| 1 | PowerDNS geobackend setup notes |
|---|
| 2 | =============================== |
|---|
| 3 | |
|---|
| 4 | These are the steps I went through to set up geobackend for PowerDNS on FreeBSD |
|---|
| 5 | -STABLE. By the time you read this maybe geobackend is part of the PowerDNS |
|---|
| 6 | main CVS so perhaps not much of this will apply. In that case you should skip |
|---|
| 7 | further down to the configuration part. |
|---|
| 8 | |
|---|
| 9 | Before I carry on I should probably point out that if you don't know how DNS |
|---|
| 10 | works much, or if you have never installed PowerDNS ever before, then you |
|---|
| 11 | probably won't understand any of this. In that case you should probably go and |
|---|
| 12 | do some reading/practicing before trying to set this up for yourself. As a |
|---|
| 13 | minimum of DNS knowledge I would say you need to understand: |
|---|
| 14 | |
|---|
| 15 | - Basically how DNS servers answer queries |
|---|
| 16 | |
|---|
| 17 | - What common DNS terms like "CNAME" "RR" and "SOA" mean |
|---|
| 18 | |
|---|
| 19 | - How to use diagnostic utilities such as "dig" to test your setup. |
|---|
| 20 | |
|---|
| 21 | So, how this should work on Linux or any general Unix system. This didn't work |
|---|
| 22 | for me so I ended up having to do it another way, but anyhow.. |
|---|
| 23 | |
|---|
| 24 | 1. Download PowerDNS source from http://www.powerdns.com/downloads/index.php - |
|---|
| 25 | you want the GPL sources. |
|---|
| 26 | |
|---|
| 27 | 2. Edit configure.in in the main directory so that where it has the list of |
|---|
| 28 | backends at the bottom, you add: |
|---|
| 29 | |
|---|
| 30 | modules/geobackend/Makefile |
|---|
| 31 | |
|---|
| 32 | 3. $ cd modules |
|---|
| 33 | |
|---|
| 34 | and get the geobackend: |
|---|
| 35 | |
|---|
| 36 | $ cvs -d :pserver:anon@cvs.blitzed.org:/data/cvs login |
|---|
| 37 | (press <return> at password prompt) |
|---|
| 38 | $ cvs -d :pserver:anon@cvs.blitzed.org:/data/cvs co -d geobackend geo-dns |
|---|
| 39 | |
|---|
| 40 | $ cd .. |
|---|
| 41 | |
|---|
| 42 | to return to top of build directory. |
|---|
| 43 | |
|---|
| 44 | 4. Regenerate the autotools with: |
|---|
| 45 | |
|---|
| 46 | $ aclocal |
|---|
| 47 | $ autoheader |
|---|
| 48 | $ automake --add-missing --copy --foreign |
|---|
| 49 | $ autoconf |
|---|
| 50 | |
|---|
| 51 | 5. Do a ./configure with the flags you normally would use, but also |
|---|
| 52 | --with-dynmodules="geo" |
|---|
| 53 | |
|---|
| 54 | 6. Install PowerDNS as normal for how you would normally use it. |
|---|
| 55 | |
|---|
| 56 | This did not work for me on FreeBSD: no matter what combination of autoconf, |
|---|
| 57 | automake, libtool was installed I would always get one error or another at the |
|---|
| 58 | stage where I was running those commands. So, after a few hours of messing |
|---|
| 59 | around I decided to just compile the geobackend outside of the PowerDNS source |
|---|
| 60 | tree. Probably if/when geobackend is made part of PowerDNS, this will "just |
|---|
| 61 | work", but in the meantime here's what I did: |
|---|
| 62 | |
|---|
| 63 | 1. Install PowerDNS from the port /usr/ports/dns/powerdns as normal. Do not do |
|---|
| 64 | a "make clean" yet though! Make sure your PowerDNS works as you'd expect |
|---|
| 65 | without any of this geobackend stuff before going further. |
|---|
| 66 | |
|---|
| 67 | 2. Somewhere else, get the source of our geobackend as above in step 3. |
|---|
| 68 | |
|---|
| 69 | 3. Compile geobackend: |
|---|
| 70 | |
|---|
| 71 | $ c++ -I/usr/ports/dns/powerdns/work/pdns-2.9.15 -O2 -Wall -c geobackend.cc |
|---|
| 72 | $ c++ -I/usr/ports/dns/powerdns/work/pdns-2.9.15 -O2 -Wall -c ippreftree.cc |
|---|
| 73 | $ c++ ippreftree.o geobackend.o -Wl,-soname -Wl,libgeobackend.so.0 -shared -o libgeobackend.so |
|---|
| 74 | |
|---|
| 75 | All of those should compile and link without error. |
|---|
| 76 | |
|---|
| 77 | 4. Now you need to copy that shared library to the system library directory, as |
|---|
| 78 | root: |
|---|
| 79 | |
|---|
| 80 | # cp libgeobackend.so /usr/local/lib/ |
|---|
| 81 | |
|---|
| 82 | Now whichever way you managed to get libgeobackend.so compiled, you are now |
|---|
| 83 | ready to configure PowerDNS. This is what Blitzed's configuration looks like |
|---|
| 84 | right now (but this is very much an experiment so things are bound to get out |
|---|
| 85 | of date quickly). |
|---|
| 86 | |
|---|
| 87 | By the way, I could not find a SysV-style startup script installed by the |
|---|
| 88 | FreeBSD port so I had to copy one from debian and put it in |
|---|
| 89 | /usr/local/etc/rc.d/pdns.sh. You can get that file here: |
|---|
| 90 | http://nubian.blitzed.org/pdns.sh |
|---|
| 91 | |
|---|
| 92 | And another thing is that the FreeBSD port doesn't add any new users for |
|---|
| 93 | PowerDNS's use. You probably don't want to run it as root even without our |
|---|
| 94 | code in it! So be sure to add some sensible user and group like "pdns". |
|---|
| 95 | |
|---|
| 96 | PowerDNS Configuration |
|---|
| 97 | ====================== |
|---|
| 98 | |
|---|
| 99 | Here is the relevant parts of my /usr/local/etc/pdns.conf file as running on |
|---|
| 100 | FreeBSD -STABLE: |
|---|
| 101 | |
|---|
| 102 | # ------------------------------------------------------------------------- |
|---|
| 103 | |
|---|
| 104 | # To make it run as user@group pdns:pdns instead of root:root |
|---|
| 105 | setgid=pdns |
|---|
| 106 | setuid=pdns |
|---|
| 107 | |
|---|
| 108 | # These totally disable query+packet caching for all zones. This is necessary |
|---|
| 109 | # because otherwise when the exact same question is asked twice in a short |
|---|
| 110 | # period of time (by default, 10 seconds), the same response will be given |
|---|
| 111 | # without any backends getting involved. |
|---|
| 112 | # |
|---|
| 113 | # This is bad for geobackend because obviously every question can potentially |
|---|
| 114 | # require a new answer based only on the IP of the user's nameserver. Now, it |
|---|
| 115 | # should be noted that if you have other zones in PowerDNS then they will have |
|---|
| 116 | # their query cache disabled as well. That's not ideal, so you probably want |
|---|
| 117 | # to run a separate instance of PowerDNS just for geobackend. Maybe one day |
|---|
| 118 | # there will be config options to set per-zone query caching time or something. |
|---|
| 119 | query-cache-ttl=0 |
|---|
| 120 | cache-ttl=0 |
|---|
| 121 | |
|---|
| 122 | # Log a lot of stuff. Logging is slow. We will disable this when we are happy |
|---|
| 123 | # things are working. :) |
|---|
| 124 | loglevel=7 |
|---|
| 125 | |
|---|
| 126 | # But these logs are not interesting at the moment |
|---|
| 127 | log-dns-details=no |
|---|
| 128 | |
|---|
| 129 | # This disables wildcards which is more efficient. geobackend doesn't use |
|---|
| 130 | # them, so if none of your backends need them, set this, otherwise comment it |
|---|
| 131 | # out. |
|---|
| 132 | wildcards=no |
|---|
| 133 | |
|---|
| 134 | # The geobackend |
|---|
| 135 | launch=geo |
|---|
| 136 | |
|---|
| 137 | # The zone that your geo-balanced RR is inside of. The whole zone has to be |
|---|
| 138 | # delegated to the PowerDNS backend, so you will generally want to make up some |
|---|
| 139 | # subzone of your main zone. We chose "geo.blitzed.org". |
|---|
| 140 | # |
|---|
| 141 | geo-zone=geo.blitzed.org |
|---|
| 142 | |
|---|
| 143 | # The only parts of the SOA for "geo.blitzed.org" that apply here are the |
|---|
| 144 | # master server name and the contact address. |
|---|
| 145 | geo-soa-values=ns0.blitzed.org,hostmaster@blitzed.org |
|---|
| 146 | |
|---|
| 147 | # List of NS records of the PowerDNS servers that are authoritative for your |
|---|
| 148 | # GLB zone. |
|---|
| 149 | geo-ns-records=ns0.blitzed.org,ns1.blitzed.org |
|---|
| 150 | |
|---|
| 151 | # The TTL of the CNAME records that geobackend will return. Since the same |
|---|
| 152 | # resolver will always get the same CNAME (apart from if the director-map |
|---|
| 153 | # changes) it is safe to return a reasonable TTL, so if you leave this |
|---|
| 154 | # commented then a sane default will be chosen. |
|---|
| 155 | #geo-ttl=3600 |
|---|
| 156 | |
|---|
| 157 | # The TTL of the NS records that will be returned. Leave this commented if you |
|---|
| 158 | # don't understand. |
|---|
| 159 | #geo-ns-ttl=86400 |
|---|
| 160 | |
|---|
| 161 | # This is the real guts of the data that drives this backend. This is a DNS |
|---|
| 162 | # zone file for RBLDNSD, a nameserver specialised for running large DNS zones |
|---|
| 163 | # typical of DNSBLs and such. We choose it for our data because it is easier |
|---|
| 164 | # to parse than the BIND-format one. |
|---|
| 165 | # |
|---|
| 166 | # Anyway, it comes from http://countries.nerd.dk/more.html - there are details |
|---|
| 167 | # there for how to rsync your own copy. You'll want to do that regularly, |
|---|
| 168 | # every couple of days maybe. We believe the nerd.dk guys take the netblock |
|---|
| 169 | # info from Regional Internet Registries (RIRs) like RIPE, ARIN, APNIC. From |
|---|
| 170 | # that they build a big zonefile of IP/prefixlen -> ISO-country-code mappings. |
|---|
| 171 | geo-ip-map-zonefile=/usr/local/etc/zz.countries.nerd.dk.rbldnsd |
|---|
| 172 | |
|---|
| 173 | # And finally this last directive tells the geobackend where to find the map |
|---|
| 174 | # files that say a) which RR to answer for, and b) what actual resource record |
|---|
| 175 | # to return for each ISO country code. The setting here is a comma-separated |
|---|
| 176 | # list of paths, each of which may either be a single map file or a directory |
|---|
| 177 | # that will contain map files. If you are only ever going to serve one RR then |
|---|
| 178 | # a single file is probably better, but if you're going to serve many then a |
|---|
| 179 | # directory would probably be better. The rest of this documentation will |
|---|
| 180 | # assume you chose a directory. |
|---|
| 181 | geo-maps=/usr/local/etc/geo-maps |
|---|
| 182 | |
|---|
| 183 | # ------------------------------------------------------------------------- |
|---|
| 184 | |
|---|
| 185 | Map configuration |
|---|
| 186 | ================= |
|---|
| 187 | |
|---|
| 188 | Above you defined a directory which should contain one file for each RR you are |
|---|
| 189 | going to serve. This section describes the format for those files. |
|---|
| 190 | |
|---|
| 191 | There is a perl script in the geo-dns module |
|---|
| 192 | (http://cvs.blitzed.org/geo-dns/iso2region.pl) which will print out a useful |
|---|
| 193 | template for starting with. There are just two lines you MUST add for your own |
|---|
| 194 | setup. The one that Blitzed is using is here: |
|---|
| 195 | http://nubian.blitzed.org/irc.geo.blitzed.org |
|---|
| 196 | |
|---|
| 197 | The first line you must add is the $RECORD line. This tells the geobackend |
|---|
| 198 | which RR within the geo-zone the file is for, so for example the file above |
|---|
| 199 | gives: |
|---|
| 200 | |
|---|
| 201 | $RECORD irc |
|---|
| 202 | |
|---|
| 203 | meaning it is for irc.geo.blitzed.org. |
|---|
| 204 | |
|---|
| 205 | The second line that must be present is the $ORIGIN line: |
|---|
| 206 | |
|---|
| 207 | $ORIGIN iso.blitzed.org. |
|---|
| 208 | |
|---|
| 209 | The rest of this file is a list of mappings of ISO country to RR, and the |
|---|
| 210 | $ORIGIN line tells the geobackend how to qualify the RRs. Any relative RR with |
|---|
| 211 | be qualified by adding a dot and then this $ORIGIN string onto it. So all the |
|---|
| 212 | relative RRs that follow are actually in the "iso.blitzed.org." zone. If you want to refer to an RR outside the $ORIGIN, put a trailing dot. |
|---|
| 213 | |
|---|
| 214 | The final mandatory line is the 0 mapping: |
|---|
| 215 | |
|---|
| 216 | 0 pool.blitzed.org. |
|---|
| 217 | |
|---|
| 218 | This is the "default" mapping. It's possible that you will get a query from an |
|---|
| 219 | IP that is not represented in the nerd.dk zone. Maybe it is a new allocation |
|---|
| 220 | by a RIR, or maybe something unexpected happened like you got a query from IPv6 |
|---|
| 221 | or from an RFC1918 address. Or, there could be some error elsewhere in the |
|---|
| 222 | geobackend that makes it want to give up. In any of these cases it needs to |
|---|
| 223 | return a CNAME to a useful default. |
|---|
| 224 | |
|---|
| 225 | The default chosen for blitzed is "pool.blitzed.org." (note the trailing dot |
|---|
| 226 | puts it outside the $ORIGIN!). At the moment, pool.blitzed.org is a |
|---|
| 227 | round-robin of A records of all our connected servers, but the best way to |
|---|
| 228 | handle it is under debate. Since it is in one of our regular zone files we |
|---|
| 229 | can change it later how we want. |
|---|
| 230 | |
|---|
| 231 | The entire rest of the file is optional, and takes the format of an ISO country |
|---|
| 232 | code (number) and an RR to map it to. iso2region.pl will have helpfully added |
|---|
| 233 | comments with the country name so you can see what is what: |
|---|
| 234 | |
|---|
| 235 | # Belgium |
|---|
| 236 | 56 eu |
|---|
| 237 | |
|---|
| 238 | Lines starting with # are comments. That's mainly so you can tell what the ISO |
|---|
| 239 | code corresponds to, and maybe later when you are tweaking where all these |
|---|
| 240 | countries will go to you can add some documentation for why you did it. |
|---|
| 241 | |
|---|
| 242 | The "56" is the ISO country code (see |
|---|
| 243 | http://www.iso.ch/iso/en/prods-services/iso3166ma/02iso-3166-code-lists/index.html) |
|---|
| 244 | for Belgium. The "eu" tells the geobackend that if a query for |
|---|
| 245 | irc.geo.blitzed.org should come in from someone in Belgium, then it should |
|---|
| 246 | respond with a CNAME for "eu.iso.blitzed.org". That's it. That is the RR that |
|---|
| 247 | gets sent back. Every other line in this file is the same, just code->RR maps. |
|---|
| 248 | |
|---|
| 249 | At this point you might be wondering when the user gets the actual IP address |
|---|
| 250 | sent to them. The answer is that we have chosen to make our geobackend only |
|---|
| 251 | respond with CNAMEs to other RRs that are assumed to be hosted elsewhere in |
|---|
| 252 | DNS. Our main blitzed.org zone is hosted in bind servers like it has been for |
|---|
| 253 | years. In that zone we have entries for every one of the RRs that appears on |
|---|
| 254 | the right hand side in the director-map. A list of those RRs is as follows: |
|---|
| 255 | |
|---|
| 256 | an.iso.blitzed.org. |
|---|
| 257 | af.iso.blitzed.org. |
|---|
| 258 | as.iso.blitzed.org. |
|---|
| 259 | eu.iso.blitzed.org. |
|---|
| 260 | na.iso.blitzed.org. |
|---|
| 261 | oc.iso.blitzed.org. |
|---|
| 262 | sa.iso.blitzed.org. |
|---|
| 263 | |
|---|
| 264 | These correspond to the ISO names for the regions/continents (Antarctica, |
|---|
| 265 | Africa, Asia, Europe, North America, Oceania, South America) and are |
|---|
| 266 | represented in our DNS at the moment by either a list of A records of servers |
|---|
| 267 | "near" there, or else a CNAME to a "nearby" region. For example we have no |
|---|
| 268 | servers in Antarctica or Asia. We just CNAME those to "na.iso.blitzed.org." to |
|---|
| 269 | send those users to North American servers instead. Doing it this way is just |
|---|
| 270 | our first attempt, we are still experimenting and might decide to do it |
|---|
| 271 | different. All you need to know is that the geobackend gives the CNAMEs you |
|---|
| 272 | tell it to give, it's your business what those CNAMEs are and what they end up |
|---|
| 273 | resolving to. |
|---|
| 274 | |
|---|
| 275 | DNS configuration |
|---|
| 276 | ================= |
|---|
| 277 | |
|---|
| 278 | Time to configure the things that go into your existing DNS setup: |
|---|
| 279 | |
|---|
| 280 | 1. The geographically load-balanced zone needs to be delegated to your PowerDNS |
|---|
| 281 | servers. For Blitzed we chose "geo.blitzed.org.", so: |
|---|
| 282 | |
|---|
| 283 | geo NS ns0.blitzed.org. |
|---|
| 284 | geo NS ns1.blitzed.org. |
|---|
| 285 | geo NS ns2.blitzed.org. |
|---|
| 286 | |
|---|
| 287 | 2. The lists of servers that correspond to each CNAME that your director-map |
|---|
| 288 | can possibly come up with. The above configuration can only answer one of: |
|---|
| 289 | |
|---|
| 290 | pool.blitzed.org. |
|---|
| 291 | an.iso.blitzed.org. |
|---|
| 292 | af.iso.blitzed.org. |
|---|
| 293 | as.iso.blitzed.org. |
|---|
| 294 | eu.iso.blitzed.org. |
|---|
| 295 | na.iso.blitzed.org. |
|---|
| 296 | oc.iso.blitzed.org. |
|---|
| 297 | sa.iso.blitzed.org. |
|---|
| 298 | |
|---|
| 299 | We have no servers in Antarctica so we just send that to North America: |
|---|
| 300 | |
|---|
| 301 | an.iso CNAME na.iso.blitzed.org. |
|---|
| 302 | |
|---|
| 303 | Other regions that actually have servers light look like: |
|---|
| 304 | |
|---|
| 305 | na.iso A 1.2.3.4 ; Some server in North America |
|---|
| 306 | na.iso A 2.3.4.1 ; Some other server in North America |
|---|
| 307 | |
|---|
| 308 | 3. Eventually you probably will want to use a more friendly name than something |
|---|
| 309 | like "irc.geo.blitzed.org.". At that point you could just do the equivalent |
|---|
| 310 | of: |
|---|
| 311 | |
|---|
| 312 | irc CNAME irc.geo.blitzed.org. |
|---|
| 313 | |
|---|
| 314 | BEAR IN MIND THAT THERE MIGHT BE BUGS IN THIS BACKEND AND IF YOU DO THIS TO |
|---|
| 315 | YOUR MAIN POOL AND IT STOPS RESPONDING, SENDS YOUR USERS TO THE WRONG SERVERS, |
|---|
| 316 | OR EVEN TO THE WRONG NETWORKS, OR ANYTHING ELSE UNFORTUNATE HAPPENS AT ALL, |
|---|
| 317 | THEN THAT'S JUST TOUGH LUCK AS THIS CODE COMES WITH NO WARRANTY, GUARANTEE |
|---|
| 318 | OR ASSURANCE OF ANY KIND! |
|---|
| 319 | |
|---|
| 320 | Testing |
|---|
| 321 | ======= |
|---|
| 322 | |
|---|
| 323 | OK! If you're still awake after all that, it is ready to test. |
|---|
| 324 | |
|---|
| 325 | By the way, at the moment some of the logging from the geobackend is a severity |
|---|
| 326 | "debug" (facility "daemon" if using FreeBSD port). The default FreeBSD -STABLE |
|---|
| 327 | install does not log daemon.debug to any file. If you don't add daemon.debug |
|---|
| 328 | to your /etc/syslog.conf then you might not see some of the logs I shall talk |
|---|
| 329 | about later. Most logging will be removed or made optional anyway as it slows |
|---|
| 330 | things down. |
|---|
| 331 | |
|---|
| 332 | - Start PowerDNS |
|---|
| 333 | |
|---|
| 334 | # /usr/local/etc/rc.d/pdns.sh start |
|---|
| 335 | |
|---|
| 336 | - Check your logs! You should see something like this: |
|---|
| 337 | |
|---|
| 338 | Feb 26 16:07:45 nubian pdns[4661]: PowerDNS 2.9.15 (C) 2001-2004 PowerDNS.COM BV (Feb 9 2004, 23:40:35) starting up |
|---|
| 339 | Feb 26 16:07:45 nubian pdns[4661]: PowerDNS comes with ABSOLUTELY NO WARRANTY. This is free software, and you are welcome to redistribute it according to the terms of the GPL version 2. |
|---|
| 340 | Feb 26 16:07:45 nubian pdns[4661]: Set effective group id to 1023 |
|---|
| 341 | Feb 26 16:07:45 nubian pdns[4661]: Set effective user id to 1023 |
|---|
| 342 | Feb 26 16:07:45 nubian pdns[4661]: Creating backend connection for TCP |
|---|
| 343 | Feb 26 16:07:45 nubian pdns[4661]: [geobackend] Parsing IP map zonefile |
|---|
| 344 | Feb 26 16:07:47 nubian pdns[4661]: [geobackend] Finished parsing IP map zonefile: added 53072 prefixes, stored in 132525 nodes using 1590300 bytes of memory |
|---|
| 345 | Feb 26 16:07:47 nubian pdns[4661]: [geobackend] Parsing director map /usr/local/etc/geo-maps/irc.geo.blitzed.org |
|---|
| 346 | Feb 26 16:07:47 nubian pdns[4661]: [geobackend] Finished parsing 2 director map files, 0 failures |
|---|
| 347 | Feb 26 16:07:47 nubian pdns[4661]: About to create 3 backend threads |
|---|
| 348 | Feb 26 16:07:47 nubian pdns[4661]: Done launching threads, ready to distribute questions |
|---|
| 349 | |
|---|
| 350 | As long as there were no errors, the server is ready, geobackend is probably |
|---|
| 351 | working. You should now test with a suitable diagnostic tool: |
|---|
| 352 | |
|---|
| 353 | $ dig irc.geo.blitzed.org. |
|---|
| 354 | |
|---|
| 355 | ; <<>> DiG 9.2.2 <<>> irc.geo.blitzed.org. |
|---|
| 356 | ;; global options: printcmd |
|---|
| 357 | ;; Got answer: |
|---|
| 358 | ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 59602 |
|---|
| 359 | ;; flags: qr rd ra; QUERY: 1, ANSWER: 6, AUTHORITY: 5, ADDITIONAL: 5 |
|---|
| 360 | |
|---|
| 361 | ;; QUESTION SECTION: |
|---|
| 362 | ;irc.geo.blitzed.org. IN A |
|---|
| 363 | |
|---|
| 364 | ;; ANSWER SECTION: |
|---|
| 365 | irc.geo.blitzed.org. 0 IN CNAME eu.iso.blitzed.org. |
|---|
| 366 | eu.iso.blitzed.org. 1 IN A 195.92.253.3 |
|---|
| 367 | eu.iso.blitzed.org. 1 IN A 213.193.225.252 |
|---|
| 368 | eu.iso.blitzed.org. 1 IN A 62.80.124.155 |
|---|
| 369 | eu.iso.blitzed.org. 1 IN A 80.196.158.72 |
|---|
| 370 | eu.iso.blitzed.org. 1 IN A 195.22.74.199 |
|---|
| 371 | |
|---|
| 372 | ;; AUTHORITY SECTION: |
|---|
| 373 | blitzed.org. 3600 IN NS sou.nameserver.net. |
|---|
| 374 | blitzed.org. 3600 IN NS bos.nameserver.net. |
|---|
| 375 | blitzed.org. 3600 IN NS iad.nameserver.net. |
|---|
| 376 | blitzed.org. 3600 IN NS phl.nameserver.net. |
|---|
| 377 | blitzed.org. 3600 IN NS sjc.nameserver.net. |
|---|
| 378 | |
|---|
| 379 | ;; ADDITIONAL SECTION: |
|---|
| 380 | bos.nameserver.net. 43200 IN A 203.20.52.5 |
|---|
| 381 | iad.nameserver.net. 43200 IN A 192.148.252.171 |
|---|
| 382 | phl.nameserver.net. 7200 IN A 203.56.139.102 |
|---|
| 383 | sjc.nameserver.net. 43200 IN A 205.158.174.201 |
|---|
| 384 | sou.nameserver.net. 34825 IN A 194.196.163.7 |
|---|
| 385 | |
|---|
| 386 | ;; Query time: 389 msec |
|---|
| 387 | ;; SERVER: 192.168.0.5#53(192.168.0.5) |
|---|
| 388 | ;; WHEN: Tue Feb 10 03:10:14 2004 |
|---|
| 389 | ;; MSG SIZE rcvd: 322 |
|---|
| 390 | |
|---|
| 391 | What does this show? Well first of all it tells us that we looked for the A |
|---|
| 392 | record of "irc.geo.blitzed.org." (A records are the default RR for dig). |
|---|
| 393 | What we actually got back was a CNAME to "eu.iso.blitzed.org." At that point |
|---|
| 394 | the work of geobackend and our PowerDNS server as a whole is done. All it is |
|---|
| 395 | designed to do is return a CNAME based on the location of the server doing |
|---|
| 396 | the query. The server I did that from is in UK, so a response of |
|---|
| 397 | "eu.iso.blitzed.org." is correct. |
|---|
| 398 | |
|---|
| 399 | The rest of the data comes from the normal BIND9 nameservers that are |
|---|
| 400 | authoritative for the "blitzed.org." zone, in this case a list of A records |
|---|
| 401 | corresponding to our EU servers. Finally the list of authoritative servers |
|---|
| 402 | for "blitzed.org." is given, those same BIND9 boxes. |
|---|
| 403 | |
|---|
| 404 | Meanwhile in the syslog of nubian, we have: |
|---|
| 405 | |
|---|
| 406 | Feb 10 03:07:02 nubian pdns[32106]: [geobackend] Serving irc.geo.blitzed.org CNAME eu.iso.blitzed.org to 82.195.224.5 (756) |
|---|
| 407 | |
|---|
| 408 | Here you can see that 82.195.224.5 asked for "irc.geo.blitzed.org." and was |
|---|
| 409 | served the mapping for ISO code 756: "eu.iso.blitzed.org.". This log notice |
|---|
| 410 | will be useful for debugging and refining the director-map by hand, but |
|---|
| 411 | later it will probably be removed or made optional. |
|---|
| 412 | |
|---|
| 413 | Ongoing maintenance |
|---|
| 414 | =================== |
|---|
| 415 | |
|---|
| 416 | New IPs are regularly allocated, also there may end up being corrections to the |
|---|
| 417 | nerd.dk zones, so you should arrange to rsync this file every so often. I'm |
|---|
| 418 | guessing once a week would be adequate. You may not be satisfied with your |
|---|
| 419 | first try at the director-map either, so from time to time you may make |
|---|
| 420 | changes. You might also add more map files to your geo-maps directory. |
|---|
| 421 | Anytime those changes happen you will need to tell the geobackend to reread |
|---|
| 422 | them. At the moment the best way to do this is: |
|---|
| 423 | |
|---|
| 424 | # pdns_control rediscover |
|---|
| 425 | |
|---|
| 426 | Feb 26 16:10:57 nubian pdns[4661]: Rediscovery was requested |
|---|
| 427 | Feb 26 16:10:57 nubian pdns[4661]: [geobackend] Parsing IP map zonefile |
|---|
| 428 | Feb 26 16:10:58 nubian pdns[4661]: [geobackend] Finished parsing IP map zonefile: added 53072 prefixes, stored in 132525 nodes using 1590300 bytes of memory |
|---|
| 429 | Feb 26 16:10:58 nubian pdns[4661]: [geobackend] Parsing director map /usr/local/etc/geo-maps/irc.geo.blitzed.org |
|---|
| 430 | Feb 26 16:10:58 nubian pdns[4661]: [geobackend] Parsing director map /usr/local/etc/geo-maps/irc.strugglers.net |
|---|
| 431 | Feb 26 16:10:58 nubian pdns[4661]: [geobackend] Finished parsing 2 director map files, 0 failures |
|---|
| 432 | |
|---|
| 433 | About recursive nameservers |
|---|
| 434 | =========================== |
|---|
| 435 | |
|---|
| 436 | There is a small but potentially confusing gotcha in all this regarding |
|---|
| 437 | recursive nameservers. |
|---|
| 438 | |
|---|
| 439 | Normally the authoritative nameservers for your regular domain will not allow |
|---|
| 440 | recursion, that is, they will return data only for the domains they are |
|---|
| 441 | authoritative for, returning pointers to nameservers for everything else. |
|---|
| 442 | |
|---|
| 443 | Here's an example of an authoritative server for blitzed.org that does not |
|---|
| 444 | allow recursion: |
|---|
| 445 | |
|---|
| 446 | $ dig www.bbc.co.uk @sou.nameserver.net |
|---|
| 447 | |
|---|
| 448 | ; <<>> DiG 9.2.3 <<>> www.bbc.co.uk @sou.nameserver.net |
|---|
| 449 | ;; global options: printcmd |
|---|
| 450 | ;; Got answer: |
|---|
| 451 | ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 38059 |
|---|
| 452 | ;; flags: qr rd; QUERY: 1, ANSWER: 0, AUTHORITY: 7, ADDITIONAL: 0 |
|---|
| 453 | |
|---|
| 454 | ;; QUESTION SECTION: |
|---|
| 455 | ;www.bbc.co.uk. IN A |
|---|
| 456 | |
|---|
| 457 | ;; AUTHORITY SECTION: |
|---|
| 458 | uk. 166680 IN NS NS1.NIC.uk. |
|---|
| 459 | uk. 166680 IN NS NS2.NIC.uk. |
|---|
| 460 | uk. 166680 IN NS NS3.NIC.uk. |
|---|
| 461 | uk. 166680 IN NS NS4.NIC.uk. |
|---|
| 462 | uk. 166680 IN NS NS5.NIC.uk. |
|---|
| 463 | uk. 166680 IN NS NSA.NIC.uk. |
|---|
| 464 | uk. 166680 IN NS NSB.NIC.uk. |
|---|
| 465 | |
|---|
| 466 | ;; Query time: 56 msec |
|---|
| 467 | ;; SERVER: 194.196.163.7#53(sou.nameserver.net) |
|---|
| 468 | ;; WHEN: Mon Feb 23 02:22:16 2004 |
|---|
| 469 | ;; MSG SIZE rcvd: 161 |
|---|
| 470 | |
|---|
| 471 | Note in the flags part the "rd", which means "recursion desired". Since this |
|---|
| 472 | server does not offer recursion to me, all it does is pass back the hostname of |
|---|
| 473 | the nameservers that can further answer my question (in this case the list of |
|---|
| 474 | nic.uk servers). My own resolver would then carry on asking questions, which |
|---|
| 475 | is how it should be. |
|---|
| 476 | |
|---|
| 477 | Look what happens if I pick a server that does allow recursion: |
|---|
| 478 | |
|---|
| 479 | $ dig www.bbc.co.uk @bos.nameserver.net |
|---|
| 480 | |
|---|
| 481 | ; <<>> DiG 9.2.3 <<>> www.bbc.co.uk @bos.nameserver.net |
|---|
| 482 | ;; global options: printcmd |
|---|
| 483 | ;; Got answer: |
|---|
| 484 | ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 51313 |
|---|
| 485 | ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 2, ADDITIONAL: 2 |
|---|
| 486 | |
|---|
| 487 | ;; QUESTION SECTION: |
|---|
| 488 | ;www.bbc.co.uk. IN A |
|---|
| 489 | |
|---|
| 490 | ;; ANSWER SECTION: |
|---|
| 491 | www.bbc.co.uk. 899 IN CNAME www.bbc.net.uk. |
|---|
| 492 | www.bbc.net.uk. 300 IN A 212.58.240.121 |
|---|
| 493 | |
|---|
| 494 | ;; AUTHORITY SECTION: |
|---|
| 495 | bbc.net.uk. 148544 IN NS ns0.thny.bbc.co.uk. |
|---|
| 496 | bbc.net.uk. 148544 IN NS ns0.thdo.bbc.co.uk. |
|---|
| 497 | |
|---|
| 498 | ;; ADDITIONAL SECTION: |
|---|
| 499 | ns0.thny.bbc.co.uk. 62144 IN A 38.160.150.20 |
|---|
| 500 | ns0.thdo.bbc.co.uk. 62144 IN A 212.58.224.20 |
|---|
| 501 | |
|---|
| 502 | ;; Query time: 321 msec |
|---|
| 503 | ;; SERVER: 203.20.52.5#53(bos.nameserver.net) |
|---|
| 504 | ;; WHEN: Mon Feb 23 02:27:14 2004 |
|---|
| 505 | ;; MSG SIZE rcvd: 164 |
|---|
| 506 | |
|---|
| 507 | Note the extra flag that appeared, "ra". This is "recursion available", and |
|---|
| 508 | the server true to its word has gone and got the information for us. |
|---|
| 509 | |
|---|
| 510 | Normally this probably would not be noticeable. In most cases it does not |
|---|
| 511 | matter if your own resolver does the work or if some other server does. This |
|---|
| 512 | geo-dns project is based totally on the IP address of the server that asks the |
|---|
| 513 | question, however, so for this application it is critical. |
|---|
| 514 | |
|---|
| 515 | As far as geo-dns is concerned, you cannot have any nameserver that allows |
|---|
| 516 | recursion be authoritative for your main domain. If you do, then any query |
|---|
| 517 | which hits this server will cause it to go out and get the answer itself. It |
|---|
| 518 | will hand back answers that are based on its own location, instead of the |
|---|
| 519 | location of the client. After wondering why too many people were getting |
|---|
| 520 | answers based on the location of bos.nameserver.net instead of their own |
|---|
| 521 | location, we finally worked out it had recursion enabled. This note is to save |
|---|
| 522 | you the same hassle. |
|---|
| 523 | |
|---|
| 524 | It is generally recommended anyway that nameservers which are meant to be |
|---|
| 525 | authoritative for domains do not have recursion enabled |
|---|
| 526 | (http://www.isc.org/pubs/tn/isc-tn-2002-2.txt), but in this case it is an |
|---|
| 527 | absolute requirement if you wish to get any sensible results. Check they are |
|---|
| 528 | not allowing recursion by use of dig as above before setting up this backend. |
|---|
| 529 | |
|---|
| 530 | (The specific example of bos.nameserver.net has since been fixed (had recursion |
|---|
| 531 | disabled) so you will not be able to repeat this example) |
|---|
| 532 | |
|---|
| 533 | Hosting multiple domains |
|---|
| 534 | ======================== |
|---|
| 535 | |
|---|
| 536 | You may have noticed that all of the instructions so far have talked only of |
|---|
| 537 | the single example domain, geo.blitzed.org, and may be wondering how to serve |
|---|
| 538 | multiple zones. The answer is, you don't need to. The only thing you want to |
|---|
| 539 | serve is individual RRs, and this geobackend does allow you to serve multiple |
|---|
| 540 | of these just by adding files to the geo-maps directory. |
|---|
| 541 | |
|---|
| 542 | So, assume you now want to apply geo-dns to the RR irc.strugglers.net, You might |
|---|
| 543 | add a new director map to your geo-maps directory named "irc.strugglers.net" |
|---|
| 544 | (name doesn't matter, just an example). Inside this file you would have |
|---|
| 545 | something like: |
|---|
| 546 | |
|---|
| 547 | $RECORD irc.strugglers.net |
|---|
| 548 | $ORIGIN iso.strugglers.net. |
|---|
| 549 | 0 bar.example.com. |
|---|
| 550 | # List of code->RR maps follow, unqualified RRs are inside |
|---|
| 551 | # iso.strugglers.net |
|---|
| 552 | |
|---|
| 553 | Once you now do a "pdns_control rediscover", your geobackend will be configured |
|---|
| 554 | to answer for irc.strugglers.net.geo.blitzed.org. Now the people who run |
|---|
| 555 | strugglers.net's dns just need to add a handy CNAME in their example.com zone |
|---|
| 556 | file: |
|---|
| 557 | |
|---|
| 558 | irc CNAME irc.strugglers.net.geo.blitzed.org. |
|---|
| 559 | |
|---|
| 560 | You can check it will work before they add the CNAME: |
|---|
| 561 | |
|---|
| 562 | $ dig +norecurse irc.strugglers.net.geo.blitzed.org. @ns0.blitzed.org |
|---|
| 563 | |
|---|
| 564 | ; <<>> DiG 9.2.3 <<>> +norecurse irc.strugglers.net.geo.blitzed.org. @ns0.blitzed.org |
|---|
| 565 | ;; global options: printcmd |
|---|
| 566 | ;; Got answer: |
|---|
| 567 | ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3592 |
|---|
| 568 | ;; flags: qr aa; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0 |
|---|
| 569 | |
|---|
| 570 | ;; QUESTION SECTION: |
|---|
| 571 | ;irc.strugglers.net.geo.blitzed.org. IN A |
|---|
| 572 | |
|---|
| 573 | ;; ANSWER SECTION: |
|---|
| 574 | irc.strugglers.net.geo.blitzed.org. 3600 IN CNAME eu.iso.blitzed.org. |
|---|
| 575 | |
|---|
| 576 | ;; Query time: 234 msec |
|---|
| 577 | ;; SERVER: 82.195.234.5#53(ns0.blitzed.org) |
|---|
| 578 | ;; WHEN: Thu Feb 26 15:45:36 2004 |
|---|
| 579 | ;; MSG SIZE rcvd: 73 |
|---|
| 580 | |
|---|
| 581 | so that is how you can geo-dnsify RRs in multiple zones. |
|---|
| 582 | |
|---|
| 583 | Other questions, corrections, etc.? |
|---|
| 584 | =================================== |
|---|
| 585 | |
|---|
| 586 | Please subscribe to our "geo" list from http://lists.blitzed.org/listinfo/geo |
|---|
| 587 | and tell us about it. Tell us about how you use this stuff too, we're |
|---|
| 588 | interested in people's experiences. |
|---|
| 589 | |
|---|
| 590 | All this is too much for you? |
|---|
| 591 | ============================= |
|---|
| 592 | |
|---|
| 593 | We're still beta-testing this ourselves. As long as it doesn't place too much |
|---|
| 594 | load on our servers we are open to the idea of doing the PowerDNS part of this |
|---|
| 595 | for you as well. We're not ready to do that yet; it will require more |
|---|
| 596 | development and specification of how you would provide your own director-maps |
|---|
| 597 | etc.. If you are interested then please subscribe to the geo list and help us |
|---|
| 598 | work out the details. |
|---|
| 599 | |
|---|
| 600 | FAQ |
|---|
| 601 | === |
|---|
| 602 | |
|---|
| 603 | Q1. My IRC network is based mostly in one country (e.g. Australia, Brazil, ...) |
|---|
| 604 | so all my users are from one ISO code and this isn't so useful to me, what |
|---|
| 605 | can I do? |
|---|
| 606 | |
|---|
| 607 | A1. Well, one of the assumptions made during the design of this backend is that |
|---|
| 608 | latency and geographical distance doesn't have a good relationship at small |
|---|
| 609 | scales, even within North America or Europe it probably does not *always* |
|---|
| 610 | follow that short distance == low latency. |
|---|
| 611 | |
|---|
| 612 | To improve things further I think you will need to do actual measurements. |
|---|
| 613 | Or you could look at the amount of hops in the AS path from your client to |
|---|
| 614 | each of your servers. Maybe you can come up with more ideas, if so we'd |
|---|
| 615 | like to hear. |
|---|
| 616 | |
|---|
| 617 | However, if you are able to get more detailed geographic info from |
|---|
| 618 | somewhere then you could still feed it into this backend, you'd just have |
|---|
| 619 | to give up on the ISO country code model. I can really only see this |
|---|
| 620 | working for very big countries like Australia and North America. |
|---|
| 621 | |
|---|
| 622 | Remember also that your servers probably are not 100% reliable! Say you |
|---|
| 623 | have 2 US servers; one in California and one in New York. You find some |
|---|
| 624 | way to split the US up into two halves with one half going to California |
|---|
| 625 | and one to New York. Now suppose the New York server dies. Half your US |
|---|
| 626 | users are still being directed there because that's the nearest one! Worse |
|---|
| 627 | still, those users probably think that what they are connecting to is a |
|---|
| 628 | *random* server, yet every single time they get will directed to a dead |
|---|
| 629 | server. They may conclude your entire network is dead. |
|---|
| 630 | |
|---|
| 631 | You would be better advised IMHO to just have one US pool with both servers |
|---|
| 632 | in. If either dies, the irc clients will make a few more attempts and get |
|---|
| 633 | the other. It might add a few more ms to the RTT, but it's better than not |
|---|
| 634 | being able to get on at all, and it's better than ending up on a non-US |
|---|
| 635 | server. Maybe one day we'll do a high-availability backend too, though. :) |
|---|
| 636 | |
|---|
| 637 | |
|---|
| 638 | Q2. Why is this tied to IRC? Can I use it for other things? |
|---|
| 639 | |
|---|
| 640 | A2. Yes! The code itself is not restricted in usefulness to only IRC, it's |
|---|
| 641 | just that IRC is a good example of a situation where you have a service |
|---|
| 642 | that is commonly split across many widely-separated servers, lots of widely |
|---|
| 643 | spread clients, and a serious lack of money for "real" solutions. |
|---|
| 644 | |
|---|
| 645 | It wouldn't be any harder to use it for things like HTTP or many other |
|---|
| 646 | protocols. |
|---|
| 647 | |
|---|
| 648 | |
|---|
| 649 | Q3. I don't want to run PowerDNS, will you make a custom backend for BIND9? |
|---|
| 650 | |
|---|
| 651 | A3. Probably not unless you're willing to pay or induce us in other ways! |
|---|
| 652 | PowerDNS is fun, you should try it, maybe you can make it work with only |
|---|
| 653 | one IP address by use of a forward zone in BIND and putting PowerDNS on a |
|---|
| 654 | strange port. |
|---|
| 655 | |
|---|
| 656 | |
|---|
| 657 | Q4. I have some ideas for other metrics you could use instead of origin |
|---|
| 658 | country, like server load or user count, or ... will you implement that |
|---|
| 659 | too? |
|---|
| 660 | |
|---|
| 661 | A4: In another backend maybe. Tell us about your ideas, if you can make them |
|---|
| 662 | sound interesting and useful then maybe we will. |
|---|
| 663 | |
|---|
| 664 | Q5. I serve lots of RRs out of my geobackend and with lots of nameservers I |
|---|
| 665 | find it hard to keep my geo-maps in sync between them all. Any hints? |
|---|
| 666 | |
|---|
| 667 | A5. At the moment we use rsync like this: |
|---|
| 668 | |
|---|
| 669 | 07 04 * * 0 pdns rsync -t rsync://calzone.hestdesign.com/countries/zz.countries.nerd.dk.rbldnsd /usr/local/etc/powerdns/zz.countries.nerd.dk.rbldnsd && /usr/local/bin/pdns_control rediscover > /dev/null |
|---|
| 670 | */15 * * * * pdns NR=$(rsync -rt --delete --stats rsync://rsync.blitzed.org/geo/ /usr/local/etc/powerdns/directormaps | awk '/Number of files transferred/ { print $5 }'); [ $NR != "0" ] && /usr/local/bin/pdns_control rediscover > /dev/null |
|---|
| 671 | |
|---|
| 672 | Which will take care of getting a new nerd.dk zone weekly and will sync the |
|---|
| 673 | geo-maps every 15 minutes, doing a rediscover if files were transferred. |
|---|
| 674 | |
|---|
| 675 | Contact |
|---|
| 676 | ======= |
|---|
| 677 | |
|---|
| 678 | Author: Andy Smith <grifferz@blitzed.org> but please direct any questions to |
|---|
| 679 | the geo list thanks! |
|---|
| 680 | |
|---|
| 681 | http://lists.blitzed.org/listinfo/geo |
|---|
| 682 | |
|---|