Mercurial > dnsbl
annotate xml/dnsbl.in @ 167:9b129ed78d7d stable-6-0-6
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
| author | carl |
|---|---|
| date | Mon, 27 Aug 2007 20:49:19 -0700 |
| parents | 5809bcdc325b |
| children | 6bac960af6b4 |
| rev | line source |
|---|---|
| 108 | 1 <reference> |
| 2 <title>@PACKAGE@ Sendmail milter - Version @VERSION@</title> | |
| 3 <partintro> | |
| 4 <title>Packages</title> | |
| 5 <para>The various source and binary packages are available at <ulink | |
|
114
f4f5fb263072
cleanup list of tlds, add trailing / on http package directory reference
carl
parents:
111
diff
changeset
|
6 url="http://www.five-ten-sg.com/@PACKAGE@/packages/">http://www.five-ten-sg.com/@PACKAGE@/packages/</ulink> |
| 108 | 7 The most recent documentation is available at <ulink |
| 8 url="http://www.five-ten-sg.com/@PACKAGE@/">http://www.five-ten-sg.com/@PACKAGE@/</ulink> | |
| 9 </para> | |
| 94 | 10 |
| 108 | 11 </partintro> |
| 94 | 12 |
| 108 | 13 <refentry id="@PACKAGE@.1"> |
| 14 <refentryinfo> | |
|
167
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
15 <date>2007-07-27</date> |
| 108 | 16 </refentryinfo> |
| 94 | 17 |
| 108 | 18 <refmeta> |
| 19 <refentrytitle>@PACKAGE@</refentrytitle> | |
| 20 <manvolnum>1</manvolnum> | |
| 21 <refmiscinfo>@PACKAGE@ @VERSION@</refmiscinfo> | |
| 22 </refmeta> | |
| 23 | |
| 24 <refnamediv id='name.1'> | |
| 25 <refname>@PACKAGE@</refname> | |
| 26 <refpurpose>a sendmail milter with per-user dnsbl filtering</refpurpose> | |
| 27 </refnamediv> | |
| 94 | 28 |
| 108 | 29 <refsynopsisdiv id='synopsis.1'> |
| 30 <title>Synopsis</title> | |
| 31 <cmdsynopsis> | |
| 32 <command>@PACKAGE@</command> | |
| 33 <arg><option>-c</option></arg> | |
| 34 <arg><option>-s</option></arg> | |
| 35 <arg><option>-d <replaceable class="parameter">n</replaceable></option></arg> | |
| 36 <arg><option>-e <replaceable class="parameter">from|to</replaceable></option></arg> | |
| 37 <arg><option>-r <replaceable class="parameter">local-domain-socket</replaceable></option></arg> | |
| 38 <arg><option>-p <replaceable class="parameter">sendmail-socket</replaceable></option></arg> | |
| 39 <arg><option>-t <replaceable class="parameter">timeout</replaceable></option></arg> | |
| 40 </cmdsynopsis> | |
| 41 </refsynopsisdiv> | |
| 94 | 42 |
| 108 | 43 <refsect1 id='options.1'> |
| 44 <title>Options</title> | |
| 45 <variablelist> | |
| 46 <varlistentry> | |
| 47 <term>-c</term> | |
| 111 | 48 <listitem><para> |
| 49 Load the configuration file, print a cannonical form | |
| 50 of the configuration on stdout, and exit. | |
| 51 </para></listitem> | |
| 108 | 52 </varlistentry> |
| 53 <varlistentry> | |
| 54 <term>-s</term> | |
| 111 | 55 <listitem><para> |
| 56 Stress test the configuration loading code by repeating | |
| 57 the load/free cycle in an infinite loop. | |
| 58 </para></listitem> | |
| 108 | 59 </varlistentry> |
| 60 <varlistentry> | |
| 61 <term>-d <replaceable class="parameter">n</replaceable></term> | |
| 111 | 62 <listitem><para> |
| 63 Set the debug level to <replaceable class="parameter">n</replaceable>. | |
| 64 </para></listitem> | |
| 108 | 65 </varlistentry> |
| 66 <varlistentry> | |
| 67 <term>-e <replaceable class="parameter">from|to</replaceable></term> | |
| 111 | 68 <listitem><para> |
| 69 Print the results of looking up the from and to addresses in the | |
| 70 current configuration. The | character is used to separate the from and to | |
| 71 addresses in the argument to the -e switch. | |
| 72 </para></listitem> | |
| 108 | 73 </varlistentry> |
| 74 <varlistentry> | |
| 75 <term>-r <replaceable class="parameter">local-domain-socket</replaceable></term> | |
| 111 | 76 <listitem><para> |
| 77 Set the local socket used for the connection to our own dns resolver processes. | |
| 78 </para></listitem> | |
| 108 | 79 </varlistentry> |
| 80 <varlistentry> | |
| 81 <term>-p <replaceable class="parameter">sendmail-socket</replaceable></term> | |
| 111 | 82 <listitem><para> |
| 83 Set the socket used for the milter connection to sendmail. This is either | |
| 84 "inet:port@ip-address" or "local:local-domain-socket-file-name". | |
| 85 </para></listitem> | |
| 108 | 86 </varlistentry> |
| 87 <varlistentry> | |
| 88 <term>-t <replaceable class="parameter">timeout</replaceable></term> | |
| 111 | 89 <listitem><para> |
| 90 Set the timeout in seconds used for communication with sendmail. | |
| 91 </para></listitem> | |
| 108 | 92 </varlistentry> |
| 93 </variablelist> | |
| 94 </refsect1> | |
| 94 | 95 |
| 111 | 96 <refsect1 id='usage.1'> |
| 108 | 97 <title>Usage</title> |
| 98 <para><command>@PACKAGE@</command> -c</para> | |
| 99 <para><command>@PACKAGE@</command> -s</para> | |
| 111 | 100 <para><command>@PACKAGE@</command> -e 'someone@aol.com|localname@mydomain.tld'</para> |
| 101 <para><command>@PACKAGE@</command> -d 10 -r resolver.sock -p local:dnsbl.sock</para> | |
| 102 </refsect1> | |
| 103 | |
| 104 <refsect1 id='installation.1'> | |
| 105 <title>Installation</title> | |
| 106 <para> | |
| 107 This is now a standard GNU autoconf/automake installation, so the normal | |
| 108 "./configure; make; su; make install" works. "make chkconfig" will | |
| 109 setup the init.d runlevel scripts. Alternatively, you can use the | |
| 110 source or binary RPMs at <ulink | |
| 111 url="http://www.five-ten-sg.com/@PACKAGE@/packages">http://www.five-ten-sg.com/@PACKAGE@/packages</ulink>. | |
| 112 </para> | |
| 113 <para> | |
| 114 Note that this has ONLY been tested on Linux, specifically RedHat Linux. | |
| 115 In particular, this milter makes no attempt to understand IPv6. Your | |
| 116 mileage will vary. You will need at a minimum a C++ compiler with a | |
| 117 minimally thread safe STL implementation. The distribution includes a | |
| 118 test.cpp program. If it fails this milter won't work. If it passes, | |
| 119 this milter might work. | |
| 120 </para> | |
| 121 <para> | |
| 122 Modify your sendmail.mc by removing all the "FEATURE(dnsbl" lines, add | |
| 123 the following line in your sendmail.mc and rebuild the .cf file | |
| 124 </para> | |
| 125 <para><screen>INPUT_MAIL_FILTER(`dnsbl', `S=local:/var/run/dnsbl/dnsbl.sock, F=T, T=C:30s;S:5m;R:5m;E:5m')</screen></para> | |
| 126 <para> | |
| 127 Modify the default <citerefentry> | |
| 128 <refentrytitle>@PACKAGE@.conf</refentrytitle> <manvolnum>5</manvolnum> | |
| 129 </citerefentry> configuration. | |
| 130 </para> | |
| 131 </refsect1> | |
| 132 | |
| 133 <refsect1 id='configuration.1'> | |
| 134 <title>Configuration</title> | |
| 135 <para> | |
| 136 The configuration file is documented in <citerefentry> | |
| 137 <refentrytitle>@PACKAGE@.conf</refentrytitle> <manvolnum>5</manvolnum> | |
| 138 </citerefentry>. Any change to the config file, or any file included | |
| 139 from that config file, will cause it to be reloaded within three | |
| 140 minutes. | |
| 141 </para> | |
| 108 | 142 </refsect1> |
| 94 | 143 |
| 108 | 144 <refsect1 id='introduction.1'> |
| 145 <title>Introduction</title> | |
| 146 <para> | |
| 147 Consider the case of a mail server that is acting as secondary MX for a | |
| 148 collection of clients, each of which has a collection of mail domains. | |
| 149 Each client may use their own collection of DNSBLs on their primary mail | |
| 150 server. We present here a mechanism whereby the backup mail server can | |
| 151 use the correct set of DNSBLs for each recipient for each message. As a | |
| 152 side-effect, it gives us the ability to customize the set of DNSBLs on a | |
| 153 per-recipient basis, so that fred@example.com could use SPEWS and the | |
| 154 SBL, where all other users @example.com use only the SBL. | |
| 155 </para> | |
| 156 <para> | |
| 157 This milter can also verify the envelope from/recipient pairs with the | |
| 158 primary MX server. This allows the backup mail servers to properly | |
| 159 reject mail sent to invalid addresses. Otherwise, the backup mail | |
| 160 servers will accept that mail, and then generate a bounce message when | |
| 161 the message is forwarded to the primary server (and rejected there with | |
| 127 | 162 no such user). These rejections are the primary cause of such backscatter. |
| 108 | 163 </para> |
| 164 <para> | |
| 165 This milter will also decode (uuencode, base64, mime, html entity, url | |
| 166 encodings) and scan for HTTP and HTTPS URLs and bare hostnames in the | |
| 167 body of the mail. If any of those host names have A or NS records on | |
| 168 the SBL (or a single configurable DNSBL), the mail will be rejected | |
| 169 unless previously whitelisted. This milter also counts the number of | |
| 170 invalid HTML tags, and can reject mail if that count exceeds your | |
| 171 specified limit. | |
| 172 </para> | |
| 173 <para> | |
| 136 | 174 This milter can also impose hourly rate limits on the number of |
| 175 recipients accepted from SMTP AUTH connections, that would otherwise be | |
| 176 allowed to relay thru this mail server with no spam filtering. | |
| 177 </para> | |
| 178 <para> | |
| 162 | 179 Consider the case of a message from A to B passing thru this milter. If |
| 180 that message is not blocked, then we might eventually see a reply | |
| 156 | 181 message from B to A. If the filtering context for A includes an |
| 162 | 182 autowhite entry, and that context does <emphasis>not</emphasis> cover B |
| 183 as a recipient, then this milter will add an entry in that file to | |
| 184 whitelist such replies for a configurable time period. Suppose A and B | |
| 185 are in the same domain, or at least use the same filtering context. In | |
| 186 that case we don't want to add a whitelist entry for B, since that would | |
| 187 then allow spammers to send mail from B (forged) to B. Such autowhite | |
| 160 | 188 files need to be writeable by the dnsbl user, where all the other dnsbl |
| 189 configuration files only need to be readable by the dnsbl user. | |
| 156 | 190 </para> |
| 191 <para> | |
| 108 | 192 The DNSBL milter reads a text configuration file (dnsbl.conf) on |
| 193 startup, and whenever the config file (or any of the referenced include | |
| 194 files) is changed. The entire configuration file is case insensitive. | |
| 195 If the configuration cannot be loaded due to a syntax error, the milter | |
| 196 will log the error and quit. If the configuration cannot be reloaded | |
| 197 after being modified, the milter will log the error and send an email to | |
| 152 | 198 root from dnsbl@$hostname. You probably want to add dnsbl@$hostname |
| 108 | 199 to your /etc/mail/virtusertable since otherwise sendmail will reject |
| 200 that message. | |
| 201 </para> | |
| 202 </refsect1> | |
| 94 | 203 |
| 111 | 204 <refsect1 id='dcc.1'> |
| 108 | 205 <title>DCC Issues</title> |
| 206 <para> | |
| 207 If you are also using the <ulink | |
| 208 url="http://www.rhyolite.com/anti-spam/dcc/">DCC</ulink> milter, there | |
| 209 are a few considerations. You may need to whitelist senders from the | |
| 210 DCC bulk detector, or from the DNS based lists. Those are two very | |
| 211 different reasons for whitelisting. The former is done thru the DCC | |
| 212 whiteclnt config file, the later is done thru the DNSBL milter config | |
| 213 file. | |
| 214 </para> | |
| 215 <para> | |
| 216 You may want to blacklist some specific senders or sending domains. | |
| 217 This could be done thru either the DCC (on a global basis, or for a | |
| 218 specific single recipient). We prefer to do such blacklisting via the | |
| 219 DNSBL milter config, since it can be done for a collection of recipient | |
| 220 mail domains. The DCC approach has the feature that you can capture the | |
| 221 entire message in the DCC log files. The DNSBL milter approach has the | |
| 222 feature that the mail is rejected earlier (at RCPT TO time), and the | |
| 223 sending machine just gets a generic "550 5.7.1 no such user" message. | |
| 224 </para> | |
| 225 <para> | |
| 226 The DCC whiteclnt file can be included in the DNSBL milter config by the | |
| 227 dcc_to and dcc_from statements. This will import the (env_to, env_from, | |
| 228 and substitute mail_host) entries from the DCC config into the DNSBL | |
| 229 config. This allows using the DCC config as the single point for | |
| 230 white/blacklisting. | |
| 231 </para> | |
| 232 <para> | |
| 233 Consider the case where you have multiple clients, each with their own | |
| 234 mail servers, and each running their own DCC milters. Each client is | |
| 235 using the DCC facilities for envelope from/to white/blacklisting. | |
| 236 Presumably you can use rsync or scp to fetch copies of your clients DCC | |
| 237 whiteclnt files on a regular basis. Your mail server, acting as a | |
| 238 backup MX for your clients, can use the DNSBL milter, and include those | |
| 239 client DCC config files. The envelope from/to white/blacklisting will | |
| 240 be appropriately tagged and used only for the domains controlled by each | |
| 241 of those clients. | |
| 242 </para> | |
| 243 </refsect1> | |
| 94 | 244 |
| 111 | 245 <refsect1 id='definitions.1'> |
| 108 | 246 <title>Definitions</title> |
| 247 <para> | |
| 248 CONTEXT - a collection of parameters that defines the filtering context | |
| 249 to be used for a collection of envelope recipient addresses. The | |
| 250 context includes such things as the list of DNSBLs to be used, and the | |
| 251 various content filtering parameters. | |
| 252 </para> | |
| 253 <para> | |
| 254 DNSBL - a named DNS based blocking list is defined by a dns suffix (e.g. | |
| 255 sbl-xbl.spamhaus.org) and a message string that is used to generate the | |
| 256 "550 5.7.1" smtp error return code. The names of these DNSBLs will be | |
| 257 used to define the DNSBL-LISTs. | |
| 258 </para> | |
| 259 <para> | |
| 260 DNSBL-LIST - a named list of DNSBLs that will be used for specific | |
| 261 recipients or recipient domains. | |
| 262 </para> | |
| 263 </refsect1> | |
| 94 | 264 |
| 111 | 265 <refsect1 id='filtering.1'> |
| 108 | 266 <title>Filtering Procedure</title> |
| 267 <para> | |
| 152 | 268 The SMTP envelope 'from' and 'to' values are used in various checks. |
| 269 The first check is to see if a reply message (swapping the env_from and | |
| 160 | 270 env_to values) would be unconditionally blocked (just based on the |
| 271 envelope from address). That check is similar to the main check | |
| 272 described below, but there is no body content to be scanned, and there | |
| 273 is no client connection ip address to be checked against DNSBLs. If | |
| 274 such a reply message would be blocked, we also block the original | |
| 275 outgoing message. This prevents folks from sending mail to recipients | |
| 276 that are unable to reply. | |
| 152 | 277 </para> |
| 278 <para> | |
| 136 | 279 If the client has authenticated with sendmail, the rate limits are |
| 280 checked. If the authenticated user has not exceeded the hourly rate | |
|
144
31ff00ea6bfb
allow parent/child to share a fully qualified env_to address
carl
parents:
140
diff
changeset
|
281 limit, then the mail is accepted, the filtering contexts are not used, |
| 136 | 282 the dns lists are not checked, and the body content is not scanned. If |
| 283 the client has not authenticated with sendmail, we follow these steps | |
| 284 for each recipient. | |
| 108 | 285 </para> |
| 286 <orderedlist> | |
| 111 | 287 <listitem><para> |
| 108 | 288 The envelope to email address is used to find an initial filtering |
| 289 context. We first look for a context that specified the full email | |
| 290 address in the env_to statement. If that is not found, we look for a | |
| 291 context that specified the entire domain name of the envelope recipient | |
| 292 in the env_to statement. If that is not found, we look for a context | |
| 293 that specified the user@ part of the envelope recipient in the env_to | |
| 294 statement. If that is not found, we use the first top level context | |
| 295 defined in the config file. | |
| 111 | 296 </para></listitem> |
| 297 <listitem><para> | |
| 108 | 298 The initial filtering context may redirect to a child context based on |
| 299 the values in the initial context's env_from statement. We look for [1) | |
| 300 the full envelope from email address, 2) the domain name part of the | |
| 301 envelope from address, 3) the user@ part of the envelope from address] | |
| 302 in that context's env_from statement, with values that point to a child | |
| 303 context. If such an entry is found, we switch to that child filtering | |
| 304 context. | |
| 111 | 305 </para></listitem> |
| 306 <listitem><para> | |
| 108 | 307 We lookup [1) the full envelope from email address, 2) the domain name |
| 308 part of the envelope from address, 3) the user@ part of the envelope | |
| 309 from address] in the filtering context env_from statement. That results | |
| 310 in one of (white, black, unknown, inherit). | |
| 111 | 311 </para></listitem> |
| 312 <listitem><para> | |
| 108 | 313 If the answer is black, mail to this recipient is rejected with "no such |
| 314 user", and the dns lists are not checked. | |
| 111 | 315 </para></listitem> |
| 316 <listitem><para> | |
| 108 | 317 If the answer is white, mail to this recipient is accepted and the dns |
| 318 lists are not checked. | |
| 111 | 319 </para></listitem> |
| 320 <listitem><para> | |
| 108 | 321 If the answer is unknown, we don't reject yet, but the dns lists will be |
| 322 checked, and the content may be scanned. | |
| 111 | 323 </para></listitem> |
| 324 <listitem><para> | |
| 108 | 325 If the answer is inherit, we repeat the envelope from search in the |
| 326 parent context. | |
| 111 | 327 </para></listitem> |
| 328 <listitem><para> | |
| 108 | 329 The dns lists specified in the filtering context are checked and the |
| 330 mail is rejected if any list has an A record for the standard dns based | |
| 331 lookup scheme (reversed octets of the client followed by the dns | |
| 332 suffix). | |
| 111 | 333 </para></listitem> |
| 334 <listitem><para> | |
| 108 | 335 If the mail has not been accepted or rejected yet, we look for a |
| 336 verification context, which is the closest ancestor of the filtering | |
| 337 context that both specifies a verification host, and which covers the | |
| 338 envelope to address. If we find such a verification context, and the | |
| 339 verification host is not our own hostname, we open an smtp conversation | |
| 340 with that verification host. The current envelope from and recipient to | |
| 341 values are passed to that verification host. If we receive a 5xy | |
| 342 response those commands, we reject the current recipient with "no such | |
| 343 user". | |
| 111 | 344 </para></listitem> |
| 345 <listitem><para> | |
| 108 | 346 If the mail has not been accepted or rejected yet, and the filtering |
| 347 context enables content filtering, and this is the first such recipient | |
| 348 in this smtp transaction, we set the content filtering parameters from | |
| 349 this context, and enable content filtering for the body of this message. | |
| 111 | 350 </para></listitem> |
| 108 | 351 </orderedlist> |
| 352 <para> | |
| 160 | 353 For each recipient that was accepted, we search for an autowhite entry |
| 354 starting in the reply filtering context. If an autowhite entry is found, | |
| 355 we add the recipient to that auto whitelist file. This will prevent reply | |
| 356 messages from being blocked by the dnsbl or content filtering. | |
| 357 </para> | |
| 358 <para> | |
| 108 | 359 If content filtering is enabled for this body, the mail text is decoded |
| 119 | 360 (uuencode, base64, mime, html entity, url encodings), and scanned for HTTP |
| 361 and HTTPS URLs or bare host names. Hostnames must be either ip address | |
| 362 literals, or must end in a string defined by the TLD list. The first | |
| 363 <configurable> host names are checked as follows. | |
| 364 </para> | |
| 365 <para> | |
| 366 The only known list that is suitable for the content filter DNSBL is the | |
| 367 SBL. If the content filter DNSBL is defined, and any of those host | |
| 368 names resolve to ip addresses that are on that DNSBL (or have | |
| 369 nameservers that are on that list), and the host name is not on the | |
| 370 <configurable> ignore list, the mail is rejected. | |
| 371 </para> | |
| 372 <para> | |
| 373 If the content uribl DNSBL is defined, and any of those host names are | |
| 374 on that DNSBL, and the host name is not on the <configurable> | |
| 375 ignore list, the mail is rejected. | |
| 376 </para> | |
| 377 <para> | |
|
167
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
378 If any non-whitelisted recipient has a filtering context with a non-zero |
|
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
379 spamassassin limit, then the message is passed thru spamassassin (via |
|
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
380 spamc), and the message is rejected for those recipients with spamassassin |
|
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
381 limits less than the resulting spamassassin score. |
| 163 | 382 </para> |
| 383 <para> | |
| 119 | 384 We also scan for excessive bad html tags, and if a <configurable> |
| 385 limit is exceeded, the mail is rejected. | |
| 108 | 386 </para> |
| 387 </refsect1> | |
| 94 | 388 |
| 111 | 389 <refsect1 id='access.1'> |
| 108 | 390 <title>Sendmail access vs. DNSBL</title> |
| 391 <para> | |
| 392 With the standard sendmail.mc dnsbl FEATURE, the dnsbl checks may be | |
| 393 suppressed by entries in the /etc/mail/access database. For example, | |
| 394 suppose you control a /18 of address space, and have allocated some /24s | |
| 395 to some clients. You have access entries like | |
| 111 | 396 <literallayout class="monospaced"><![CDATA[ |
| 397 192.168.4 OK | |
| 398 192.168.17 OK]]></literallayout> | |
| 108 | 399 </para> |
| 400 <para> | |
| 401 to allow those clients to smarthost thru your mail server. Now if one | |
| 402 of those clients happens get infected with a virus that turns a machine | |
| 403 into an open proxy, and their 192.168.4.45 lands on the SBL-XBL, you | |
| 404 will still wind up allowing that infected machine to smarthost thru your | |
| 405 mail servers. | |
| 406 </para> | |
| 407 <para> | |
| 408 With this DNSBL milter, the sendmail access database cannot override the | |
| 409 dnsbl checks, so that machine won't be able to send mail to or thru your | |
| 410 smarthost mail server (unless the virus/proxy can use smtp-auth). | |
| 411 </para> | |
| 412 <para> | |
| 413 Using the standard sendmail features, you would add access entries to | |
| 414 allow hosts on your local network to relay thru your mail server. Those | |
| 415 OK entries in the sendmail access database will override all the dnsbl | |
| 416 checks. With this DNSBL milter, you will need to have the local users | |
| 417 authenticate with smtp-auth to get the same effect. You might find | |
| 418 <ulink | |
| 419 url="http://www.ists.dartmouth.edu/classroom/sendmail-ssl-how-to.php"> | |
| 420 these directions</ulink> helpful for setting up smtp-auth if you are on | |
| 421 RH Linux. | |
| 422 </para> | |
| 423 </refsect1> | |
| 94 | 424 |
| 111 | 425 <refsect1 id='performance.1'> |
| 108 | 426 <title>Performance Issues</title> |
| 427 <para> | |
| 428 Consider a high volume high performance machine running sendmail. Each | |
| 429 sendmail process can do its own dns resolution. Typically, such dns | |
| 430 resolver libraries are not thread safe, and so must be protected by some | |
| 431 sort of mutex in a threaded environment. When we add a milter to | |
| 432 sendmail, we now have a collection of sendmail processes, and a | |
| 433 collection of milter threads. | |
| 434 </para> | |
| 435 <para> | |
| 436 We will be doing a lot of dns lookups per mail message, and at least | |
| 437 some of those will take many tens of seconds. If all this dns work is | |
| 438 serialized inside the milter, we have an upper limit of about 25K mail | |
| 439 messages per day. That is clearly not sufficient for many sites. | |
| 440 </para> | |
| 441 <para> | |
| 442 Since we want to do parallel dns resolution across those milter threads, | |
| 443 we add another collection of dns resolver processes. Each sendmail | |
| 444 process is talking to a milter thread over a socket, and each milter | |
| 445 thread is talking to a dns resolver process over another socket. | |
| 446 </para> | |
| 447 <para> | |
| 448 Suppose we are processing 20 messages per second, and each message | |
| 449 requires 20 seconds of dns work. Then we will have 400 sendmail | |
| 450 processes, 400 milter threads, and 400 dns resolver processes. Of | |
| 451 course that steady state is very unlikely to happen. | |
| 452 </para> | |
| 453 </refsect1> | |
| 94 | 454 |
| 455 | |
| 111 | 456 <refsect1 id='rejected.1'> |
| 108 | 457 <title>Rejected Ideas</title> |
| 458 <para> | |
| 459 The following ideas have been considered and rejected. | |
| 460 </para> | |
| 461 <para> | |
| 111 | 462 Add max_recipients setting to the context configuration. Recipients in |
| 463 excess of that limit will be rejected, and all the non-whitelisted | |
| 464 recipients will be removed. Current spammers *very* rarely send more | |
| 465 than ten recipients in a single smtp transaction, so this won't stop any | |
| 108 | 466 significant amount of spam. |
| 467 </para> | |
| 468 <para> | |
| 469 Add poison addresses to the configuration. If any recipient is | |
| 470 poison, all recipients are rejected even if they would be whitelisted, | |
| 471 and the data is rejected if sent. I have a collection of spam trap | |
| 472 addresses that would be suitable for such use. Based on my log files, | |
| 473 any mail to those spam trap addresses is rejected based on either dnsbl | |
| 474 lookups or the DCC. So this won't result in blocking any additional | |
| 475 spam. | |
| 476 </para> | |
| 477 <para> | |
| 478 Add an option to only allow one recipient if the return path is | |
| 479 empty. Based on my log files, there is no mail that violates this | |
| 480 check. | |
| 481 </para> | |
| 482 <para> | |
| 483 Reject the mail if the envelope from domain name contains any MX | |
| 484 records pointing to 127.0.0.0/8. I don't see any significant amount of | |
| 485 spam sent with such domain names. | |
| 486 </para> | |
| 487 </refsect1> | |
| 94 | 488 |
| 108 | 489 <refsect1 id='todo.1'> |
| 490 <title>TODO</title> | |
| 491 <para> | |
| 492 The following ideas are under consideration. | |
| 493 </para> | |
| 494 <para> | |
| 495 Add a per-context option to reject mail if the number of digits in | |
| 496 the reverse dns client name exceeds some threshold. | |
| 497 </para> | |
| 115 | 498 <para> |
| 499 Look for href="hostname/path" strings that are missing the required | |
| 500 http:// protocol header. Such references are still clickable in common | |
| 501 mail software. | |
| 502 </para> | |
| 108 | 503 </refsect1> |
| 94 | 504 |
| 111 | 505 <refsect1 id='copyright.1'> |
| 108 | 506 <title>Copyright</title> |
| 507 <para> | |
| 163 | 508 Copyright (C) 2007 by 510 Software Group <carl@five-ten-sg.com> |
| 108 | 509 </para> |
| 510 <para> | |
| 511 This program is free software; you can redistribute it and/or modify it | |
| 512 under the terms of the GNU General Public License as published by the | |
| 160 | 513 Free Software Foundation; either version 3, or (at your option) any |
| 108 | 514 later version. |
| 515 </para> | |
| 516 <para> | |
| 517 You should have received a copy of the GNU General Public License along | |
| 518 with this program; see the file COPYING. If not, please write to the | |
| 519 Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. | |
| 520 </para> | |
| 521 </refsect1> | |
| 94 | 522 |
| 111 | 523 <refsect1 id='version.1'> |
| 524 <title>CVS Version</title> | |
| 108 | 525 <para> |
| 526 $Id$ | |
| 527 </para> | |
| 528 </refsect1> | |
| 529 </refentry> | |
| 530 | |
| 531 | |
| 532 <refentry id="@PACKAGE@.conf.5"> | |
| 533 <refentryinfo> | |
|
167
9b129ed78d7d
actually use spamassassin result, allow build without spam assassin, only call it if some recipient needs it.
carl
parents:
164
diff
changeset
|
534 <date>2007-07-27</date> |
| 108 | 535 </refentryinfo> |
| 94 | 536 |
| 108 | 537 <refmeta> |
| 538 <refentrytitle>@PACKAGE@.conf</refentrytitle> | |
| 539 <manvolnum>5</manvolnum> | |
| 540 <refmiscinfo>@PACKAGE@ @VERSION@</refmiscinfo> | |
| 541 </refmeta> | |
| 94 | 542 |
| 108 | 543 <refnamediv id='name.5'> |
| 544 <refname>@PACKAGE@.conf</refname> | |
| 111 | 545 <refpurpose>configuration file for @PACKAGE@ sendmail milter</refpurpose> |
| 108 | 546 </refnamediv> |
| 547 | |
| 548 <refsynopsisdiv id='synopsis.5'> | |
| 549 <title>Synopsis</title> | |
| 550 <cmdsynopsis> | |
| 551 <command>@PACKAGE@.conf</command> | |
| 552 </cmdsynopsis> | |
| 553 </refsynopsisdiv> | |
| 94 | 554 |
| 108 | 555 <refsect1 id='description.5'> |
| 556 <title>Description</title> | |
| 557 <para>The <command>@PACKAGE@.conf</command> configuration file is | |
|
148
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
558 specified by this partial bnf description. Comments start with // |
|
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
559 or # and extend to the end of the line. To include the contents |
|
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
560 of some file verbatim in the dnsbl.conf file, use |
|
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
561 <literallayout class="monospaced"><![CDATA[include "<file>";]]></literallayout> |
|
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
562 </para> |
| 108 | 563 |
| 564 <literallayout class="monospaced"><![CDATA[ | |
| 565 CONFIG = {CONTEXT ";"}+ | |
| 566 CONTEXT = "context" NAME "{" {STATEMENT}+ "}" | |
| 153 | 567 STATEMENT = (DNSBL | DNSBLLIST | CONTENT | ENV-TO | VERIFY | |
| 568 AUTOWHITE | CONTEXT | ENV-FROM | RATE-LIMIT) ";" | |
| 108 | 569 |
| 124 | 570 DNSBL = "dnsbl" NAME DNSPREFIX ERROR-MSG1 |
| 108 | 571 |
| 572 DNSBLLIST = "dnsbl_list" {NAME}+ | |
| 94 | 573 |
| 108 | 574 CONTENT = "content" ("on" | "off") "{" {CONTENT-ST}+ "}" |
| 119 | 575 CONTENT-ST = (FILTER | URIBL | IGNORE | TLD | CCTLD | HTML-TAGS | |
| 163 | 576 HTML-LIMIT | HOST-LIMIT | SPAMASS) ";" |
| 577 SPAMASS = "spamassassin" INTEGER | |
| 124 | 578 FILTER = "filter" DNSPREFIX ERROR-MSG2 |
| 579 URIBL = "uribl" DNSPREFIX ERROR-MSG3 | |
| 108 | 580 IGNORE = "ignore" "{" {HOSTNAME [";"]}+ "}" |
| 581 TLD = "tld" "{" {TLD [";"]}+ "}" | |
| 119 | 582 CCTLD = "cctld" "{" {TLD [";"]}+ "}" |
| 108 | 583 HTML-TAGS = "html_tags" "{" {HTMLTAG [";"]}+ "}" |
| 124 | 584 ERROR-MSG1 = string containing exactly two %s replacement tokens |
| 585 both are replaced with the client ip address | |
| 586 ERROR-MSG2 = string containing exactly two %s replacement tokens | |
| 587 the first is replaced with the hostname, and the second | |
| 588 is replaced with the ip address | |
| 589 ERROR-MSG3 = string containing exactly two %s replacement tokens | |
| 590 both are replaced with the hostname | |
| 108 | 591 |
| 592 HTML-LIMIT = "html_limit" ("on" INTEGER ERROR-MSG | "off") | |
| 593 | |
| 111 | 594 HOST-LIMIT = "host_limit" ("on" INTEGER ERROR-MSG | "off" | |
| 595 "soft" INTEGER) | |
| 94 | 596 |
| 108 | 597 ENV-TO = "env_to" "{" {(TO-ADDR | DCC-TO)}+ "}" |
| 598 TO-ADDR = ADDRESS [";"] | |
| 599 DCC-TO = "dcc_to" ("ok" | "many") "{" DCCINCLUDEFILE "}" ";" | |
| 600 | |
| 601 VERIFY = "verify" HOSTNAME ";" | |
| 153 | 602 AUTOWHITE = "autowhite" DAYS FILENAME ";" |
| 108 | 603 |
| 604 ENV_FROM = "env_from" [DEFAULT] "{" {(FROM-ADDR | DCC-FROM)}+ "}" | |
| 605 FROM-ADDR = ADDRESS VALUE [";"] | |
| 606 DCC-FROM = "dcc_from" "{" DCCINCLUDEFILE "}" ";" | |
| 136 | 607 |
| 140 | 608 RATE-LIMIT = "rate_limit" [DEFAULTLIMIT] "{" (RATE)+ "}" |
| 136 | 609 RATE = USER LIMIT [";"] |
| 610 | |
| 108 | 611 DEFAULT = ("white" | "black" | "unknown" | "inherit" | "") |
| 612 ADDRESS = (USER@ | DOMAIN | USER@DOMAIN) | |
|
148
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
613 VALUE = ("white" | "black" | "unknown" | "inherit" | CHILD-CONTEXT-NAME)]]></literallayout> |
| 108 | 614 </refsect1> |
| 94 | 615 |
| 108 | 616 <refsect1 id='sample.5'> |
| 617 <title>Sample</title> | |
| 618 <literallayout class="monospaced"><![CDATA[ | |
| 127 | 619 context main-default { |
| 620 // outbound dnsbl filtering to catch our own customers that end up on the sbl | |
|
144
31ff00ea6bfb
allow parent/child to share a fully qualified env_to address
carl
parents:
140
diff
changeset
|
621 dnsbl localp partial.blackholes.five-ten-sg.com "Mail from %s rejected - local; see http://www.five-ten-sg.com/blackhole.php?%s"; |
| 127 | 622 dnsbl local blackholes.five-ten-sg.com "Mail from %s rejected - local; see http://www.five-ten-sg.com/blackhole.php?%s"; |
| 623 dnsbl sbl sbl-xbl.spamhaus.org "Mail from %s rejected - sbl; see http://www.spamhaus.org/query/bl?ip=%s"; | |
| 624 dnsbl dul dul.dnsbl.sorbs.net "Mail from %s rejected - dul; see http://www.sorbs.net/lookup.shtml?%s"; | |
| 625 dnsbl_list local sbl dul; | |
| 626 | |
| 627 // outbound content filtering to prevent our own customers from sending spam | |
| 628 content on { | |
| 629 filter sbl-xbl.spamhaus.org "Mail containing %s rejected - sbl; see http://www.spamhaus.org/query/bl?ip=%s"; | |
| 630 uribl multi.surbl.org "Mail containing %s rejected - surbl; see http://www.rulesemporium.com/cgi-bin/uribl.cgi?bl0=1&domain0=%s"; | |
| 631 #uribl black.uribl.com "Mail containing %s rejected - uribl; see http://l.uribl.com/?d=%s"; | |
| 632 ignore { include "hosts-ignore.conf"; }; | |
| 633 tld { include "tld.conf"; }; | |
| 634 cctld { include "cctld.conf"; }; | |
| 635 html_tags { include "html-tags.conf"; }; | |
| 636 html_limit on 20 "Mail containing excessive bad html tags rejected"; | |
| 637 html_limit off; | |
| 638 host_limit on 20 "Mail containing excessive host names rejected"; | |
| 639 host_limit soft 20; | |
| 640 }; | |
| 641 | |
| 642 // backscatter prevention - don't send bounces for mail that we accepted but could not forward | |
| 643 // we only send bounces to our own customers | |
| 644 env_from unknown { | |
| 645 "<>" black; | |
| 646 }; | |
| 136 | 647 |
|
144
31ff00ea6bfb
allow parent/child to share a fully qualified env_to address
carl
parents:
140
diff
changeset
|
648 // hourly recipient rate limit by smtp auth client id |
| 140 | 649 rate_limit 30 { // default |
| 136 | 650 fred 100; // override default limits |
| 140 | 651 joe 10; // "" |
| 136 | 652 }; |
| 127 | 653 }; |
| 654 | |
| 108 | 655 context sample { |
| 656 dnsbl local blackholes.five-ten-sg.com "Mail from %s rejected - local; see http://www.five-ten-sg.com/blackhole.php?%s"; | |
| 657 dnsbl sbl sbl-xbl.spamhaus.org "Mail from %s rejected - sbl; see http://www.spamhaus.org/query/bl?ip=%s"; | |
| 658 dnsbl xbl xbl.spamhaus.org "Mail from %s rejected - xbl; see http://www.spamhaus.org/query/bl?ip=%s"; | |
| 659 dnsbl dul dul.dnsbl.sorbs.net "Mail from %s rejected - dul; see http://www.sorbs.net/lookup.shtml?%s"; | |
| 660 dnsbl_list local sbl dul; | |
| 94 | 661 |
| 108 | 662 content on { |
| 663 filter sbl-xbl.spamhaus.org "Mail containing %s rejected - sbl; see http://www.spamhaus.org/query/bl?ip=%s"; | |
| 122 | 664 uribl multi.surbl.org "Mail containing %s rejected - surbl; see http://www.rulesemporium.com/cgi-bin/uribl.cgi?bl0=1&domain0=%s"; |
| 119 | 665 #uribl black.uribl.com "Mail containing %s rejected - uribl; see http://l.uribl.com/?d=%s"; |
| 108 | 666 ignore { include "hosts-ignore.conf"; }; |
| 667 tld { include "tld.conf"; }; | |
| 119 | 668 cctld { include "cctld.conf"; }; |
| 108 | 669 html_tags { include "html-tags.conf"; }; |
| 670 html_limit on 20 "Mail containing excessive bad html tags rejected"; | |
| 671 html_limit off; | |
| 672 host_limit on 20 "Mail containing excessive host names rejected"; | |
| 673 host_limit soft 20; | |
| 674 }; | |
| 94 | 675 |
| 108 | 676 env_to { |
| 677 # child contexts are not allowed to specify recipient addresses outside these domains | |
| 678 # leave this outer global context env_to empty to allow arbitrary recipients in child contexts | |
| 679 mydomain.com; | |
| 680 customer1.com; | |
| 681 customer1a.com; | |
| 682 customer1b.com; | |
| 683 customer2.com; | |
| 684 customer2a.com; | |
| 685 customer2b.com; | |
| 686 }; | |
| 94 | 687 |
| 108 | 688 context whitelist { |
| 689 content off {}; | |
| 690 env_to { | |
| 691 # dcc_to ok { include "/var/dcc/whitecommon"; }; # copy the dcc OK values (env_to) into this context | |
| 692 }; | |
| 693 env_from white {}; # white forces all unmatched from addresses (everyone in this case) to be whitelisted | |
| 694 # so all mail TO these env_to addresses is accepted | |
| 695 }; | |
| 94 | 696 |
| 108 | 697 context minimal { |
| 698 dnsbl_list sbl dul; | |
| 699 content on {}; | |
| 700 env_to { | |
| 701 sales@mydomain.com; | |
| 702 }; | |
| 703 }; | |
| 94 | 704 |
| 108 | 705 context blacklist { |
| 706 env_to { | |
| 707 dcc_to many { include "/var/dcc/whitecommon"; }; # copy the dcc MANY values (env_to) into this context | |
| 708 old-employee@mydomain.com; | |
| 709 }; | |
| 710 env_from black {}; # black forces all unmatched from addresses (everyone in this case) to be blacklisted | |
| 711 # so all mail TO these env_to addresses is rejected | |
| 712 }; | |
| 94 | 713 |
| 108 | 714 context vp { # special context for the vp |
| 715 env_to { | |
| 716 vp@mydomain.com; | |
| 717 }; | |
| 718 env_from inherit { | |
| 719 nai.com black; # the vp does not like nai | |
| 720 yahoo.com unknown; # override parent context blacklisting | |
| 721 mother@spammyisp.com white; # suppress dnsbl checking | |
| 722 }; | |
| 723 }; | |
| 724 | |
| 725 context customer1 { | |
| 726 dnsbl_list sbl dul; | |
| 727 env_to { | |
| 728 customer1.com; | |
| 729 customer1a.com; | |
| 730 customer1b.com; | |
| 731 }; | |
| 94 | 732 |
| 153 | 733 # we can reject unknown users |
| 108 | 734 verify mail.customer1.com; |
| 94 | 735 |
| 153 | 736 # whitelist anyone to whom we have sent mail in the last 90 days |
| 737 autowhite 90 "autowhite/customer1"; | |
| 738 | |
| 108 | 739 context customer1a { |
| 740 env_to { | |
| 741 customer1a.com; | |
|
148
9330b8d6a56b
add documentation fixes, allow env_from target of inherit
carl
parents:
144
diff
changeset
|
742 }; |
| 108 | 743 env_from black { # blacklist everything |
| 744 first@acceptable.com unknown; # except these specific envelope senders | |
| 745 second@another.com unknown; | |
| 746 yahoo.com inherit; # delegate to the parent | |
| 747 }; | |
| 748 }; | |
| 749 | |
| 750 env_from { # default value of the default is inherit | |
| 751 yahoo.com black; # no mail from yahoo | |
| 752 first@yahoo.com unknown; # except this one | |
| 753 }; | |
| 754 }; | |
| 94 | 755 |
| 108 | 756 context customer2 { |
| 757 dnsbl_list sbl; | |
| 758 env_to { | |
| 759 customer2.com; | |
| 760 customer2a.com; | |
| 761 customer2b.com; | |
| 762 }; | |
| 763 }; | |
| 104 | 764 |
| 149 | 765 # this is at the end, so that these abuse@ and postmaster@ entries will |
| 766 # override any conflicting entries inside the customer contexts. | |
| 767 context abuse { | |
| 768 dnsbl_list xbl; | |
| 769 content off {}; | |
| 770 env_to { | |
| 771 abuse@; # no content filtering on abuse reports | |
| 772 postmaster@; # "" | |
| 773 }; | |
| 774 env_from unknown {}; # ignore all parent white/black listing | |
| 775 }; | |
| 776 | |
| 108 | 777 env_from unknown { |
| 778 dcc_from { include "/var/dcc/whitecommon"; }; # copy the dcc OK/MANY values (env_from, substitute mail_host) into this context | |
| 779 abuse@ abuse; # replies to abuse reports use the abuse context | |
| 780 yahoo.com black; # don't take mail from yahoo | |
| 781 spammer@example.com black; | |
| 782 }; | |
| 783 };]]></literallayout> | |
| 784 </refsect1> | |
| 94 | 785 |
| 111 | 786 <refsect1 id='version.5'> |
| 787 <title>CVS Version</title> | |
| 108 | 788 <para> |
| 789 $Id$ | |
| 790 </para> | |
| 791 </refsect1> | |
| 792 | |
| 793 </refentry> | |
| 794 </reference> |