating system, which version of PowerDNS you use and where you got it from (RPM, .DEB, tar.gz). If you compiled it yourself, what were the ./configure parameters. If at *all* possible, supply the actual name of your domain and the IP address of your server(s). Q: Where should I send my question? A: To a mailinglist. Please email the authors directly only if you previously entered a support contract with them, or are considering doing so. For mailing list details, see the mailinglists page. Questions about using PowerDNS should be sent to the pdns-users list, questions about compiler errors or feature requests to pdns-dev. Before posting, read all FAQs. Q: My information is confidential, must I send it to the mailing list? If you desire privacy, please consider entering a support relationship with us, in which case you can email . __________________________________________________________ 18.2. Using and Compiling PowerDNS FAQ In the course of compiling and using PowerDNS, many questions may arise. Here are some we've heard earlier or questions we expect people may have. Please read this list before mailing us! If you don't see your question answered here, please check out the Wiki FAQ, but do note that it is user-editable and not under our constant control. Q: I get this entry a lot of times in my log file: Authoritative empty NO ERROR to 1.2.3.4 for 'powerdns.nl' (AAAA).. As the name implies, this is not an error. It tells you there are questions for a domain which exists in your database, but for which no record of the requested type exists. To get rid of this error, add log-dns-details=off to your configuration. Q: Can I launch multiple backends simultaneously? A: You can. This might for example be useful to keep an existing BIND configuration around but to store new zones in, say MySQL. The syntax to use is 'launch=bind,gmysql'. Q: PowerDNS does not give authoritative answers, how come? A: This is almost always not the case. An authoritative answer is recognized by the 'AA' bit being set. Many tools prominently print the number of Authority records included in an answer, leading users to conclude that the absence or presence of these records indicates the authority of an answer. This is not the case. Verily, many misguided country code domain operators have fallen into this trap and demand authority records, even though these are fluff and quite often misleading. Invite such operators to look at section 6.2.1 of RFC 1034, which shows a correct authoritative answer without authority records. In fact, none of the non-deprecated authoritative answers shown have authority records! Sorry for sounding like DJB on this, but we get so many misguided questions about authority.. Q: Which backend should I use? There are so many! A: If you have no external constraints, the Generic MySQL (gmysql) and Generic PostgreSQL (gpgsql) ones are probably the most used and complete. By all means do not use the non-generic MySQL backend, which is deprecated and only available for older installations. The Oracle backend also has happy users, we know of no deployments of the DB2 backend. The BIND backend is pretty capable too in fact, but many prefer a relational database. Q: I'm trying to build from SVN but I get lots of weird errors! A: Read the 'HACKING' file, it lists the build requirements (mostly autoconf, automake, libtool). In many cases, it may be easier to build from the source distribution though. More information for developers is available on the PowerDNS Open Source Community Wiki. Q: When compiling I get errors about 'sstream' and 'ostringstream', or BITSPERCHAR A: Your gcc is too old. Versions 2.95.2 and older are not supported. Many distributions have improved gcc 2.95.2 with an ostringstream implementation, in which case their 2.95.2 is also supported. Q: PowerDNS crashes when I install the pdns-static .deb on Debian SID A: Indeed. Install the .debs that come with Debian or recompile PowerDNS yourself. If not using MySQL, the crashes will go away if you remove setuid and setgid statements from the configuration. Q: Why don't my slaves act on notifications and transfer my updated zone? A: Raise the serial number of your zone. In most backends, this is the first digit of the SOA contents field. If this number is lower to equal to that on a slave, it will not consider your zone updated. Q: Master or Slave support is not working, PDNS is not picking up changes A: The Master/Slave apparatus is off by default. Turn it on by adding a slave and/or master statement to the configuration file. Also, check that the configured backend is master or slave capable. Q: My masters won't allow PowerDNS to access zones as it is using the wrong local IP address A: Mark Bergsma contributed the query-local-address setting to tell PowerDNS which local IP address to use. Q: I compiled PowerDNS myself and I see weird problems, especially on SMP A: There are known issues between gcc <3.2 and PowerDNS on Linux SMP systems. The exact cause is not known but moving to our precompiled version always fixes the problems. If you compile yourself, use a recent gcc! Q: I see this a lot: Backend error: Failed to execute mysql_query, perhaps connection died? A: Check your MySQL timeout, it may be set too low. This can be changed in the my.cnf file. Q: PowerDNS does not answer queries on all my IP addresses and I've ignored the warning I got about that at startup A: Please don't ignore what PowerDNS says to you. Furthermore, read Chapter 15 about the local-address setting, and use it to specify which IP addresses PowerDNS should listen on. Q: Can I use a MySQL database with the Windows version of PowerDNS? A: You can. MySQL support is supplied through the ODBC backend, which is compiled into the main binary. So if you want to use MySQL you can change the pdns.conf file, which is located in the PowerDNS for Windows directory, to use the correct ODBC data sources. If you don't know how to use ODBC with MySQL: + Download MyODBC from http://www.mysql.com/ + Install the MySQL ODBC driver. Then you can follow the instructions located in Chapter 3. But instead of selecting the Microsoft Access Driver you select the MySQL ODBC Driver and configure it to use your MySQL database. Note For other databases for which an ODBC driver is available, the procedure is the same as this example. __________________________________________________________ 18.3. Backend developer HOWTO Writing backends without access to the full PDNS source means that you need to write code that can be loaded by PDNS at runtime. This in turn means that you need to use the same compiler that we do. Furthermore, your pdns_server executable must be dynamically linked. The default .rpm PDNS contains a static binary so you need to retrieve the dynamic rpm or the dynamic tar.gz or the Debian unstable ('Woody') deb. FreeBSD dynamic releases are forthcoming. Q: Will PDNS drivers work with other PDNS versions than they were compiled for? A: 'Probably'. We make no guarantees. Efforts have been made to keep the interface between the backend and PDNS as thin as possible. For example, a backend compiled with the 1.99.11 backend development kit works with 1.99.10. But don't count on it. We will notify when we think an incompatible API change has occured but you are best off recompiling your driver for each new PDNS release. Q: What is in that DNSPacket * pointer passed to lookup! A: For reasons outlined above, you should treat that pointer as opaque and only access it via the getRemote() functions made available and documented above. The DNSPacket class changes a lot and this level of indirection allows for greater changes to be made without changing the API to the backend coder. Q: How is the PowerDNS Open Source Backend Development Kit licensed? A: MIT X11, a very liberal license permitting basically everything. Q: Can I release the backend I wrote? A: Please do! If you tell us about it we will list you on our page. Q: Can I sell backends I wrote? A: You can. Again, if you tell us about them we will list your backend on the site. You can keep the source of your backend secret if you want, or you can share it with the world under any license of your chosing. Q: Will PowerDNS use my code in the PDNS distribution? A: If your license permits it and we like your backend, we sure will. If your license does not permit it but we like your backend anyway we may contact you. Q: My backend compiles but when I try to load it, it says 'undefined symbol: BackendMakers__Fv' A: You compiled with the wrong GCC. Use GCC 3.x for Linux, 2.95.x for FreeBSD. You may want to change g++ to g++-3.0 in the Makefile, or change your path so that 3.x is used. Q: I downloaded a dynamic copy of pdns_server but it doesn't run, even without my backend A: Run 'ldd' on the pdns_server binary and figure out what libraries you are missing. Most likely you need to install gcc 3.0 libraries, RedHat 7.1 and 7.2 have packages available, Debian installs these by default if you use the 'unstable deb' of PDNS. Q: What is this 'AhuException' I keep reading about? A: This name has historical reasons and has no significance. Q: I need a backend but I can't write it, can you help? A: Yes, we also do custom development. Contact us at pdns.bd@trilab.com. __________________________________________________________ 18.4. About PowerDNS.COM BV, 'the company' As of 25 November 2002, the PowerDNS nameserver and its modules are open source. This has led to a lot of questions on the future of both PowerDNS, the company and the products. This FAQ attempts to address these questions. Q: Is PowerDNS 2.9 really open source? What license? A: PowerDNS 2.9 is licensed under the GNU General Public License version two, the same license that covers the Linux kernel. Q: Is the open source version crippled? A: It is not. Not a single byte has been omitted. Q: Is the nameserver abandoned? A: Far from it. In fact, we expect development to speed up now that we have joined the open source community. Q: Can I buy support contracts for PowerDNS? Sure, to do so, please contact us at Q: Will you accept patches? We've added a feature Probably - in general, it is best to discuss your intentions and needs on the (subscribe) mailinglist before doing the work. We may have suggestions or guidelines on how you should implement the feature. Q: PowerDNS doesn't work on my platform, will you port it?, Q: PowerDNS doesn't have feature I need, will you add it? Be sure to ask on the (subscribe) mailinglist. You can even hire us to do work on PowerDNS if plain asking is not persuasive enough. This might be the case if we don't currently have time for your feature, but you need it quickly anyhow, and are not in a position to submit a patch implementing it. Q: Will PowerDNS Express be open sourced? Perhaps, we're not yet sure. Q: We are a Linux/Unix vendor, can we include PowerDNS? A: Please do. In fact, we'd be very happy to work with you to make this happen. Contact if you have specific upstream needs. __________________________________________________________ Chapter 19. Other tools included with PowerDNS PowerDNS comes with several tools that can be used to do various DNS related things. __________________________________________________________ 19.1. Notification proxy (nproxy) Available in PowerDNS 2.9.22 and later. For additional security, operators may prefer to have a 'hidden slave' that sits behind a strong firewall. This slave pulls in zones from the outside world, and stores them in a database. This database is then used by publicly accessible nameservers to publish zone data. For proper slave operation, master nameservers send out notifications to inform slaves of updates. This is not normally a problem, but when operating with a hidden slave behind a firewall, notification packets can't reach the slave. For this purpose, the PowerDNS also supplies a notification proxy. It sits outside the firewall, and accepts notifications from remote master servers. It interprets and validates these packets, and then sends on a new notification to the hidden slave. The hidden slave then promptly retrieves an updated zone from the master. The notification proxy, called nproxy, can be configured using the following settings: chroot Change root to this directory for additional security. daemon Run in the background. Defaults to true, can be turned off using '--daemon=no'. listen-address Public addresses (IPv4 and IPv6) to listen on for incoming notification packets. Defaults to "all addresses", but it is highly recommended to specify addresses here. origin-address Can be used to pin the address the nproxy uses to communicate with the hidden slave. Highly recommended. Corresponds to the PowerDNS settting trusted-notification-proxy. powerdns-address IP address (IPv4 or IPv6) of the hidden slave, to which notifications should be relayed. This setting is mandatory, and has no default. setuid and setgid Change to these numerical user-id and/or group-id, dropping root privileges, for additional security. __________________________________________________________ Chapter 20. Tools to analyse DNS traffic DNS is highly mission critical, it is therefore necessary to be able to study and compare DNS traffic. Since 2.9.18, PowerDNS comes with three tools to aid in analysis: Warning As of 2.9.18 these tools are somewhat rough - they have no help messages for example. They do work though. dnsreplay pcapfile [ipaddress] [port number] This program takes recorded questions and answers and replays them to a specified nameserver and reporting afterwards which percentage of answers matched, were worse or better. dnswasher pcapfile output Anonymises recorded traffic, making sure it only contains DNS, and that the originating IP addresses of queries are stripped, which may allow you to send traces to our company or mailing list without violating obligations towards your customers or privacy laws. dnsscope pcapfile Calculates statistics without replaying traffic __________________________________________________________ Appendix A. Backends in detail This appendix lists several of the available backends in more detail __________________________________________________________ A.1. PipeBackend Table A-1. PipeBackend capabilities Native Yes Master No Slave No Superslave No Autoserial No Case Depends Module name pipe Launch name pipe The PipeBackend allows for easy dynamic resolution based on a 'Coprocess' which can be written in any programming language that can read a question on standard input and answer on standard output. To configure, the following settings are available: pipe-command Command to launch as backend. Mandatory. pipe-timeout Number of milliseconds to wait for an answer from the backend. If this time is ever exceeded, the backend is declared dead and a new process is spawned. Available since 2.7. pipe-regex If set, only questions matching this regular expression are even sent to the backend. This makes sure that most of PowerDNS does not slow down if you you reploy a slow backend. A query for the A record of 'www.powerdns.com' would be presented to the regex as 'www.powerdns.com;A'. A matching regex would be '^www.powerdns.com;.*$'. To match only ANY and A queries for www.powerdns.com, use '^www.powerdns.com;(A|ANY)$'. Please be aware that the single quotes used in this document should not be present in the configuration file, and only on the command line. In the configuration file, the previous example would be stored as: pipe-regex=^www.powerdns.com;(A|ANY)$ Available since 2.8. pipebackend-abi-version This is the version of the question format that is sent to the co-process (pipe-command) for the pipe backend. If not set the default pipebackend-abi-version is 1. When set to 2, the local-ip-address field is added after the remote-ip-address. (the local-ip-address refers to the IP address the question was received on) __________________________________________________________ A.1.1. PipeBackend protocol Questions come in over a file descriptor, by default standard input. Answers are sent out over another file descriptor, standard output by default. __________________________________________________________ A.1.1.1. Handshake PowerDNS sends out 'HELO\t1', indicating that it wants to speak the protocol as defined in this document, version 1. A PowerDNS CoProcess must then send out a banner, prefixed by 'OK\t', indicating it launched successfully. If it does not support the indicated version, it should respond with FAIL, but not exit. Suggested behaviour is to try and read a further line, and wait to be terminated. __________________________________________________________ A.1.1.2. Questions Questions come in three forms and are prefixed by a tag indicating the kind: Q Regular queries AXFR List requests, which mean that an entire zone should be listed PING Check if the coprocess is functioning The question format: pipebackend-abi-version = 1 [default] type qname qclass qtype id remote-ip-address pipebackend-abi-version = 2 type qname qclass qtype id remote-ip-address local-ip-address Fields are tab separated, and terminated with a single \n. The remote-ip-address is the IP address of the nameserver asking the question; the local-ip-address is the IP address on which the question was received. Type is the tag above, qname is the domain the question is about. qclass is always 'IN' currently, denoting an INternet question. qtype is the kind of information desired, the record type, like A, CNAME or AAAA. id can be specified to help your backend find an answer if the id is already known from an earlier query. You can ignore it. remote-ip-address is the ip-address of the nameserver asking the question. local-ip-address is the ip-address that was querried locally. __________________________________________________________ A.1.1.3. Answers Each answer starts with a tag, possibly followed by a TAB and more data. DATA Indicating a successful line of DATA END Indicating the end of an answer - no further data FAIL Indicating a lookup failure. Also serves as 'END'. No further data. LOG For specifying things that should be logged. Can only be sent after a query and before an END line. After the tab, the message to be logged So letting it be known that there is no data consists if sending 'END' without anything else. The answer format: DATA qname qclass qtype ttl id content 'content' is as specified in Chapter 17. A sample dialogue may look like this: Q www.ds9a.nl IN CNAME -1 213.244.168.210 DATA www.ds9a.nl IN CNAME 3600 1 ws1.ds9a.nl Q ws1.ds9a.nl IN CNAME -1 213.244.168.210 END Q wd1.ds9a.nl IN A -1 213.244.168.210 DATA ws1.ds9a.nl IN A 3600 1 1.2.3.4 DATA ws1.ds9a.nl IN A 3600 1 1.2.3.5 DATA ws1.ds9a.nl IN A 3600 1 1.2.3.6 END This would correspond to a remote webserver 213.244.168.210 wanting to resolve the IP address of www.ds9a.nl, and PowerDNS traversing the CNAMEs to find the IP addresses of ws1.ds9a.nl Another dialogue might be: Q ds9a.nl IN SOA -1 213.244.168.210 DATA ds9a.nl IN SOA 86400 1 ahu.ds9a.nl ... END AXFR 1 DATA ds9a.nl IN SOA 86400 1 ahu.ds9a.nl ... DATA ds9a.nl IN NS 86400 1 ns1.ds9a.nl DATA ds9a.nl IN NS 86400 1 ns2.ds9a.nl DATA ns1.ds9a.nl IN A 86400 1 213.244.168.210 DATA ns2.ds9a.nl IN A 86400 1 63.123.33.135 . . END This is a typical zone transfer. __________________________________________________________ A.1.1.4. Sample perl backend #!/usr/bin/perl -w # sample PowerDNS Coprocess backend # use strict; $|=1; # no buffering my $line=<>; chomp($line); unless($line eq "HELO\t1") { print "FAIL\n"; print STDERR "Recevied '$line'\n"; <>; exit; } print "OK Sample backend firing up\n"; # print our banner while(<>) { print STDERR "$$ Received: $_"; chomp(); my @arr=split(/\t/); if(@arr<6) { print "LOG PowerDNS sent unparseable line\n"; print "FAIL\n"; next; } my ($type,$qname,$qclass,$qtype,$id,$ip)=split(/\t/); if(($qtype eq "A" || $qtype eq "ANY") && $qname eq "webserver.ex ample.com") { print STDERR "$$ Sent A records\n"; print "DATA $qname $qclass A 3600 -1 1.2.3.4\n"; print "DATA $qname $qclass A 3600 -1 1.2.3.5\n"; print "DATA $qname $qclass A 3600 -1 1.2.3.6\n"; } elsif(($qtype eq "CNAME" || $qtype eq "ANY") && $qname eq "www.e xample.com") { print STDERR "$$ Sent CNAME records\n"; print "DATA $qname $qclass CNAME 3600 -1 webserver.example.com\n"; } elsif($qtype eq "MBOXFW") { print STDERR "$$ Sent MBOXFW records\n"; print "DATA $qname $qclass MBOXFW 3600 -1 powerdns\@example.com\n"; } print STDERR "$$ End of data\n"; print "END\n"; } __________________________________________________________ A.2. MySQL backend Warning This backend is deprecated! Use the Generic MySQL backend which is better in all respects. It does support master/slave operation, this backend does not. See Section A.5. So stop reading here unless you already have a database filled with 'mysql' records. Table A-2. MySQL backend capabilities Native Yes Master No Slave Yes Superslave No Autoserial Yes Case Insensitive Module name mysql Launch name mysql The MySQL Backend as present in PDNS is fixed - it requires a certain database schema to function. This schema corresponds to this create statement: CREATE TABLE records ( id int(11) NOT NULL auto_increment, domain_id int(11) NOT NULL, name varchar(255) NOT NULL, type varchar(6) NOT NULL, content varchar(255) NOT NULL, ttl int(11) NOT NULL, prio int(11) default NULL, change_date int(11) default NULL, PRIMARY KEY (id), KEY name_index(name), KEY nametype_index(name,type), KEY domainid_index(domain_id) ); Every domain should have a unique domain_id, which should remain identical for all records in a domain. Records with a domain_id that differs from that in the domain SOA record will not appear in a zone transfer. The change_date may optionally be updated to the time_t (the number of seconds since midnight UTC at the start of 1970), and is in that case used to auto calculate the SOA serial number in case that is unspecified. __________________________________________________________ A.2.1. Configuration settings WARNING! Make sure that you can actually resolve the hostname of your database without accessing the database! It is advised to supply an IP address here to prevent chicken/egg problems! mysql-dbname Database name to connect to mysql-host Database host to connect to mysql-password Password to connect with mysql-socket MySQL socket to use for connecting mysql-table MySQL table name. Defaults to 'records'. mysql-user MySQL user to connect as __________________________________________________________ A.2.2. Notes It has been observed that InnoDB tables outperform the default MyISAM tables by a large margin. Furthermore, the default number of backends (3) should be raised to 10 or 15 for busy servers. __________________________________________________________ A.3. Random Backend Table A-3. Random Backend capabilities Native Yes Master No Slave No Superslave No Autoserial No Case Depends Module name built in Launch name random This is a very silly backend which is discussed in Section C.1 as a demonstration on how to write a PowerDNS backend. This backend knows about only one hostname, and only about its IP address at that. With every query, a new random IP address is generated. It only makes sense to load the random backend in combination with a regular backend. This can be done by prepending it to the launch= instruction, such as launch=random,gmysql. Variables: random-hostname Hostname for which to supply a random IP address. __________________________________________________________ A.4. MySQL PDNS backend Table A-4. MySQL backend capabilities Native Yes Master No Slave No Superslave No Autoserial Yes Case Insensitive Module name pdns Launch name pdns This is the driver that corresponds to the set of XML-RPC tools available from PowerDNS. The schema: CREATE TABLE MailForwards ( Id int(10) unsigned NOT NULL auto_increment, ZoneId int(10) unsigned NOT NULL default '0', Name varchar(255) NOT NULL default '', Destination varchar(255) NOT NULL default '', Flags int(11) NOT NULL default '0', ChangeDate timestamp(14) NOT NULL, CreateDate timestamp(14) NOT NULL, Active tinyint(4) NOT NULL default '0', PRIMARY KEY (Id), KEY NameIndex (Name), KEY ZoneIdIndex (ZoneId) ); -- -- Table structure for table 'Mailboxes' -- CREATE TABLE Mailboxes ( Id int(10) unsigned NOT NULL auto_increment, ZoneId int(10) unsigned NOT NULL default '0', Name varchar(255) NOT NULL default '', Password varchar(255) NOT NULL default '', Quota int(10) unsigned NOT NULL default '0', Flags int(11) NOT NULL default '0', ChangeDate timestamp(14) NOT NULL, CreateDate timestamp(14) NOT NULL, Active tinyint(4) NOT NULL default '0', PRIMARY KEY (Id), UNIQUE KEY Name (Name), KEY ZoneIdIndex (ZoneId), KEY NameIndex (Name) ); -- -- Table structure for table 'Records' -- CREATE TABLE Records ( Id int(10) unsigned NOT NULL auto_increment, ZoneId int(10) unsigned NOT NULL default '0', Name varchar(255) NOT NULL default '', Type varchar(8) NOT NULL default '', Content varchar(255) NOT NULL default '', TimeToLive int(11) NOT NULL default '60', Priority int(11) NOT NULL default '0', Flags int(11) NOT NULL default '0', ChangeDate timestamp(14) NOT NULL, CreateDate timestamp(14) NOT NULL, Active tinyint(4) NOT NULL default '0', PRIMARY KEY (Id), KEY NameIndex (Name) ); -- -- Table structure for table 'WebForwards' -- CREATE TABLE WebForwards ( Id int(10) unsigned NOT NULL auto_increment, ZoneId int(10) unsigned NOT NULL default '0', Name varchar(255) NOT NULL default '', Destination varchar(255) NOT NULL default '', Type varchar(7) NOT NULL default 'NORMAL', Title varchar(255) NOT NULL default '', Description varchar(255) NOT NULL default '', Keywords varchar(255) NOT NULL default '', FavIcon varchar(255) NOT NULL default '', Flags int(11) NOT NULL default '0', ChangeDate timestamp(14) NOT NULL, CreateDate timestamp(14) NOT NULL, Active tinyint(4) NOT NULL default '0', PRIMARY KEY (Id), KEY NameIndex (Name), KEY ZoneIdIndex (ZoneId) ); -- -- Table structure for table 'Zones' -- CREATE TABLE Zones ( Id int(10) unsigned NOT NULL auto_increment, Name varchar(255) NOT NULL default '', Hostmaster varchar(255) NOT NULL default '', Serial int(10) unsigned NOT NULL default '0', AutoSerial tinyint(4) NOT NULL default '0', Flags int(11) NOT NULL default '0', ChangeDate timestamp(14) NOT NULL, CreateDate timestamp(14) NOT NULL, Active tinyint(4) NOT NULL default '0', TimeToLive int(11) NOT NULL default '0', OwnerId varchar(255) NOT NULL default '', Master varchar(255) NOT NULL default '', PRIMARY KEY (Id), UNIQUE KEY Name (Name), KEY NameIndex (Name) ); It takes a number of parameters: pdns-dbname Database name to connect to pdns-host Database host to connect to pdns-password Password to connect with pdns-socket MySQL socket to use for connecting pdns-user MySQL user to connect as pdns-soa-refresh Pdns SOA refresh in seconds pdns-max-slave-records Maximal records to transfer __________________________________________________________ A.4.1. Notes It has been observed that InnoDB tables outperform the default MyISAM tables by a large margin. Furthermore, the default number of backends (3) should be raised to 10 or 15 for busy servers. __________________________________________________________ A.5. Generic MySQL and PgSQL backends Table A-5. Generic PgSQL and MySQL backend capabilities Native Yes - but PostgreSQL does not replicate Master Yes Slave Yes Superslave Yes Autoserial NO Case All lower Module name < 2.9.3 pgmysql Module name > 2.9.2 gmysql and gpgsql Launch name gmysql and gpgsql2 and gpgsql PostgreSQL and MySQL backend with easily configurable SQL statements, allowing you to graft PDNS on any PostgreSQL or MySQL database of your choosing. Because all database schemas will be different, a generic backend is needed to cover all needs. The template queries are expanded using the C function 'snprintf' which implies that substitutions are performed on the basis of %-place holders. To place a % in a query which will not be substituted, use %%. Make sure to fill out the search key, often called 'name' in lower case! There are in fact two backends, one for PostgreSQL and one for MySQL but they accept the same settings and use almost exactly the same database schema. __________________________________________________________ A.5.1. MySQL specifics Warning If using MySQL with 'slave' support enabled in PowerDNS you must run MySQL with a table engine that supports transactions. In practice, great results are achieved with the 'InnoDB' tables. PowerDNS will silently function with non-transaction aware MySQLs but at one point this is going to harm your database, for example when an incoming zone transfer fails. The default setup conforms to the following schema: create table domains ( id INT auto_increment, name VARCHAR(255) NOT NULL, master VARCHAR(128) DEFAULT NULL, last_check INT DEFAULT NULL, type VARCHAR(6) NOT NULL, notified_serial INT DEFAULT NULL, account VARCHAR(40) DEFAULT NULL, primary key (id) )type=InnoDB; CREATE UNIQUE INDEX name_index ON domains(name); CREATE TABLE records ( id INT auto_increment, domain_id INT DEFAULT NULL, name VARCHAR(255) DEFAULT NULL, type VARCHAR(6) DEFAULT NULL, content VARCHAR(255) DEFAULT NULL, ttl INT DEFAULT NULL, prio INT DEFAULT NULL, change_date INT DEFAULT NULL, primary key(id) )type=InnoDB; CREATE INDEX rec_name_index ON records(name); CREATE INDEX nametype_index ON records(name,type); CREATE INDEX domain_id ON records(domain_id); create table supermasters ( ip VARCHAR(25) NOT NULL, nameserver VARCHAR(255) NOT NULL, account VARCHAR(40) DEFAULT NULL ); GRANT SELECT ON supermasters TO pdns; GRANT ALL ON domains TO pdns; GRANT ALL ON records TO pdns; Zone2sql with the --gmysql flag also assumes this layout is in place. This schema contains all elements needed for master, slave and superslave operation. Depending on which features will be used, the 'GRANT' statements can be trimmed to make sure PDNS cannot subvert the contents of your database. When using the InnoDB storage engine, we suggest adding the following lines to the 'create table records' command above: CONSTRAINT `records_ibfk_1` FOREIGN KEY (`domain_id`) REFERENCES `domain s` (`id`) ON DELETE CASCADE This automates deletion of records on deletion of a domain from the domains table. __________________________________________________________ A.5.2. PostgresSQL specifics The default setup conforms to the following schema, which you should add to a PostgreSQL database. create table domains ( id SERIAL PRIMARY KEY, name VARCHAR(255) NOT NULL, master VARCHAR(128) DEFAULT NULL, last_check INT DEFAULT NULL, type VARCHAR(6) NOT NULL, notified_serial INT DEFAULT NULL, account VARCHAR(40) DEFAULT NULL ); CREATE UNIQUE INDEX name_index ON domains(name); CREATE TABLE records ( id SERIAL PRIMARY KEY, domain_id INT DEFAULT NULL, name VARCHAR(255) DEFAULT NULL, type VARCHAR(6) DEFAULT NULL, content VARCHAR(255) DEFAULT NULL, ttl INT DEFAULT NULL, prio INT DEFAULT NULL, change_date INT DEFAULT NULL, CONSTRAINT domain_exists FOREIGN KEY(domain_id) REFERENCES domains(id) ON DELETE CASCADE ); CREATE INDEX rec_name_index ON records(name); CREATE INDEX nametype_index ON records(name,type); CREATE INDEX domain_id ON records(domain_id); create table supermasters ( ip VARCHAR(25) NOT NULL, nameserver VARCHAR(255) NOT NULL, account VARCHAR(40) DEFAULT NULL ); GRANT SELECT ON supermasters TO pdns; GRANT ALL ON domains TO pdns; GRANT ALL ON domains_id_seq TO pdns; GRANT ALL ON records TO pdns; GRANT ALL ON records_id_seq TO pdns; This schema contains all elements needed for master, slave and superslave operation. Depending on which features will be used, the 'GRANT' statements can be trimmed to make sure PDNS cannot subvert the contents of your database. Zone2sql with the --gpgsql flag also assumes this layout is in place. With PostgreSQL, you may have to run 'createdb powerdns' first and then connect to that database with 'psql powerdns', and feed it the schema above. __________________________________________________________ A.5.3. Oracle specifics Generic Oracle support is only available since version 2.9.18. The default setup conforms to the following schema, which you should add to an Oracle database. You may need or want to add 'namespace' statements. create table domains ( id NUMBER, name VARCHAR(255) NOT NULL, master VARCHAR(128) DEFAULT NULL, last_check INT DEFAULT NULL, type VARCHAR(6) NOT NULL, notified_serial INT DEFAULT NULL, account VARCHAR(40) DEFAULT NULL, primary key (id) ); create sequence DOMAINS_ID_SEQUENCE; create index DOMAINS$NAME on Domains (NAME); CREATE TABLE records ( id number(11) not NULL, domain_id INT DEFAULT NULL REFERENCES Domains(ID) ON DELET E CASCADE, name VARCHAR(255) DEFAULT NULL, type VARCHAR(6) DEFAULT NULL, content VARCHAR(255) DEFAULT NULL, ttl INT DEFAULT NULL, prio INT DEFAULT NULL, change_date INT DEFAULT NULL, primary key (id) ); create index RECORDS$NAME on RECORDS (NAME); create sequence RECORDS_ID_SEQUENCE; create table supermasters ( ip VARCHAR(25) NOT NULL, nameserver VARCHAR(255) NOT NULL, account VARCHAR(40) DEFAULT NULL ); This schema contains all elements needed for master, slave and superslave operation. Depending on which features will be used, 'GRANT' statements can be trimmed to make sure PDNS cannot subvert the contents of your database. Zone2sql with the --gpgsql flag also assumes this layout is in place. Inserting records is a bit different compared to MySQL and PostgreSQL, you should use: insert into domains (id,name,type) values (domains_id_sequence.nextval,' netherlabs.nl','NATIVE'); Furthermore, use the goracle-tnsname setting to specify which TNSNAME the Generic Oracle Backend should be connectiong to. There are no goracle-dbname, goracle-host or goracle-port settings, their equivalent is in /etc/tnsnames.ora. __________________________________________________________ A.5.4. Basic functionality 4 queries are needed for regular lookups, 4 for 'fancy records' which are disabled by default and 1 is needed for zone transfers. The 4+4 regular queries must return the following 6 fields, in this exact order: content This is the 'right hand side' of a DNS record. For an A record, this is the IP address for example. ttl TTL of this record, in seconds. Must be a real value, no checking is performed. prio For MX records, this should be the priority of the mail exchanger specified. qtype The ASCII representation of the qtype of this record. Examples are 'A', 'MX', 'SOA', 'AAAA'. Make sure that this field returns an exact answer - PDNS won't recognise 'A ' as 'A'. This can be achieved by using a VARCHAR instead of a CHAR. domain_id Each domain must have a unique domain_id. No two domains may share a domain_id, all records in a domain should have the same. A number. name Actual name of a record. Must not end in a '.' and be fully qualified - it is not relative to the name of the domain! Please note that the names of the fields are not relevant, but the order is! As said earlier, there are 8 SQL queries for regular lookups. To configure them, set 'gmysql-basic-query' or 'gpgsql-basic-query', depending on your choice of backend. If so called 'MBOXFW' fancy records are not used, four queries remain: basic-query Default: select content,ttl,prio,type,domain_id,name from records where type='%s' and name='%s' This is the most used query, needed for doing 1:1 lookups of qtype/name values. First %s is replaced by the ASCII representation of the qtype of the question, the second by the name. id-query Default: select content,ttl,prio,type,domain_id,name from records where type='%s' and name='%s' and domain_id=%d Used for doing lookups within a domain. First %s is replaced by the qtype, the %d which should appear after the %s by the numeric domain_id. any-query For doing ANY queries. Also used internally. Default: select content,ttl,prio,type,domain_id,name from records where name='%s' The %s is replaced by the qname of the question. any-id-query For doing ANY queries within a domain. Also used internally. Default: select content,ttl,prio,type,domain_id,name from records where name='%s' and domain_id=%d The %s is replaced by the name of the domain, the %d by the numerical domain id. The last query is for listing the entire contents of a zone. This is needed when performing a zone transfer, but sometimes also internally: list-query To list an entire zone. Default: select content,ttl,prio,type,domain_id,name from records where domain_id=%d __________________________________________________________ A.5.5. Master/slave queries Most installations will have zero need to change the following settings, but should the need arise, here they are: master-zone-query Called to determine the master of a zone. Default: select master from domains where name='%s' and type='SLAVE' info-zone-query Called to retrieve (nearly) all information for a domain: Default: select id,name,master,last_check,notified_serial,type from domains where name='%s' info-all-slaves-query Called to retrieve all slave domains Default: select id,name,master,last_check,type from domains where type='SLAVE' supermaster-query Called to determine if a certain host is a supermaster for a certain domain name. Default: select account from supermasters where ip='%s' and nameserver='%s'); insert-slave-query Called to add a domain as slave after a supermaster notification. Default: insert into domains (type,name,master,account) values('SLAVE','%s','%s','%s') insert-record-query Called during incoming AXFR. Default: insert into records (content,ttl,prio,type,domain_id,name) values ('%s',%d,%d,'%s',%d,'%s') update-serial-query Called to update the last notified serial of a master domain. Default: update domains set notified_serial=%d where id=%d update-lastcheck-query Called to update the last time a slave domain was checked for freshness. Default: update domains set notified_serial=%d where id=%d info-all-master-query Called to get data on all domains for which the server is master. Default: select id,name,master,last_check,notified_serial,type from domains where type='MASTER' delete-zone-query Called to delete all records of a zone. Used before an incoming AXFR. Default: delete from records where domain_id=%d __________________________________________________________ A.5.6. Fancy records If PDNS is used with so called 'Fancy Records', the 'MBOXFW' record exists which specifies an email address forwarding instruction, wildcard queries are sometimes needed. This is not enabled by default. A wildcard query is an internal concept - it has no relation to *.domain-type lookups. You can safely leave these queries blank. wildcard-query Can be left blank. See above for an explanation. Default: select content,ttl,prio,type,domain_id,name from records where type='%s' and name like '%s' wildcard-id-query Can be left blank. See above for an explanation. Default: select content,ttl,prio,type,domain_id,name from records where type='%s' and name like '%s' and domain_id=%d Used for doing lookups within a domain. wildcard-any-query For doing wildcard ANY queries. Default: select content,ttl,prio,type,domain_id,name from records where name like '%s' wildcard-any-id-query For doing wildcard ANY queries within a domain. Default: select content,ttl,prio,type,domain_id,name from records where name like '%s' and domain_id=%d __________________________________________________________ A.5.7. Settings and specifying queries The queries above are specified in pdns.conf. For example, the basic-query would appear as: gpgsql-basic-query=select content,ttl,prio,type,domain_id,na me from records where type='%s' and name='%s' When using the Generic PostgreSQL backend, they appear as above. When using the generic MySQL backend, change the "gpgsql-" prefix to "gmysql-". Queries can span multiple lines, like this: gpgsql-basic-query=select content,ttl,prio,type,domain_id,na me from records \ where type='%s' and name='%s' Do not wrap statements in quotes as this will not work. Besides the query related settings, the following configuration options are available, where one should substitute 'gmysql', 'gpgsql', 'godbc' or 'goracle' for the prefix 'backend'. So 'backend-dbname' can stand for 'gpgsql-dbname' or 'gmysql-dbname' etc. backend-dbname Database name to connect to backend-host Database host to connect to. WARNING: When specified as a hostname a chicken/egg situation might arise where the database is needed to resolve the IP address of the database. It is best to supply an IP address of the database here. backend-port Database port to connect to. gmysql-socket (only for MySQL!) Filename where the MySQL connection socket resides. Often /tmp/mysql.sock or /var/run/mysqld/mysqld.sock. backend-password Password to connect with backend-user PgSQL user to connect as __________________________________________________________ A.5.8. Native operation For native operation, either drop the FOREIGN KEY on the domain_id field, or (recommended), make sure the domains table is filled properly. To add a domain, issue the following: insert into domains (name,type) values ('powerdns.com','NATI VE'); The records table can now be filled by with the domain_id set to the id of the domains table row just inserted. __________________________________________________________ A.5.9. Slave operation The PostgreSQL backend is fully slave capable. To become a slave of the 'powerdns.com' domain, execute this: insert into domains (name,master,type) values ('powerdns.com ','213.244.168.217','SLAVE'); And wait a while for PDNS to pick up the addition - which happens within one minute. There is no need to inform PDNS that a new domain was added. Typical output is: Apr 09 13:34:29 All slave domains are fresh Apr 09 13:35:29 1 slave domain needs checking Apr 09 13:35:29 Domain powerdns.com is stale, master serial 1, our serial 0 Apr 09 13:35:30 [gPgSQLBackend] Connected to database Apr 09 13:35:30 AXFR started for 'powerdns.com' Apr 09 13:35:30 AXFR done for 'powerdns.com' Apr 09 13:35:30 [gPgSQLBackend] Closing connection From now on, PDNS is authoritative for the 'powerdns.com' zone and will respond accordingly for queries within that zone. Periodically, PDNS schedules checks to see if domains are still fresh. The default slave-cycle-interval is 60 seconds, large installations may need to raise this value. Once a domain has been checked, it will not be checked before its SOA refresh timer has expired. Domains whose status is unknown get checked every 60 seconds by default. __________________________________________________________ A.5.10. Superslave operation To configure a supermaster with IP address 10.0.0.11 which lists this installation as 'autoslave.powerdns.com', issue the following: insert into supermasters ('10.0.0.11','autoslave.powerdns.co m','internal'); From now on, valid notifies from 10.0.0.11 that list a NS record containing 'autoslave.powerdns.com' will lead to the provisioning of a slave domain under the account 'internal'. See Section 13.2.1 for details. __________________________________________________________ A.5.11. Master operation The PostgreSQL backend is fully master capable with automatic discovery of serial changes. Raising the serial number of a domain suffices to trigger PDNS to send out notifications. To configure a domain for master operation instead of the default native replication, issue: insert into domains (name,type) values ('powerdns.com','MAST ER'); Make sure that the assigned id in the domains table matches the domain_id field in the records table! __________________________________________________________ A.6. Oracle backend Table A-6. Oracle backend capabilities Native Yes Master No Slave No Superslave No Autoserial Yes Module name oracle Launch name oracle Oracle backend with easily configurable SQL statements, allowing you to graft PDNS on any Oracle database of your choosing. PowerDNS is currently ascertaining if this backend can be distributed in binary form without violating Oracle licensing. In the meantime, the source code to the Oracle backend is available in the pdns distribution. The following configuration settings are available: oracle-debug-queries Output all queries to disk for debugging purposes. oracle-time-queries Output all queries to disk for timing purposes. oracle-uppercase-database Change all domain names to uppercase before querying database. oracle-database Oracle database name to connect to. oracle-home PDNS can set the ORACLE_HOME environment variable from within the executable, allowing execution of the daemon from init.d scripts where ORACLE_HOME may not yet be set. oracle-sid PDNS can set the ORACLE_SID environment variable from within the executable, allowing execution of the daemon from init.d scripts where ORACLE_SID may not yet be set. oracle-username Oracle username to connect as. oracle-password Oracle password to connect with. The generic Oracle backend can be configured to use user-specified queries. The following are the default queries and their names: oracle-forward-query select content, TimeToLive, Priority, type, ZoneId, nvl(ChangeDate,0) from Records where name = :name and type = :type oracle-forward-query-by-zone select content, TimeToLive, Priority, type, ZoneId, nvl(ChangeDate,0) from records where name = :name and type = :type and ZoneId = :id oracle-forward-any-query select content, TimeToLive, Priority, type, ZoneId, nvl(ChangeDate,0) from records where name = :name oracle-list-query select content, TimeToLive, Priority, type, ZoneId, nvl(ChangeDate, 0), name from records where ZoneId = :id __________________________________________________________ A.6.1. Setting up Oracle for use with PowerDNS To setup a database that corresponds to these default queries, issue the following as Oracle user sys: create user powerdns identified by YOURPASSWORD; grant connect to powerdns; create tablespace powerdns datafile '/opt/oracle/oradata/ora cle/powerdns.dbf' size 256M extent management local autoallocate; alter user powerdns quota unlimited on powerdns; As user 'powerdns' continue with: create table Domains ( ID number(11) NOT NULL, NAME VARCHAR(255) NOT NULL, MASTER VARCHAR(128) DEFAULT NULL, LAST_CHECK INT DEFAULT NULL, TYPE VARCHAR(6) NOT NULL, NOTIFIED_SERIAL INT DEFAULT NULL, ACCOUNT VARCHAR(40) DEFAULT NULL, primary key (ID) )tablespace POWERDNS; create index DOMAINS$NAME on Domains (NAME) tablespace POWERDNS; create sequence DOMAINS_ID_SEQUENCE; create table Records ( ID number(11) NOT NULL, ZoneID number(11) default NULL REFERENCES Domains(ID) ON DE LETE CASCADE, NAME varchar2(255) default NULL, TYPE varchar2(6) default NULL, CONTENT varchar2(255) default NULL, TimeToLive number(11) default NULL, Priority number(11) default NULL, CreateDate number(11) default NULL, ChangeDate number(11) default NULL, primary key (ID) )tablespace POWERDNS; create index RECORDS$NAME on RECORDS (NAME) tablespace POWERDNS; create sequence RECORDS_ID_SEQUENCE; To insert records, either use zone2sql with the --oracle setting, or execute sql along the lines of: insert into domains (id,name,type) values (domains_id_sequence.nextval,' netherlabs.nl','NATIVE'); insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'SOA', 'ahu.casem a.net. hostmaster.ds9a.nl. 2000081401 28800 7200 604800 86400', 3600, 0 from Domains where name='netherlabs.nl'; insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'NS', 'ahu.casema .net', 3600, 0 from Domains where name='netherlabs.nl'; insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'NS', 'ns1.pine.n l', 3600, 0 from Domains where name='netherlabs.nl'; insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'NS', 'ns2.pine.n l', 3600, 0 from Domains where name='netherlabs.nl'; insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'A', '213.244.168 .210', 3600, 0 from Domains where name='netherlabs.nl'; insert into Records (id,ZoneId, name,type,content,TimeToLive,Priority) s elect RECORDS_ID_SEQUENCE.nextval,id ,'netherlabs.nl', 'MX', 'outpost.ds 9a.nl', 3600, 10 from Domains where name='netherlabs.nl'; For performance reasons it is best to specify --transactions too! __________________________________________________________ A.7. Generic SQLite backend (2 and 3) Table A-7. Generic SQLite backend capabilities Native Yes Master Yes Slave Yes Superslave Yes Module name gsqlite and gsqlite3 Launch name gsqlite and gsqlite3 This backend retrieves all data from a SQLite database, which is a RDBMS that's embedded into the application itself, so you won't need to be running a seperate server process. It also reduces overhead, and simplifies installation. At http://www.sqlite.org you can find more information about SQLite. As this is a generic backend, built on top of the gSql framework, you can specify all queries as documented in Generic MySQL and PgSQL backends. SQLite exists in two incompatible versions, numbered 2 and 3, and from 2.9.21 onwards, PowerDNS supports both. It is recommended to go with version 3 as it is newer, has better performance and is actively maintained. To use version 3, choose 'launch=gsqlite3'. __________________________________________________________ A.7.1. Compiling the SQLite backend Before you can begin compiling PowerDNS with the SQLite backend you need to have the SQLite utility and library installed on your system. You can download these from http://www.sqlite.org/download.html, or you can use packages (if your distribution provides those). When you've installed the library you can use: ./configure --with-modules="gsqlite" or ./configure --with-modules="gsqlite3" to configure PowerDNS to use the SQLite backend. Compilation can then proceed as usual. SQLite is included in most PowerDNS binary releases. __________________________________________________________ A.7.2. Setting up the database Before you can use this backend you first have to set it up and fill it with data. The default setup conforms to the following schema: create table domains ( id INTEGER PRIMARY KEY, name VARCHAR(255) NOT NULL, master VARCHAR(128) DEFAULT NULL, last_check INTEGER DEFAULT NULL, type VARCHAR(6) NOT NULL, notified_serial INTEGER DEFAULT NULL, account VARCHAR(40) DEFAULT NULL ); CREATE UNIQUE INDEX name_index ON domains(name); CREATE TABLE records ( id INTEGER PRIMARY KEY, domain_id INTEGER DEFAULT NULL, name VARCHAR(255) DEFAULT NULL, type VARCHAR(6) DEFAULT NULL, content VARCHAR(255) DEFAULT NULL, ttl INTEGER DEFAULT NULL, prio INTEGER DEFAULT NULL, change_date INTEGER DEFAULT NULL ); CREATE INDEX rec_name_index ON records(name); CREATE INDEX nametype_index ON records(name,type); CREATE INDEX domain_id ON records(domain_id); create table supermasters ( ip VARCHAR(25) NOT NULL, nameserver VARCHAR(255) NOT NULL, account VARCHAR(40) DEFAULT NULL ); This schema contains all elements needed for master, slave and superslave operation. After you have created the database you probably want to fill it with data. If you have a BIND zonefile it's as easy as: zone2sql --zone=myzonefile --gmysql | sqlite powerdns.sqlite, but you can also use AXFR (or insert data manually). To communicate with a SQLite database, use either the 'sqlite' or 'sqlite3' program, and feed it SQL. __________________________________________________________ A.7.3. Using the SQLite backend The last thing you need to do is telling PowerDNS to use the SQLite backend. # in pdns.conf launch=gsqlite # or gsqlite3 gsqlite-database= # or gsqli te3-database Then you can start PowerDNS and it should notify you that a connection to the database was made. __________________________________________________________ A.8. DB2 backend Table A-8. DB2 backend capabilities Native Yes Master No Slave No Superslave No Autoserial Yes Module name db2 Launch name db2 PowerDNS is currently ascertaining if this backend can be distributed in binary form without violating IBM DB2 licensing. The DB2 backend executes the following queries: Forward Query select Content, TimeToLive, Priority, Type, ZoneId, 0 as ChangeDate, Name from Records where Name = ? and type = ? Forward By Zone Query select Content, TimeToLive, Priority, Type, ZoneId, 0 as ChangeDate, Name from Records where Name = ? and Type = ? and ZoneId = ? Forward Any Query select Content, TimeToLive, Priority, Type, ZoneId, 0 as ChangeDate, Name from Records where Name = ? List Query select Content, TimeToLive, Priority, Type, ZoneId, 0 as ChangeDate, Name from Records where ZoneId = ? Configuration settings: db2-server Server name to connect to. Defaults to 'powerdns'. Make sure that your nameserver is not needed to resolve an IP address needed to connect as this might lead to a chicken/egg situation. db2-user Username to connect as. Defaults to 'powerdns'. db2-password Password to connect with. Defaults to 'powerdns'. __________________________________________________________ A.9. Bind zone file backend Table A-9. Bind zone file backend capabilities Native Yes Master Yes Slave Yes Superslave No Autoserial No Module name none (built in) Launch bind The BindBackend started life as a demonstration of the versatility of PDNS but quickly gained in importance when there appeared to be demand for a Bind 'workalike'. The BindBackend parses a Bind-style named.conf and extracts information about zones from it. It makes no attempt to honour other configuration flags, which you should configure (when available) using the PDNS native configuration. --help=bind Outputs all known parameters related to the bindbackend bind-example-zones Loads the 'example.com' zone which can be queried to determine if PowerDNS is functioning without configuring database backends. This feature is no longer supported from 2.9.21 onwards. bind-config= Location of the Bind configuration file to parse. bind-check-interval= How often to check for zone changes. See 'Operation' section. bind-enable-huffman Enable Huffman compression on zone data. Currently saves around 20% of memory actually used, but slows down operation somewhat. __________________________________________________________ A.9.1. Operation On launch, the BindBackend first parses the named.conf to determine which zones need to be loaded. These will then be parsed and made available for serving, as they are parsed. So a named.conf with 100.000 zones may take 20 seconds to load, but after 10 seconds, 50.000 zones will already be available. While a domain is being loaded, it is not yet available, to prevent incomplete answers. Reloading is currently done only when a request for a zone comes in, and then only after bind-check-interval seconds have passed after the last check. If a change occurred, access to the zone is disabled, the file is reloaded, access is restored, and the question is answered. For regular zones, reloading is fast enough to answer the question which lead to the reload within the DNS timeout. If bind-check-interval is specified as zero, no checks will be performed until the pdns_control reload is given. __________________________________________________________ A.9.2. Pdns_control commands bind-domain-status domain [domain] Output status of domain or domains. Can be one of 'seen in named.conf, not parsed', 'parsed successfully at ' or 'error parsing at line ... at