1 <!doctype HTML public "-//W3O//DTD W3 HTML 3.2//EN">
4 <TITLE>The Fetchmail FAQ</TITLE>
5 <link rev=made href="mailto:esr@snark.thyrsus.com">
6 <meta name="description" content="Frequently asked questions about fetchmail.">
7 <meta name="keywords" content="fetchmail, POP, POP2, POP3, IMAP, remote mail">
10 <table width="100%" cellpadding=0><tr>
11 <td width="30%">Back to <a href="index.html">Fetchmail Home Page</a>
12 <td width="30%" align=center>To <a href="/~esr/sitemap.html">Site Map</a>
13 <td width="30%" align=right>$Date: 2001/02/12 01:14:27 $
16 <H1>Frequently Asked Questions About Fetchmail</H1>
18 Before reporting any bug, please read <a href="#G3">G3</a> for advice
19 on how to include diagnostic information that will get your bug fixed
20 as quickly as possible. <p>
22 If you have a question or answer you think ought to be added to this FAQ list,
23 mail it to fetchmail's maintainer, Eric S. Raymond, at
24 <A HREF="mailto:esr@thyrsus.com">esr@snark.thyrsus.com</A>.<p>
26 <h1>General questions:</h1>
28 <a href="#G1">G1. What is fetchmail and why should I bother?</a><br>
29 <a href="#G2">G2. Where do I find the latest FAQ and fetchmail sources?</a><br>
30 <a href="#G3">G3. I think I've found a bug. Will you fix it?</a><br>
31 <a href="#G4">G4. I have this idea for a neat feature. Will you add it?</a><br>
32 <a href="#G5">G5. Is there a mailing list for exchanging tips?</a><br>
33 <a href="#G6">G6. So, what's this I hear about a fetchmail paper?</a><br>
34 <a href="#G7">G7. What is the best server to use with fetchmail?</a><br>
35 <a href="#G8">G8. What is the best mail program to use with fetchmail?</a><br>
36 <a href="#G9">G9. How can I avoid sending my password en clair?</a><br>
37 <a href="#G10">G10. Is any special configuration needed to use a dynamic
39 <a href="#G11">G11. Is any special configuration needed to use firewalls?</a><br>
40 <a href="#G12">G12. Is any special configuration needed to <em>send</em> mail?</a><br>
41 <a href="#G13">G13. Is fetchmail Y2K-compliant?</a><br>
42 <a href="#G14">G14. Is there a way in fetchmail to support disconnected IMAP mode?</a><br>
43 <a href="#G15">G15. How will fetchmail perform under heavy loads?</a><br>
45 <h1>Build-time problems:</h1>
47 <a href="#B1">B1. Lex bombs out while building the fetchmail lexer.</a><br>
48 <a href="#B2">B2. I get link failures when I try to build fetchmail.</a><br>
50 <h1>Fetchmail configuration file grammar questions:</h1>
52 <a href="#F1">F1. Why does my old .fetchmailrc no longer work?</a><br>
53 <a href="#F2">F2. The .fetchmailrc parser won't accept my all-numeric user name.</a><br>
54 <a href="#F3">F3. The .fetchmailrc parser won't accept my host or username beginning with `no'.</a><br>
55 <a href="#F4">F4. I'm getting a `parse error' message I don't understand.</a><br>
57 <h1>Configuration questions:</h1>
59 <a href="#C1">C1. Why do I need a .fetchmailrc when running as root on my own machine?</a><br>
60 <a href="#C2">C2. How can I arrange for a fetchmail daemon to get killed when I log out?</a><br>
61 <a href="#C3">C3. How do I know what interface and address to use with --interface?</a><br>
62 <a href="#C4">C4. How can I set up support for sendmail's anti-spam features?</a><br>
63 <a href="#C5">C5. How can I poll some of my mailboxes more/less often than others?</a><br>
64 <a href="#C6">C6. Fetchmail works OK started up manually, but not from an init script.</a><br>
65 <a href="#C7">C7. How can I forward mail to another host?.</a><br>
67 <h1>How to make fetchmail play nice with various MTAs:</h1>
69 <a href="#T1">T1. How can I use fetchmail with sendmail?</a><br>
70 <a href="#T2">T2. How can I use fetchmail with qmail?</a><br>
71 <a href="#T3">T3. How can I use fetchmail with exim?</a><br>
72 <a href="#T4">T4. How can I use fetchmail with smail?</a><br>
73 <a href="#T5">T5. How can I use fetchmail with SCO's MMDF?</a><br>
74 <a href="#T6">T6. How can I use fetchmail with Lotus Notes?</a><br>
76 <h1>How to make fetchmail work with various servers:</h1>
78 <a href="#S1">S1. How can I use fetchmail with qpopper?</a><br>
79 <a href="#S2">S2. How can I use fetchmail with Microsoft Exchange?</a><br>
80 <a href="#S3">S3. How can I use fetchmail with Compuserve RPA?</a><br>
81 <a href="#S4">S4. How can I use fetchmail with Demon Internet's SDPS?</a><br>
82 <a href="#S5">S5. How can I use fetchmail with usa.net's servers?</a><br>
83 <a href="#S6">S6. How can I use fetchmail with HP OpenMail?</a><br>
84 <a href="#S7">S7. How can I use fetchmail with geocities POP3 servers?</a><br>
85 <a href="#S8">S8. How can I use fetchmail with Hotmail?</a><br>
86 <a href="#S9">S9. How can I use fetchmail with MSN?</a><br>
87 <a href="#S10">S10. How can I use fetchmail with SpryNet?</a><br>
88 <a href="#S11">S11. How can I use fetchmail with FTGate?</a><br>
89 <a href="#S12">S12. How can I use fetchmail with MailMax?</a><br>
90 <a href="#S13">S13. How can I use fetchmail with Novell GroupWise?</a><br>
91 <a href="#S14">S14. How can I use fetchmail with InterChange?</a><br>
93 <h1>How to set up well-known security and authentication methods:</h1>
95 <a href="#K1">K1. How can I use fetchmail with SOCKS?</a><br>
96 <a href="#K2">K2. How can I use fetchmail with IPv6 and IPsec?</a><br>
97 <a href="#K3">K3. How can I get fetchmail to work with ssh?</a><br>
98 <a href="#K4">K4. What do I have to do to use the IMAP-GSS protocol?</a><br>
99 <a href="#K5">K5. How can I use fetchmail with SSL?</a><br>
101 <h1>Runtime fatal errors:</h1>
103 <a href="#R1">R1. Fetchmail isn't working, and -v shows `SMTP connect failed' messages.</a><br>
104 <a href="#R2">R2. When I try to configure an MDA, fetchmail doesn't work.</a><br>
105 <a href="#R3">R3. Fetchmail dumps core when given an invalid rc file.</a><br>
106 <a href="#R4">R4. Fetchmail dumps core in -V mode, but operates normally otherwise.</a><br>
107 <a href="#R5">R5. Running fetchmail in daemon mode doesn't work.</a><br>
108 <a href="#R6">R6. Fetchmail hangs when used with pppd.</a><br>
109 <a href="#R7">R7. Fetchmail randomly dies with socket errors.</a><br>
110 <a href="#R8">R8. Fetchmail running as root stopped working after an OS upgrade</a><br>
111 <a href="#R9">R9. Fetchmail is timing out after fetching certain
112 messages but before deleting them</a><br>
113 <a href="#R10">R10. Fetchmail is timing out during message fetches</a><br>
114 <a href="#R11">R11. Fetchmail is dying with SIGPIPE.</a><br>
116 <h1>Disappearing mail</h1>
118 <a href="#D1">D1. I think I've set up fetchmail correctly, but I'm not getting any mail.</a><br>
119 <a href="#D2">D2. All my mail seems to disappear after a dropped connection.</a><br>
120 <a href="#D3">D3. Mail that was being fetched when I interrupted my fetchmail seems to have been vanished.</a><br>
122 <h1>Multidrop-mode problems:</h1>
124 <a href="#M1">M1. I've declared local names, but all my multidrop mail is going to root anyway.</a><br>
125 <a href="#M2">M2. I can't seem to get fetchmail to route to a local domain properly.</a><br>
126 <a href="#M3">M3. I tried to run a mailing list using multidrop, and I have a mail loop!</a><br>
127 <a href="#M4">M4. My multidrop fetchmail seems to be having DNS problems.</a><br>
128 <a href="#M5">M5. I'm seeing long DNS delays before each message is processed.</a><br>
129 <a href="#M6">M6. How do I get multidrop mode to work with majordomo?</a><br>
130 <a href="#M7">M7. Multidrop mode isn't parsing envelope addresses from
131 my Received headers as it should.</a><br>
132 <a href="#M8">M8. Users are getting multiple copies of messages.</a><br>
134 <h1>Mangled mail:</h1>
136 <a href="#X1">X1. Spurious blank lines are appearing in the headers of fetched mail.</a><br>
137 <a href="#X2">X2. My mail client can't see a Subject line.</a><br>
138 <a href="#X3">X3. Messages containing "From" at start of line are being split.</a><br>
139 <a href="#X4">X4. My mail is being mangled in a new and different way.</a><br>
140 <a href="#X5">X5. Using POP3, retrievals seems to be fetching too much!</a><br>
141 <a href="#X6">X6. My mail attachments are being dropped or mangled.</a><br>
142 <a href="#X7">X7. Some mail attachments are hanging fetchmail.</a><br>
144 <h1>Other problems:</h1>
146 <a href="#O1">O1. The --logfile option doesn't work if the logfile doesn't exist.</a><br>
147 <a href="#O2">O2. Every time I get a POP or IMAP message the header is
148 dumped to all my terminal sessions.</a><br>
149 <a href="#O3">O3. Does fetchmail reread its rc file every poll cycle?</a><br>
150 <a href="#O4">O4. Why do deleted messages show up again when I take
151 a line hit while downloading?</a><br>
152 <a href="#O5">O5. Why is fetched mail being logged with my name, not the real From address?</a><br>
153 <a href="#O6">O6. I'm seeing long sendmail delays or hangs near the start of each poll cycle.</a><br>
154 <a href="#O7">O7. Why doesn't fetchmail deliver mail in date-sorted order?</a><br>
155 <a href="#O8">O8. I'm using pppd. Why isn't my monitor option working?</a><br>
156 <a href="#O9">O9. Why does fetchmail keep retrieving the same messages
157 over and over?</a><br>
161 <h2><a name="G1">G1. What is fetchmail and why should I bother?</a></h2>
163 Fetchmail is a one-stop solution to the remote mail retrieval problem
164 for Unix machines, quite useful to anyone with an intermittent PPP or
165 SLIP connection to a remote mailserver. It can collect mail using any
166 variant of POP or IMAP and forwards via port 25 to the local SMTP
167 listener, enabling all the normal forwarding/filtering/aliasing
168 mechanisms that would apply to local mail or mail arriving via a
169 full-time TCP/IP connection.<p>
171 Fetchmail is not a toy or a coder's learning exercise, but an
172 industrial-strength tool capable of transparently handling every
173 retrieval demand from those of a simple single-user ISP connection up
174 to mail retrieval and rerouting for an entire client domain.
175 Fetchmail is easy to configure, unobtrusive in operation, powerful,
176 feature-rich, and well documented. <P>
178 Fetchmail is <a href="http://www.opensource.org">open-source</a>
179 software. The openness of the sources is the strongest assurance of
180 quality you can have. Extensive peer review by a large,
181 multi-platform user community has shown that fetchmail is as near
182 bulletproof as the underlying protocols permit.<p>
184 Fetchmail is licensed under the <a
185 href="http://gnu.org//copyleft/gpl.html">GNU General Public
188 If you found this FAQ in the distribution, see the README for fetchmail's
189 full feature list.<p>
192 <h2><a name="G2">G2. Where do I find the latest FAQ and fetchmail
195 The latest HTML FAQ is available alongside the latest fetchmail
196 sources at the fetchmail home page:
197 <a href="http://www.tuxedo.org/~esr/fetchmail">
198 http://www.tuxedo.org/~esr/fetchmail</a>. You can also usually find
200 href="http://sunsite.unc.edu/pub/Linux/system/mail/pop/!INDEX.html">POP
201 mail tools directory on Sunsite</a>.<p>
203 A text dump of this FAQ is included in the fetchmail
204 distribution. Because it freezes at distribution release time, it may
205 not be completely current.<p>
208 <h2><a name="G3">G3. I think I've found a bug. Will you fix it?</a></h2>
210 Yes I will, provided you include enough diagnostic information for me
211 to go on. Send bugs to <a
212 href="mailto:fetchmail-friends@ccil.org">fetchmail-friends</a>. When reporting
213 bugs, please include the following:
216 <li>Your operating system.
217 <li>Your compiler version, if you built from source; otherwise, the
218 name and origin ogf the RPM or other binary package you installed.
219 <li>A copy of your POP or IMAP server's greeting line.
220 <li>The name and version of the SMTP listener or MDA you are forwarding to.
221 <li>Any command-line options you used.
222 <li>The output of fetchmail -V called with whatever other
223 command-line options you used.
226 If you have FTP access to your remote mail account, and you have any
227 suspicion that the bug was triggered by a particular message, please
228 include a copy of the message that triggered the bug.<p>
230 Often, the first thing I will do when you report a bug is tell you to
231 upgrade to the newest version of fetchmail, and then see if the
232 problem reproduces. So you'll probably save us both time if you
233 upgrade and test with the latest version <em>before</em> sending in a
236 Another useful thing you can do, if you're using POP3, is to test for
237 IMAP4 support on your mailserver using the autoprobe function of
238 fetchmailconf. If you have IMAP4, and fetchmailconf doesn't tell you
239 it's broken, switch immediately. POP3 is a weak, poorly-designed
240 protocol with chronic problems, and the later versions after RFC1725
241 actually get worse rather than better. Changing over to IMAP4 may well
242 make your problem go away -- and if your ISP doesn't have IMAP4
243 support, bug them to supply it.<p>
245 It is helpful if you include your .fetchmailrc file, but not necessary
246 unless your symptom seems to involve an error in configuration
247 parsing. If you do send in your .fetchmailrc, mask the passwords
250 If fetchmail seems to run and fetch mail, but the headers look mangled
251 (that is, headers are missing or blank lines are inserted in the
252 headers) then read the FAQ items in section <a href="#X1">X</a>
253 before submitting a bug report. Pay special attention to the item on
254 <a href="#generic_mangling">diagnosing mail mangling</a>. There are
255 lots of ways for other programs in the mail chain to screw up that
256 look like fetchmail's fault, but you may be able to fix these by
257 tweaking your configuration.<P>
259 A transcript of the failed session with -v -v (yes, that's
260 <em>two</em> -v options, enabling debug mode) will almost always be useful.
261 It is very important that the transcript include your POP/IMAP server's
262 greeting line, so I can identify it in case of server problems. This
263 transcript will not reveal your passwords, which are specially masked
264 out precisely so the transcript can be passed around.<P>
266 If the bug involves a core dump or hang, a gdb stack trace is good to have.
267 (Bear in mind that you can attach gdb to a running but hung process by
268 giving the process ID as a second argument.) You will need to
272 CFLAGS=-g LDFLAGS=" " ./configure
275 and then rebuild in order to generate a version that can be gdb-traced.<p>
277 Best of all is a mail file which, when fetched, will reproduce the
278 bug under the latest (current) version.<p>
280 Any bug I can reproduce will usually get fixed very quickly, often
281 within 48 hours. Bugs I can't reproduce are a crapshoot. If the
282 solution isn't obvious when I first look, it may evade me for a long
283 time (or to put it another way, fetchmail is well enough tested that the
284 easy bugs have long since been found). So if you want your bug fixed
285 rapidly, it is not just sufficient but nearly <em>necessary</em> that
286 you give me a way to reproduce it.<p>
289 <h2><a name="G4">G4. I have this idea for a neat feature. Will you add it?</a></h2>
291 Probably not. Most of the feature suggestions I get are for ways to
292 set various kinds of administrative policy or add more spam filtering
293 (the most common one, which I used to get about four million times a week
294 and got <em>really</em> tired of, is for tin-like kill files).<p>
296 You can do spam filtering better with procmail or maildrop on the
297 server side and (if you're the server sysadmin) sendmail.cf domain
298 exclusions. You can do other policy things better with the
299 <CODE>mda</CODE> option and script wrappers around fetchmail. If
300 it's a prime-time-vs.-non-prime-time issue, ask yourself whether a
301 wrapper script called from crontab would do the job.<p>
303 I'm not going to do these; fetchmail's job is transport, not policy, and I
304 refuse to change it from doing one thing well to attempting many things badly.
305 One of my objectives is to keep fetchmail simple so it stays reliable.<p>
307 For reasons fetchmail doesn't have other commonly-requested features
308 (such as password encryption, or multiple concurrent polls from the
309 same instance of fetchmail) see the <a
310 href="http://www.tuxedo.org/~esr/fetchmail/design-notes.html">design notes</a>.<p>
312 Fetchmail is a mature project, no longer in constant active
313 development. It is no longer my top project, and I am going to be
314 quite reluctant to add features that might either jeopardize its
315 stability or involve me in large amounts of coding.<p>
317 All that said, if you have a feature idea that really is about a transport
318 problem that can't be handled anywhere but fetchmail, lay it on me. I'm
319 very accommodating about good ideas.<p>
322 <h2><a name="G5">G5. Is there a mailing list for exchanging tips?</a></h2>
324 There is a fetchmail-friends list for people who want to discuss fixes
325 and improvements in fetchmail and help co-develop it. It's a MailMan
326 list, which you can sign up for at <a
327 <a href="http://lists.ccil.org/mailman/listinfo/fetchmail-friends">fetchmail-friends@lists.ccil.org</a>.
328 There is also an announcements-only list, <a href="http://lists.ccil.org/mailman/listinfo/fetchmail-announce">fetchmail-announce@lists.ccil.org</em>.<P>
332 <h2><a name="G6">G6. So, what's this I hear about a fetchmail paper?</a></h2>
334 The fetchmail development was also a sociological experiment, an
335 extended test to see if my theory about the critical features of the
336 Linux development model is correct.<p>
338 The experiment was a success. I wrote a paper about it titled <a
339 href="http://www.tuxedo.org/~esr/writings/cathedral.html">The
340 Cathedral and the Bazaar</a> which was first presented at Linux
341 Kongress '97 in Bavaria and very well received there. It was also
342 given at Atlanta Linux Expo, Linux Pro '97 in Warsaw, and the first
343 Perl Conference, at UniForum '98, and was the basis of an invited
344 presentation at Usenix '98. The folks at Netscape tell me it helped
346 href="http://www.netscape.com/newsref/pr/newsrelease558.html"> give
347 away the source for Netscape Communicator</a>.<p>
349 If you're reading a non-HTML dump of this FAQ, you can find the paper
350 on the Web with a search for that title.<p>
353 <h2><a name="G7">G7. What is the best server to use with fetchmail?</a></h2>
355 The short answer: IMAP 2000 running over Unix.<P>
357 Here's a longer answer: <P>
359 Fetchmail will work with any POP, IMAP, ETRN, or ODMR server that
360 conforms to the relevant RFCs (and even some outright broken ones like
361 <a href="#S2">Microsoft Exchange</a> and <a href="#S12">Novell
362 GroupWise</a>). This doesn't mean it works equally well with all,
363 however. POP2 servers, and POP3 servers without LAST, limit
364 fetchmail's capabilities in various ways described on the manual
367 Most modern Unixes (and effectively all Linux/*BSD systems) come with
368 POP3 support preconfigured (but beware of the horribly broken POP3
369 server mentioned in <a href="#D2">D2</a>). An increasing minority
370 also feature IMAP (you can detect IMAP support by running fetchmail in
371 AUTO mode, or by using the `Probe for supported protocols' function in
372 the fetchmailconf utility).<P>
374 If you have the option, we recommend using or installing an IMAP4rev1
375 server; it has the best facilities for tracking message `seen' states.
376 It also recovers from interrupted connections more gracefully than
377 POP3, and enables some significant performance optimizations. The new
378 <a href="ftp://ftp.cac.washington.edu/imap/imap.tar.Z">IMAP 2000</a>
379 is particularly nice, as it supports CRAM-MD5 so you don't have to
380 ship your mail password over the net en clair (fetchmail autodetects
381 this capability). Older versions had support for GSSAPI giving a
384 Don't be fooled by NT/Exchange propaganda. M$ Exchange is just plain
385 broken (see item <a href="#S2">S2</a>) and NT cannot handle the
386 sustained load of a high-volume remote mail server. Even Microsoft
387 itself knows better than to try this; their own Hotmail service runs
388 over Solaris! For extended discussion, see John Kirch's excellent <a
389 href="http://unix-vs-nt.org/kirch/">white paper</a> on Unix
390 vs. NT performance.<P>
392 Source for a high-quality supported implementation of POP is available
393 from the <a href="ftp://ftp.qualcomm.com/eudora/servers/unix/popper/">Eudora
394 FTP site</a>. Don't use 2.5, which has a rather restrictive license.
395 The 2.5.2 version appears to restore the open-source license of
396 previous versions.<P>
399 <h2><a name="G8">G8. What is the best mail program to use with fetchmail?</a></h2>
401 Fetchmail will work with all popular <a href="#T1">mail transport programs</a>.
402 It also doesn't care which user agent you use, and user agents are as a
403 rule almost equally indifferent to how mail is delivered into your system
404 mailbox. So any of the popular Unix mail agents --
405 <a href="http://www.myxa.com/old/elm.html">elm</a>,
406 <a href="http://www.washington.edu/pine/">pine</a>
407 <a href="http://www.cs.indiana.edu/docproject/mail/mh.html">mh</a>,
408 or <a href="http://www.mutt.org">mutt</a>
409 -- will work fine with fetchmail.<p>
411 All this having been said, I can't resist putting in a discreet plug
412 for <a href="http://www.mutt.org">mutt</a>. My own personal mail
413 setup is sendmail plus fetchmail plus mutt. Mutt's interface is only
414 a little different from that of its now-moribund ancestor elm, but its
415 excellent handling of MIME and PGP put it in a class by itself. You
416 won't need its built-in POP3 support, though; most of the mutt
417 developers will cheerfully admit that fetchmail's is better :-).<p>
420 <h2><a name="G9">G9. How can I avoid sending my password en clair?</a></h2>
422 Depending on what your mail server you are talking to, this ranges
423 from trivial to impossible. It may even be next to useless.<P>
425 Most people use fetchmail over phone wires, which are hard to tap.
426 Anybody with the skill and resources to do this could get into your
427 server mailbox with much less effort by subverting the server host.
428 So if your provider setup is modem wires going straight into a service
429 box, you probably don't need to worry.<P>
431 In general there is little point in trying to secure your fetchmail
432 transaction unless you trust the security of the server host you are
433 retrieving mail from. Your vulnerability is more likely to be an
434 insecure local network on the server end (e.g. to somebody with a TCP/IP
435 packet sniffer intercepting Ethernet traffic between the modem
436 concentrator you dial in to and the mailserver host).<P>
438 Having realized this, you need to ask whether password encryption
439 alone will really address your security exposure. If you think you
440 might be snooped between server and client, it's better to use
441 end-to-end encryption on your whole mail stream so none of it can be
442 read. One of the advantages of fetchmail over conventional SMTP-push
443 delivery is that you may be able to arrange this by using ssh(1); see
444 <a href="#K3">K3</a>.<P>
446 Note that ssh is not a complete privacy solution either, as your mail
447 could have been snooped in transit to your POP server from wherever it
448 originated. For best security, agree with your correspondents to use
449 a tool such as <a href="http://www.gnupg.org/">GPG</a> (Gnu Privacy
450 Guard) or PGP (Pretty Good Privacy).<P>
452 If ssh/sshd isn't available, or you find it too complicated for you to
453 set up, password encryption will at least keep a malicious cracker
454 from deleting your mail, and require him to either tap your connection
455 continuously or crack root on the server in order to read it.<P>
457 You can deduce what encryptions your mail server has available
458 by looking at the server greeting line (and, for IMAP, the
459 response to a CAPABILITY query). Do a <code>fetchmail -v</code>
460 to see these, or telnet direct to the server port (110 for POP3, 143 for
463 If your mailserver is using IMAP 2000, you'll have CRAM-MD5 support
464 built in. Fetchmail autodetects this; you can skip the rest of this
467 The POP3 facility you are most likely to have available is APOP. This is a
468 POP3 feature supported by many servers (fetchmailconf's autoprobe
469 facility will detect it and tell you if you have it). If you see
470 something in the greeting line that looks like an
471 angle-bracket-enclosed Internet address with a numeric left-hand part,
472 that's an APOP challenge (it will vary each time you log in). You can
473 register a secret on the host (using <code>popauth(8)</code> or some
474 program like it). Specify the secret as your password in your
475 .fetchmailrc; it will be used to encrypt the current challenge, and
476 the encrypted form will be sent back the the server for
479 Alternatively, you may have Kerberos available. This may require you
480 to set up some magic files in your home directory on your client
481 machine, but means you can omit specifying any password at all.<P>
483 Fetchmail supports two different Kerberos schemes. One is a POP3
484 variant called KPOP; consult the documentation of your mail server to
485 see if you have it (one clue is the string "krb-IV" in the greeting
486 line on port 110). The other is an IMAP and POP3 facility described
487 by RFC1731 and RFC1734. You can tell if this one is present by looking
488 for AUTH=KERBEROS_V4 in the CAPABILITY response.<P>
490 If you are fetching mail from a CompuServe POP3 account, you can use
491 their RPA authentication (which works much like APOP). See <a
492 href="#S3">S3</a> for details. If you are fetching mail from
493 Microsoft Exchange, you will be able to use NTLM.<P>
495 Your POP3 server may have the RFC1938 OTP capability to use one-time
496 passwords (if it doesn't, you can get OTP patches for the 2.2 version
497 of the Qualcomm popper from <a href="#cmetz">Craig Metz</a>). To check
498 this, look for the string "otp-" in the greeting line. If you see it,
499 and your fetchmail was built with OPIE support compiled in (see the
500 distribution INSTALL file), fetchmail will detect it also. When using
501 OTP, you will specify a password but it will not be sent en clair.<P>
503 You can get both POP3 and IMAP OTP patches from <a name="cmetz">Craig
505 href="http://www.inner.net/pub/">http://www.inner.net/pub/</a>.<P>
506 These patches use a SASL authentication method named "X-OTP" because
507 there is not currently a standard way to do this; fetchmail also uses
508 this method, so the two will interoperate happily. They better,
509 because this is how Craig gets his mail ;-)<P>
511 Finally, you can use <a href="#K5">SSL</a> for complete
512 end-to-end encryption if you have an SSL-enabled mailserver.<P>
514 <h2><a name="G10">G10. Is any special configuration needed to use a dynamic IP address?</a></h2>
516 Yes. In order to avoid giving indigestion to certain picky MTAs
517 (notably <a href="#T3">exim</a>), fetchmail always makes the RCPT TO
518 address it feeds the MTA a fully qualified one with a hostname part.
519 Normally it does this by appending @ and "localhost", but when you are
520 using Kerberos or ETRN mode it will append @ and your machine's
521 fully-qualified domain name (FQDN).<P>
523 Appending the FQDN can create problems when fetchmail is running in daemon
524 mode and outlasts the dynamic IP address assignment your client
525 machine had when it started up.<P>
527 Since the new IP address (looked up at RCPT TO interpretation time)
528 doesn't match the original, the most benign possible result is that
529 your MTA thinks it's seeing a relaying attempt and refuses. More
530 frequently, fetchmail will try to connect to a nonexistent host
531 address and time out. Worst case, you could up forwarding your mail
532 to the wrong machine!<P>
534 Use the <code>smtpaddress</code> option to force the appended hostname
535 to one with a (fixed) IP address of 127.0.0.1 in your
536 <code>/etc/hosts</code>. (The name `localhost' will usually work; or
537 you can use the IP address itself).<P>
539 Only one fetchmail option interacts directly with your IP address,
540 `<code>interface</code>'. This option can be used to set the gateway
541 device and restrict the IP address range fetchmail will use. Such a
542 restriction is sometimes useful for security reasons, especially on
543 multihomed sites. See <a href="#C3">C3</a>.<P>
545 I recommend against trying to set up the <code>interface</code> option
546 when initially developing your poll configuration -- it's never
547 necessary to do this just to get a link working. Get the link working
548 first, observe the actual address range you see on connections, and
549 add an <code>interface</code> option (if you need one) later.<P>
551 You can't use ETRN if you have a dynamic IP address (your ISP changes
552 your IP address occasionally, possibly with every connect). You need
553 to have your own registered domain and a definite IP address
554 registered for that domain. The server needs to be configured to
555 accept mail for your domain but then queue it to forward to your
556 machine. ETRN just tells to server to flush its queue for your
557 domain. Fetchmail doesn't actually get the mail in that case.<p>
559 You can use On-Demand Mail Relay (ODMR) with a dynamic IP address;
560 that's what it was designed for, and it provides capabilities very
561 similar to ETRN. Unfortunately ODMR servers are not yet widely
562 deployed, as of early 2001.<p>
564 If you're using a dynamic-IP configuration, one other (non-fetchmail)
565 problem you may run into with outgoing mail is that some sites will
566 bounce your email because the hostname your giving them isn't real
567 (and doesn't match what they get doing a reverse DNS on your
568 dynamically-assigned IP address). If this happens, you need to hack
569 your sendmail so it masquerades as your host. Setting<P>
575 in your <code>sendmail.cf</code> will work, or you can set<P>
578 MASQUERADE_AS(smarthost.here)
581 in the m4 configuration and do a reconfigure. (In both cases, replace
582 <code>smarthost.here</code> with the actual name of your mailhost.)
583 See the <a href="http://www.lege.com/sendmail-FAQ.txt">sendmail
584 FAQ</a> for more details.<P>
587 <h2><a name="G11">G11. Is any special configuration needed to use firewalls?</a></h2>
589 No. You can use fetchmail with SOCKS, the standard tool for
590 indirecting TCP/IP through a firewall. You can find out about SOCKS,
591 and download the SOCKS software including server and client code, at
592 the <a href="http://www.socks.nec.com/">SOCKS distribution
595 The specific recipe for using fetchmail with a firewall is at <a
599 <h2><a name="G12">G12. Is any special configuration needed to <em>send</em> mail?</a></h2>
601 A user asks: but how do we send mail out to the POP3 server? Do I need
602 to implement another tool or will fetchmail do this too?<p>
604 Fetchmail only handles the receiving side. The sendmail or other
605 preinstalled MTA on your client machine will handle sending mail
606 automatically; it will ship mail that is submitted while the
607 connection is active, and put mail that is submitted while
608 the connection is inactive into the outgoing queue.<P>
610 Normally, sendmail is also run periodically (every 15 minutes on most
611 Linux systems) in a mode that tries to ship all the mail in the
612 outgoing queue. If you have set up something like pppd to
613 automatically dial out when your kernel is called to open a TCP/IP
614 connection, this will ensure that the mail gets out.<P>
617 <h2><a name="G13">G13. Is fetchmail Y2K-compliant?</a></h2>
619 Fetchmail is fully Y2K-compliant.<P>
621 Fetchmail could theoretically have problems when the 32-bit time_t
622 counters roll over in 2038, but I doubt it. Timestamps aren't used
623 for anything but log entry generation. Anyway, if you aren't running
624 on a 64-bit machine by then, you'll deserve to lose.<P>
627 <h2><a name="G14">G14. Is there a way in fetchmail to support disconnected IMAP mode?</a></H2>
629 No. Fetchmail is a mail transport agent, best understood as a protocol
630 gateway between POP3/IMAP servers and SMTP. Disconnected operation
631 requires an elaborate interactive client. It's a very different problem.<p>
634 <h2><a name="G15">G15. How will fetchmail perform under heavy loads?</a></h2>
636 Fetchmail streams message bodies line-by-line; the most core it
637 ever requires per message is enough memory to hold the RFC822 header, and
638 that storage is freed when body processing begins. It is, accordingly,
639 quite economical in its use of memory.<p>
641 After startup time, a fetchmail running in daemon mode stats its
642 configuration file once per poll cycle to see whether it has changed
643 and should be rescanned. Other than that, a fetchmail in normal
644 operation doesn't touch the disk at all; that job is left up to the
645 MTA or MDA the fetchmail talks to.<p>
647 Fetchmail's performance is usually bottlenecked by latency on the POP
648 server or (less often) on the TCP/IP link to the server. This is not
649 a problem readily solved by tuning fetchmail, or even by buying more
650 TCP/IP capacity (which tends to improve bandwidth but not necessarily
654 <h2><a name="B1">B1. Lex bombs out while building the fetchmail lexer.</a></h2>
656 In the immortal words of Alan Cox the last time this came up: ``Take
657 the Solaris lex and stick it up the backside of a passing Sun
658 salesman, then install <a
659 href="ftp://ftp.gnu.org/pub/non-gnu/flex/">flex</a> and use that. All
660 will be happier.''<P>
662 I couldn't have put it better myself, and ain't going to try now.<P>
664 (The same problem has been reported under HP-UX v10.20 and IRIX)<P>
667 <h2><a name="B2">B2. I get link failures when I try to build fetchmail.</a></h2>
669 If you get errors resembling these<P>
672 mxget.o(.text+0x35): undefined referenceto `__res_search'
673 mxget.o(.text+0x99): undefined reference to`__dn_skipname'
674 mxget.o(.text+0x11c): undefined reference to`__dn_expand'
675 mxget.o(.text+0x187): undefined reference to`__dn_expand'
676 make: *** [fetchmail] Error 1
679 then you must add "-lresolv" to the LOADLIBS line in your Makefile
680 once you have installed the `bind' package.<P>
683 <h2><a name="F1">F1. Why does my old .fetchmailrc file no longer work?</a></h2>
685 <h3>If your file predates 5.6.5</h3>
687 The <tt>imap-gss</tt>, <tt>imap-k4</tt>, and <tt>imap-login</tt>
688 protocol types are gone. This is a result of a major re-factoring
689 of the authentication machinery; fetchmail can now use Kerberos V4
690 and GSSAPI not just with IMAP but with POP3 servers that have RFC1734
691 support for the AUTH command.<p>
693 When trying to identify you to an IMAP or POP mailserver, fetchmail
694 now first tries methods that don't require a password (GSSAPI,
695 KERBEROS_IV); then it looks for methods that mask your password
696 (CRAM-MD5, X-OTP); and only if it the server doesn't support any of
697 those will it ship your password en clair.<p>
699 Setting the <tt>preauth</tt> option to any value other than `password'
700 will prevent from looking for a password in your <tt>.netrc</tt> file
701 or querying for it at startup time.<p>
703 <h3>If your file predates 5.1.0</h3>
705 In 5.1.0, the <tt>auth</tt> keyword and option were changed to
708 <h3>If your file predates 4.5.5</h3>
710 If the <code>dns</code> option is on (the default), you may need to
711 make sure that any hostname you specify (for mail hosts or for an SMTP
712 target) is a canonical fully-qualified hostname). In order to avoid
713 DNS overhead and complications, fetchmail no longer tries to derive
714 the fetchmail client machine's canonical DNS name at startup.<P>
716 <h3>If your file predates 4.0.6:</h3>
718 Just after the `<CODE>via</CODE>' option was introduced, I realized
719 that the interactions between the `<CODE>via</CODE>',
720 `<CODE>aka</CODE>', and `<CODE>localdomains</CODE>' options were out
721 of control. Their behavior had become complex and confusing, so much so
722 that I was no longer sure I understood it myself. Users were being
723 unpleasantly surprised.<P>
725 Rather than add more options or crock the code, I re-thought it. The
726 redesign simplified the code and made the options more orthogonal, but
727 may have broken some complex multidrop configurations.
729 Any multidrop configurations that depended on the name just after the
730 `<CODE>poll</CODE>' or `<CODE>skip</CODE>' keyword being still
731 interpreted as a DNS name for address-matching purposes, even in the
732 presence of a `<CODE>via</CODE>' option, will break.<P>
734 It is theoretically possible that other unusual configurations (such
735 as those using a non-FQDN poll name to generate Kerberos IV tickets) might
736 also break; the old behavior was sufficiently murky that we can't be
737 sure. If you think this has happened to you, contact the maintainer.<P>
739 <h3>If your file predates 3.9.5:</h3>
741 The `<code>remote</code>' keyword has been changed to `<code>folder</code>'.
742 If you try to use the old keyword, the parser will utter a warning.<P>
744 <h3>If your file predates 3.9:</h3>
746 It could be because you're using a .fetchmailrc that's written in the
747 old popclient syntax without an explicit `<CODE>username</CODE>'
748 keyword leading the first user entry attached to a server entry.
750 This error can be triggered by having a user option such as `<CODE>keep</CODE>'
751 or `<CODE>fetchall</CODE>' before the first explicit username. For
752 example, if you write<p>
755 poll openmail protocol pop3
756 keep user "Hal DeVore" there is hdevore here
759 the `<CODE>keep</CODE>' option will generate an entire user entry with
760 the default username (the name of fetchmail's invoking user).<p>
762 The popclient compatibility syntax was removed in 4.0. It complicated
763 the configuration file grammar and confused users.<p>
765 <h3>If your file predates 2.8:</h3>
767 The `<CODE>interface</CODE>', `<CODE>monitor</CODE>' and
768 `<CODE>batchlimit</CODE>' options changed after 2.8.<p>
770 They used to be global options with `<CODE>set</CODE>' syntax like the
771 batchlimit and logfile options. Now they're per-server options, like
772 `<CODE>protocol</CODE>'.<p>
774 If you had something like<p>
777 set interface = "sl0/10.0.2.15"
780 in your .fetchmailrc file, simply delete that line and insert
781 `interface sl0/10.0.2.15' in the server options part of your `defaults'
784 Do similarly for any `<CODE>monitor</CODE>' or `<CODE>batchlimit</CODE>' options.<p>
787 <h2><a name="F2">F2. The .fetchmailrc parser won't accept my all-numeric user name.</a></h2>
789 Either upgrade to a post-5.0.5 fetchmail or put string quotes around it. :-)<p>
791 The configuration file parser in older fetchmail versions treated any
792 all-numeric token as a number, which confused it when it was
793 expecting a name. String quoting forces the token's class.<p>
795 The lexical analyzer in 5.0.6 and beyond is smarter and assumes
796 any token following "username" or "password" is a string.
799 <h2><a name="F3">F3. The .fetchmailrc parser won't accept my host or username beginning with `no'.</a></h2>
801 See <a href="#F2">F2</a> You're caught in an unfortunate crack between
802 the newer-style syntax for negated options (`no keep', `no rewrite'
803 etc.) and the older style run-on syntax (`nokeep', `norewrite'
806 Upgrade to a 5.0.6 or later fetchmail, or put string quotes around your
810 <h2><a name="F4">F4. I'm getting a `parse error' message I don't understand.</a></h2>
812 The most common cause of mysterious parse errors is putting a server
813 option after a user option. Check the manual page; you'll probably
814 find that by moving one or more options closer to the `poll' keyword
815 you can eliminate the problem.<p>
817 Yes, I know these ordering restrictions are hard to understand.
818 Unfortunately, they're necessary in order to allow the `defaults'
822 <h2><a name="C1">C1. Why do I need a .fetchmailrc when running as root on my own machine?</a></h2>
824 Ian T. Zimmerman <itz@rahul.net> asked:<p>
826 On the machine where I'm the only real user, I run fetchmail as root
827 from a cron job, like this:<p>
830 fetchmail -u "itz" -p POP3 -s bolero.rahul.net
833 This used to work as is (with no .fetchmailrc file in root's home
834 directory) with the last version I had (1.7 or 1.8, I don't
835 remember). But with 2.0, it RECPs all mail to the local root user,
836 unless I create a .fetchmailrc in root's home directory containing:<p>
839 skip bolero.rahul.net proto POP3
843 It won't work if the second line is just "<CODE>user itz</CODE>". This is silly.<p>
845 It seems fetchmail decides to RECP the `default local user' (i.e. the
846 uid running fetchmail) unless there are local aliases, and the
847 `default' aliases (itz->itz) don't count. They should.<p>
851 No they shouldn't. I thought about this for a while, and I don't much
852 like the conclusion I reached, but it's unavoidable. The problem is
853 that fetchmail has no way to know, in general, that a local user `itz'
856 "Ah!" you say, "Why doesn't it check the password file to see if the remote
857 name matches a local one?" Well, there are two reasons.<p>
859 One: it's not always possible. Suppose you have an SMTP host declared
860 that's not the machine fetchmail is running on? You lose.<p>
862 Two: How do you know server itz and SMTP-host itz are the same person?
863 They might not be, and fetchmail shouldn't assume they are unless
864 local-itz can explicitly produce credentials to prove it (that is, the
865 server-itz password in local-itz's .fetchmailrc file.).<p>
867 Once you start running down possible failure modes and thinking about
868 ways to tinker with the mapping rules, you'll quickly find that all the
869 alternatives to the present default are worse or unacceptably
870 more complicated or both.<p>
873 <h2><a name="C2">C2. How can I arrange for a fetchmail daemon to get killed when I log out?</a></h2>
875 The easiest way to dispatch fetchmail on logout (which will work
876 reliably only if you have just one login going at any time) is to
877 arrange for the command `fetchmail -q' to be called on logout. Under
878 bash, you can arrange this by putting `fetchmail -q' in the file
879 `~/.bash_logout'. Most csh variants execute `~/.logout' on logout.
880 For other shells, consult your shell manual page.<p>
882 Automatic startup/shutdown of fetchmail is a little harder to arrange
883 if you may have multiple login sessions going. In the contrib
884 subdirectory of the fetchmail distribution there is some shell code
885 you can add to your .bash_login and .bash_logout profiles that will
886 accomplish this. Thank James Laferriere <babydr@nwrain.net> for
889 Some people start up and shut down fetchmail using the ppp-up and
890 ppp-down scripts of pppd.<p>
893 <h2><a name="C3">C3. How do I know what interface and address to use with --interface?</a></h2>
895 This depends a lot on your local networking configuration (and right
896 now you can't use it at all except under Linux and the newer BSDs). However,
897 here are some important rules of thumb that can help. If they don't
898 work, ask your local sysop or your Internet provider.<p>
900 First, you may not need to use --interface at all. If your machine
901 only ever does SLIP or PPP to one provider, it's almost certainly by a
902 point to point modem connection to your provider's local subnet that's
903 pretty secure against snooping (unless someone can tap your phone or
904 the provider's local subnet!). Under these circumstances, specifying
905 an interface address is fairly pointless.<p>
907 What the option is really for is sites that use more than one
908 provider. Under these circumstances, typically one of your provider
909 IP addresses is your mailserver (reachable fairly securely via the
910 modem and provider's subnet) but the others might ship your packets
911 (including your password) over unknown portions of the general
912 Internet that could be vulnerable to snooping. What you'll use
913 --interface for is to make sure your password only goes over the
916 To determine the device:<p>
919 <li> If you're using a SLIP link, the correct device is probably sl0.
920 <li> If you're using a PPP link, the correct device is probably ppp0.
921 <li> If you're using a direct connection over a local network such as
922 an ethernet, use the command `netstat -r' to look at your routing table.
923 Try to match your mailserver name to a destination entry; if you don't
924 see it in the first column, use the `default' entry. The device name
925 will be in the rightmost column.
928 To determine the address and netmask:<p>
931 <li> If you're talking to slirp, the correct address is probably 10.0.2.15,
932 with no netmask specified. (It's possible to configure slirp to present
933 other addresses, but that's the default.)
935 <li> If you have a static IP address, run `ifconfig <device>', where <device>
936 is whichever one you've determined. Use the IP address given after
937 "inet addr:". That is the IP address for your end of the link, and is
938 what you need. You won't need to specify a netmask.
940 <li> If you have a dynamic IP address, your connection IP will vary randomly
941 over some given range (that is, some number of the least significant bits
942 change from connection to connection). You need to declare an address
943 with the variable bits zero and a complementary netmask that sets
947 To illustrate the rule for dynamic IP addresses, let's suppose you're
948 hooked up via SLIP and your IP provider tells you that the dynamic
949 address pool is 255 addresses ranging from 205.164.136.1 to
950 205.164.136.255. Then<p>
953 interface "sl0/205.164.136.0/255.255.255.0"
956 would work. To range over any value of the last two octets
957 (65536 addresses) you would use<p>
960 interface "sl0/205.164.0.0/255.255.0.0"
964 <h2><a name="C4">C4. How can I set up support for sendmail's anti-spam features?</a></h2>
966 This answer covers versions of sendmail from 8.8.7 (the version
967 installed in Red Hat 5.1) upwards. If you have an older version,
968 upgrade to sendmail 8.9.<P>
970 Stock sendmails can now do anti-spam exclusions based on a database of
971 filter rules. The human-readable form of the database is at
972 <tt>/etc/mail/deny</tt>. The database itself is at
973 <tt>/etc/mail/deny.db</tt>.<P>
975 The table itself uses email addresses, domain names, and network
976 numbers as keys. For example,</P>
978 spammer@aol.com REJECT
979 cyberspammer.com REJECT
982 <P>would refuse mail from spammer@aol.com, any user from
983 cyberspammer.com (or any host within the cyberspammer.com domain), and
984 any host on the 192.168.212.* network. (This feature can be used to
985 do other things as well; see the <a
986 href="http://www.sendmail.org/m4/anti-spam.html">sendmail
987 documentation</a> for details)</P>
989 To actually set up the database, run
992 makemap hash deny <deny
996 To test, send a message to your mailing address from that host and
997 then pop off the message with fetchmail, using the -v argument. You
998 can monitor the SMTP transaction, and when the FROM address is parsed,
999 if sendmail sees that it is an address in spamlist, fetchmail will
1000 flush and delete it.<p>
1002 Under no circumstances put your <strong>mailhost</strong> or <strong>any host
1003 you accept mail from</strong> using fetchmail into your reject file. You
1004 <strong>will</strong> lose mail if you do this!!!<p>
1007 <h2><a name="C5">C5. How can I poll some of my mailboxes more/less
1008 often than others?</a></h2>
1010 Use the <cite>interval</cite> keyword on the ones that should be
1011 checked less often. For example, if you do a poll every 5 minutes,
1012 and want to poll some mailboxes every 5 minutes and some every 30
1013 minutes, use something like this:<p>
1016 poll mainsite.example.com proto pop3 user ....
1017 poll secondary.example.com proto pop3 interval 6 user ...
1020 Then secondary.example.com will be polled every 6th time that
1021 mainsite.example.com is polled, which with a polling interval of every
1022 5 minutes means that secondary.example.com will be polled every 30
1026 <h2><a name="C6">Fetchmail works OK started up manually, but not from an init script.</a></h2>
1028 Often, startup scripts have a different environment than an interactive
1029 login shell. For instance, $HOME might point to "/root" when you are
1030 logged in as root, but it might be either unset, or set to "/" when the
1031 startup scripts are running. That means fetchmail at startup can't find
1032 the .fetchmailrc.<p>
1034 Pick a location (such as /etc/fetchmailrc) and use fetchmail's -f
1035 option to point fetchmail at it. That should solve the problem.<p>
1038 <h2><a name="C7">C7. How can I forward mail to another host?</a></h2>
1040 To forward mail to a host other than the one you are running fetchmail
1041 on, use the <code>smtphost</code> or <code>smtpname</code> option.
1042 See the manual page for details.<p>
1045 <h2><a name="T1">T1. How can I use fetchmail with sendmail?</a></h2>
1047 For most sendmails, no special configuration is required. Eric Allman
1048 tells me that if <code>FEATURE(always_add_domain)</code> is included
1049 in sendmail's configuration, you can leave the <code>rewrite</code>
1052 If your sendmail complains ``sendmail does not relay'', make sure
1053 your sendmail.cf file says <code>Cwlocalhost</code>
1054 so that sendmail recognizes `localhost' as a name of its host.<p>
1056 If you're mailing from another machine on your local network, also
1057 ensure that its IP address is listed in ip_allow or name in name_allow
1058 (usually in /etc/mail/)<p>
1060 If you find that your sendmail doesn't like the address
1061 `FETCHMAIL-DAEMON@localhost' (which is used in the bouncemail
1062 that fetchmail generates), you may have to set
1063 <code>FEATURE(accept_unqualified_senders)</code>.<P>
1065 Günther Leber reports that Digital Unix sendmails won't work with
1066 fetchmail. The symptom is an error message "<code>553 Local configuration
1067 error, hostname not recognized as local</code>". The problem is that
1068 fetchmail normally feeds sendmail with the client machine's host
1069 address in the MAIL FROM line. These sendmails think this means
1070 they're seeing the result of a mail loop and suppress the mail. You
1071 may be able to work around this by running in <code>--invisible</code> mode.<P>
1073 If you want to support multidrop mode, and you can get access to your
1074 mailserver's sendmail.cf file, it's a good idea to add this rule:<P>
1077 H?l?Delivered-To: $h
1080 This will cause the mailserver's sendmail to reliably write the
1081 appropriate envelope address into each message before fetchmail sees
1082 it, and tell fetchmail which header it is. With this change,
1083 multidrop mode should work reliably even when the Received header
1084 omits the envelope address (which will typically be the case when
1085 the message has multiple recipients). However it will still not
1086 distinguish the recipients, your only advantage is that no bounce
1087 will be sent if a message is BCC addressed to multiple users at
1088 your site. To fix even that problem, you might want to try the
1089 following hack, which is however untested and quite experimental:<P>
1092 H?J?Delivered-To: $u
1094 Mmdrop, P=/usr/bin/procmail, F=lsDFMqSPfhnu9J,
1095 S=EnvFromSMTP/HdrFromSMTP, R=EnvToSMTP/HdrToSMTP,
1096 T=DNS/RFC822/X-Unix,
1097 A=procmail -Y -a $u -d $h
1100 For both hacks, you have to declare `<CODE>envelope "Delivered-To:"</CODE>' on
1101 the fetchmail side, to put the virtual domain (e.g. `domain.com')
1102 with RELAY permission into your access file and to add a line
1103 reading `<CODE>domain.com local:local-pop-user</CODE>' for the first and
1104 `<CODE>domain.com mdrop:local-pop-user</CODE>' for the second hack to your
1107 You will notice that if the mail already has a Delivered-To header,
1108 sendmail will not add another. Further, editing sendmail.cf
1109 directly is not very comfortable. Solutions for both problems
1110 can be found in Peter `Rattacresh' Backes' `hybrid' patch against
1111 sendmail. Have a look at it, you can find it in the contrib
1114 Feel free to try Martijn Lievaart's detailed recipe in the contrib
1115 subdirectory of the fetchmail source distribution, it attempts
1116 to realize multidrop mailboxes with an external script.<P>
1119 <h2><a name="T2">T2. How can I use fetchmail with qmail?</a></h2>
1121 Turn on the <CODE>forcecr</CODE> option; qmail's listener mode doesn't like
1122 header or message lines terminated with bare linefeeds.<p>
1124 (This information is thanks to Robert de Bath
1125 <robert@mayday.cix.co.uk>.)<p>
1127 If a mailhost is using the qmail package (see <a
1128 href="http://pobox.com/~djb/qmail.html">http://pobox.com/~djb/qmail.html</a>)
1129 then, providing the local hosts are also using qmail, it is possible
1130 to set up one fetchmail link to be reliably collect the mail for an
1133 One of the basic features of qmail is the `Delivered-To:' message
1134 header. Whenever qmail delivers a message to a local mailbox it puts
1135 the username and hostname of the envelope recipient on this line. The
1136 major reason for this is to prevent mail loops. <p>
1138 To set up qmail to batch mail for a disconnected site the ISP-mailhost
1139 will have normally put that site in its `virtualhosts' control file so
1140 it will add a prefix to all mail addresses for this site. This results
1141 in mail sent to 'username@userhost.userdom.dom.com' having a
1142 'Delivered-To:' line of the form:<p>
1145 Delivered-To: mbox-userstr-username@userhost.userdom.dom.com
1148 A single host maildrop will be slightly simpler:
1151 Delivered-To: mbox-userstr-username@userhost.dom.com
1154 The ISP can make the 'mbox-userstr-' prefix anything they choose
1155 but a string matching the user host name is likely.<p>
1157 To use this line you must:<p>
1160 <li>Ensure the option `envelope Delivered-To:' is in the fetchmail
1163 <li>Ensure you have a localdomains containing 'userdom.dom.com' or
1164 `userhost.dom.com' respectively.
1167 So far this reliably delivers messages to the correct machine of the
1168 local network, to deliver to the correct user the 'mbox-userstr-'
1169 prefix must be stripped off of the user name. This can be done by
1170 setting up an alias within the qmail MTA on each local machine.
1171 Simply create a dot-qmail file called '.qmail-mbox-userstr-default'
1172 in the alias directory (normally /var/qmail/alias) with the contents:<p>
1175 | ../bin/qmail-inject -a -f"$SENDER" "${LOCAL#mbox-userstr-}@$HOST"
1178 Note this <em>does</em> require a modern /bin/sh.<p>
1180 Peter Wilson adds: <P>
1182 ``My ISP uses "alias-unzzippedcom-" as the prefix, which means that I
1183 need to name my file ".qmail-unzzippedcom-default". This is due to
1184 qmail's assumption that a message sent to user-xyz is handled by the
1185 file ~user/.qmail-xyz (or ~user/.qmail-default).''<p>
1187 Luca Olivetti adds:<P>
1189 If you aren't using qmail locally, or you don't want to set up the
1190 alias mechanism described above, you can use the option `<code>qvirtual
1191 "mbox-userstr-"</code>' in your fetchmail config file to strip the prefix
1192 from the local user name.<p>
1195 <h2><a name="T3">T3. How can I use fetchmail with exim?</a></h2><p>
1197 If you have <CODE>rewrite</CODE> on: <P>
1199 There is an RFC1123 requirement that MAIL FROM and RCPT TO addresses
1200 you pass to it have to be canonical (e.g. with a fully qualified
1201 hostname part). Therefore fetchmail tries to pass fully qualified
1202 RCPT TO addresses. But exim does not by default accept `localhost' as
1203 a fully qualified domain. This can be fixed.<P>
1205 In exim.conf, add `localhost' to your local_domains declaration if it's not
1206 already present. For example, the author's site at thyrsus.com would
1207 have a line reading:<P>
1210 local_domains = thyrsus.com:localhost
1213 If you have <CODE>rewrite</CODE> off:<P>
1215 MAIL FROM is a potential problem if the MTAs upstream from your fetchmail
1216 don't necessarily pass canonicalized From and Return-Path addresses,
1217 and fetchmail's <CODE>rewrite</CODE> option is off. The specific case
1218 where this has come up involves bounce messages generated by sendmail
1219 on your mailer host, which have the (un-canonicalized) origin address
1222 The right way to fix this is to enable the <CODE>rewrite</CODE> option and
1223 have fetchmail canonicalize From and Return-Path addresses with the
1224 mailserver hostname before exim sees them. This option is enabled by
1225 default, so it won't be off unless you turned it off.<p>
1227 If you must run with <CODE>rewrite</CODE> off, there is a switch in exim's
1228 configuration files that allows it to accept domainless MAIL FROM
1229 addresses; you will have to flip it by putting the line <p>
1232 sender_unqualified_hosts = localhost
1235 in the main section of the exim configuration file. Note that this
1236 will result in such messages having an incorrect domain name attached
1237 to their return address (your SMTP listener's hostname rather than
1238 that of the remote mail server). <p>
1241 <h2><a name="T4">T4. How can I use fetchmail with smail?</a></h2><p>
1243 Smail 3.2 is very nearly plug-compatible with sendmail, and may work
1244 fine out of the box.<P>
1246 We have one report that when processing multiple messages from a
1247 single fetchmail session, smail sometimes delivers them in an
1248 order other than received-date order. This can be annoying because it
1249 scrambles conversational threads. This is not fetchmail's problem,
1250 it is an smail `feature' and has been reported to the maintainers
1253 Very recent smail versions require an <code>-smtp_hello_verify</code>
1254 option in the smail config file. This overrides smail's check to see
1255 that the HELO address is actually that of the client machine, which
1256 is never going to be the case when fetchmail is in the picture.
1257 According to RFC1123 an SMTP listener <em>must</em> allow this
1258 mismatch, so smail's new behavior (introduced sometime between
1259 3.2.0.90 and 3.2.0.95) is a bug.<P>
1262 <h2><a name="T5">T5. How can I use fetchmail with SCO's MMDF?</a></h2><p>
1264 MMDF itself is difficult to configure, but it turns out that
1265 connecting fetchmail to MMDF's SMTP channel isn't that hard.
1267 href="http://www.aplawrence.com/Unixart/uucptofetch.html">
1268 MMDF recipe</a> that describes replacing a UUCP link with
1269 fetchmail feeding MMDF.<P>
1272 <h2><a name="T6">T6. How can I use fetchmail with Lotus Notes?</a></h2><p>
1274 The Lotus Notes SMTP gateway tries to deduce when it should convert \n
1275 to \r\n, but its rules are not the intuitive and correct-for-RFC822
1276 ones. Use `forcecr'.<P>
1279 <h2><a name="S1">S1. How can I use fetchmail with qpopper?</a></h2>
1281 Qualcomm's qpopper is probably the best-of-breed among POP3 servers, and
1282 is very widely deployed. Nevertheless, it has some problems which
1283 fetchmail exposes. We recommend using <a href="#G7">IMAP</a> instead if at
1284 all possible. If you must talk to qpopper, here are some problems to
1287 <h3>Problems with retrieving large messages from qpopper 2.53</h3>
1289 Tony Tang <a href="mailto:tony@atn.com.hk"><tony@atn.com.hk></a>
1290 reports that there is a bad intercation between fetchmail and qpopper
1291 2.5.3 under Red Hat Linux versions 5.0 to 5.2, kernels 2.0.34 to
1292 2.0.35. When fetching very large messages (over 700K) from 2.5.3,
1293 fetchmail will hang with a socket error.<p>
1295 This is probably not a fetchmail bug, but rather a symptom of some
1296 problem in the networking stack that qpopper's transmission pattern is
1297 tickling, as fetchpop (another Linux POP client) also displays the hang
1298 but Netscape running under Win95 does not. The problem can also be
1300 href="http://www.eudora.com/freeware/qpop.html">upgrading to qpopper
1303 <h3>Bad interaction with fetchmail 4.4.2 to 4.4.7</h3>
1305 Versions of fetchmail from 4.4.2 through 4.4.7 had a bad interaction
1306 with Eudora qpopper versions 2.3 and later. See <a href="#X5">X5</a>
1307 for details. The solution is to upgrade your fetchmail.<p>
1310 <h2><a name="S2">S2. How can I use fetchmail with Microsoft Exchange?</a></h2>
1312 Fetchmail now supports the proprietary NTLM mode used with M$ Exchange
1313 servers. To enable this, configure fetchmail with the --enable-NTLM
1314 option and recompile it. Note: if you specify a user option value
1315 that looks like `user@domain', the part to the left of the @ will
1316 be passed as the username and the part to the right as the NTLM domain.<P>
1318 M$ Exchange violates the POP3 RFCs. Its LIST command does not reveal
1319 the real sizes of mail in the pop mailbox, but the sizes of the
1320 compressed versions in the exchange mail database (thanks to Arjan De
1321 Vet and Guido Van Rooij for alerting us to this problem).<P>
1323 Fetchmail works with M$ Exchange, despite this brain damage. Two
1324 features are compromised. One is that the --limit option will not
1325 work right (it will check against compressed and not actual sizes).
1326 The other is that a too-small SIZE argument may be passed to your
1327 ESMTP listener, assuming you're using one (this should not be a
1328 problem unless the actual size of the message is above the listener's
1329 configured length limit).<P>
1331 Somewhat belatedly, I've learned that there's supposed to be a
1332 registry bit that can fix this breakage:<P>
1335 HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\MsExchangeIs\Parameters
1336 System\Pop3 Compatibility
1339 This is a bitmask that controls the variations from the standard protocol.
1340 The bits defined are:<P>
1344 <DD>Report exact message sizes for the LIST command
1346 <DD>Allow arbitrary linear whitespace between commands and arguments
1348 <DD>Enable the LAST command
1350 <DD>Allow an empty PASS command (needed for users with blank
1351 passwords, but illegal in the protocol)
1353 <DD>Relax the length restrictions for arguments to commands (protocol
1354 requires 40, but some user names may be longer than that).
1356 <DD>Allow spaces in the argument to the USER command.
1359 There's another one that may be useful to know about:<P>
1362 KEY_LOCAL_MACHINE\System\CurrentControlSet\Services\MsExchangeIs\Parameters
1363 System\Pop3 Performance
1368 <DD>Render messages to a temporary stream instead of sending directly
1369 from the database (should always be on)
1371 Flag unrenderable messages (instead of just failing commands)
1372 (should only be on if you are seeing the problems reported
1375 <DD>Return from the QUIT command before all messages have been deleted.
1378 The Microsoft pod-person who revealed this information to me admitted
1379 that he couldn't find it anywhere in their public knowledge base.<P>
1381 Another specific problem we have seen with Exchange servers has as its
1382 symptom a response to LOGIN that says "NO Ambiguous Alias". Grant
1385 This means that Exchange Server is too f*&#ing stupid to figure
1386 out which mailbox belongs to you. Instead of actually keeping
1387 track of which inbox belongs to which user, it uses some
1388 half-witted, guess-o-matic heuristic to try to guess your
1389 mailbox name from your username.<p>
1391 In your case it doesn't work because your username maps to more
1392 than one mailbox. For some people it doesn't work because
1393 their username maps to zero mailboxes. This is yet another
1394 inept, lame, almost criminally negligent design decision from
1395 our friends in Redmond.<p>
1397 You've got several options:
1401 Try giving fetchmail a username of "/NTDomain/NTUsername/MailboxName".
1403 Get your administrator to configure the server so that
1404 usernames and mailbox names are the same.
1406 Get your administrator to add an alias that maps your
1407 username explicitly to your mailbox name.
1410 But, the best option involves a tactical nuclear weapon (an old
1411 ASROC will do), pissing off a lot people who live downwind from
1412 Redmond, and your choice of any Linux, NetBSD, FreeBSD, or
1415 I'll provide the CD.
1418 <h2><a name="S3">S3. How can I use fetchmail with CompuServe RPA?</a></h2>
1420 First, make sure your fetchmail has the RPA support compiled in.
1421 Stock fetchmail binaries (such as you might get from an RPM) don't.
1422 You can check this by looking at the output of <code>fetchmail -V</code>;
1423 if you see the string "+RPA" after the version ID you're good to go,
1424 otherwise you'll have to build your own from sources (see the INSTALL
1425 file in the source distribution for directions).<P>
1427 Give your CompuServe pass-phrase in lower case as your password. Add
1428 `@compuserve.com' to your user ID so that it looks like `user
1429 <UserID>@compuserve.com', where <UserID> can be either
1430 your numerical userID or your E-mail nickname. An RPA-enabled
1431 fetchmail will automatically check for csi.com in the POP server's
1432 greeting line. If that's found, and your user ID ends with
1433 `@compuserve.com', it will query the server to see if it
1434 is RPA-capable, and if so do an RPA transaction rather than a
1435 plain-text password handshake.<P>
1437 <strong>Warning:</strong> the debug (-v -v) output of fetchmail will show
1438 your pass-phrase in Unicode!<P>
1440 These two .fetchmailrc entries show the difference between an RPA and
1441 non-RPA configuration:
1444 # This version will use RPA
1445 poll csi.com via "pop.site1.csi.com" with proto POP3 and options no dns
1446 user "CSERVE_USER@compuserve.com" there with password "CSERVE_PASSWORD"
1447 is LOCAL_USER here options fetchall stripcr
1449 # This version will not use RPA
1450 poll non-rpa.csi.com via "pop.site1.csi.com" with proto POP3 and options no dns
1451 user "CSERVE_USER" there with password "CSERVE_POP3_PASSWORD"
1452 is LOCAL_USER here options fetchall stripcr
1456 <h2><a name="S4">S4. How can I use fetchmail with Demon Internet's SDPS?</a></h2>
1458 <h3>Single-drop mode</h3>
1460 You can get fetchmail to download the email for just one user from
1461 Demon Internet's POP3 server by giving it a username consisting of your
1462 Demon user name followed by your account name, with an at-sign between
1465 For example, to download email for the user <philh@vision25.demon.co.uk>,
1466 you could use the following .fetchmailrc file:<P>
1469 set postmaster "philh"
1470 poll pop3.demon.co.uk with protocol POP3:
1471 user "philh@vision25" is philh
1474 <h3>Multi-drop mode</h3>
1476 Demon Internet's SDPS service is an implementation of POP3. All messages
1477 have a Received: header added when they enter the maildrop, like this:
1480 Received: from punt-1.mail.demon.net by mailstore.com for fred@xyz.demon.co.uk
1481 id 899963657:10:27896:0; Thu, 09 Jul 98 05:54:17 GMT
1484 To enable multi-drop mode you need to tell fetchmail that 'mailstore.com' is
1485 the name of the host which accepted the mail, and let it know the
1486 hostname part(s) of your E-mail address. The following example assumes
1487 that your hostname is xyz.demon.co.uk, and that you have also bought
1488 "mail forwarding" for the domain my-company.co.uk (in which case your
1489 MTA must also be configured to accept mail sent to user@my-company.co.uk)
1492 poll pop3.demon.co.uk proto pop3 aka mailstore.com no dns:
1493 localdomains xyz.demon.co.uk my-company.co.uk
1494 user xyz is * fetchall
1497 The `fetchall' command ensures that all mail is downloaded. If you
1498 want to leave mail on the server use `uidl' and `keep'; Demon does not
1499 implement the obsolete `top' command, because SDPS combines messages
1500 residing on two separate punt clusters into a single POP3 maildrop.
1501 If you do use UIDL, be aware that the "user@host" form for fetching
1502 mail from a particular Demon host will confuse fetchmail's UIDL code;
1505 Note that Demon may delete mail on the server which is more than 30
1506 days old; see their <a
1507 href="http://www.demon.net/info/helpdesk/demon_products/mail/sdps-tech.shtml">
1508 POP3 page</a> for details.<P>
1510 <h3>The SDPS extension</h3>
1512 There's a different way to do multidrop. It's not necessary on Demon
1513 Internet, since fetchmail can parse Received addresses, but the person
1514 who implemented this didn't know that. It may be useful if Demon
1515 Internet ever changes mail transports.<P>
1517 SDPS includes a non-standard extension for retrieving the envelope of a
1518 message (*ENV), which fetchmail optionally supports if compiled with the
1519 --enable-SDPS option. If you have it, the first line of the fetchmail -V
1520 response will include the string "+SDPS".<P>
1522 Once you have SDPS compiled in, fetchmail in POP3 mode will
1523 automatically detect when it's talking to a Demon Internet host in
1524 multidrop mode, and use the *ENV extension to get an envelope To address.<P>
1526 The autodetection works by looking at the hostname in the POP3
1527 greeting line; if you're accessing Demon Internet through a proxy it
1528 may fail. To force SDPS mode, pick "sdps" as your protocol.<P>
1531 <h2><a name="S5">S5. How can I use fetchmail with usa.net's servers?</a></h2>
1533 Enable `<CODE>fetchall</CODE>'. A user reports that the 2.2 version
1534 of USA.NET's POP server reports that you must use the
1535 `<CODE>fetchall</CODE>' option to make sure that all of the mail is
1536 retrieved, otherwise some may be left on the server. This is almost
1537 certainly a server bug.<P>
1539 The usa.net servers (at least in their 2.2 version, June 1998) don't
1540 handle the TOP command properly, either. Regardless of the argument
1541 you give it, they retrieve only about 10 lines of the message.
1542 Fetchmail normally uses TOP for message retrieval in order to avoid
1543 marking messages seen, but `<CODE>fetchall</CODE>' forces it to use
1546 (Note: Other failure modes have been reported on usa.net's servers.
1547 They seem to be chronically flaky. We recommend finding another
1551 <h2><a name="S6">S6. How can I use fetchmail with HP OpenMail?</a></h2>
1553 No special configuration is required, but OpenMail versions prior to
1554 6.0 have an annoying bug similar to the big one in <a
1555 href="#S2">Microsoft Exchange</a>. The message sizes it gives in the
1556 LIST are rounded to the nearest 1024 bytes. It also has a nasty habit
1557 of discarding headers it doesn't recognize, such as X- and Resent-
1560 As with M$ Exchange, the only real fix for these problems is to get a
1561 POP (or preferably IMAP) server that isn't brain-dead. OpenMail's
1562 project manager claims these bugs have been fixed in 6.0.<P>
1565 <h2><a name="S8">S8. How can I use fetchmail with Hotmail?</a></h2>
1567 You can't, yet. But <a
1568 href="http://hawkins.emu.id.au/gotmail/">gotmail</a>
1569 might be what you need.<P>
1572 <h2><a name="S9">S9. How can I use fetchmail with MSN?</a></h2>
1574 You can't. MSN uses something that looks like POP3, except the
1575 authentication part is nonstandard. And of course they don't
1576 document it, so nobody but their Windows clients can speak it.<p>
1578 This is a customer lock-in tactic; we recommend boycotting MSN as the
1579 only appropriate response.<p>
1581 As of 5.0.8, we have support for the client side of NTLM
1582 authentication. It's possible this may enable fetchmail to talk to
1583 MSN; if so, somebody should report it so this FAQ can be corrected.<p>
1586 <h2><a name="S10">S10. How can I use fetchmail with SpryNet?</a></h2>
1588 The SpryNet POP3 servers mark a message queried with TOP as seen.
1589 This means that if your connection drops in mid-message, it may end
1590 up invisibly stuck on your mail spool. Use the <code>fetchall</code>
1591 flag to ensure that it's recovered on the next cycle.<p>
1594 <h2><a name="S11">S11. How can I use fetchmail with FTGate?</a></h2>
1596 The FTGate V2 server (and possibly older versions as well) has a weird
1597 bug. It answers OK twice to a TOP request! Use the
1598 <code>fetchall</code> option to force use of RETR and work around this
1602 <h2><a name="S12">S12. How can I use fetchmail with MailMax?</a></h2>
1604 You can't. At least not if you want to be able to see attachments.
1605 MailMax has a bug; it reports the message length with attachments
1606 but doesn't download them on TOP or RETR.<P>
1608 Also, we're told that TOP sometimes fails to retrieve the entire
1609 message even when enough lines have been specified. The MailMax
1610 developers have acknowledged this bug as of 4 May 2000, but there is
1611 no fix yet. If you must use this server, force RETR with the
1612 <tt>fetchall</tt> option.<p>
1615 <h2><a name="S13">S13. How can I use fetchmail with Novell GroupWise?</a></h2>
1617 The Novell GroupWise IMAP server would be better named GroupFoolish;
1618 it is (according to the designer of IMAP) unusably broken. Among
1619 other things, it doesn't include a required content length in its
1620 BODY[TEXT] response.<p>
1622 Fetchmail works around this problem, but we strongly recommend voting
1623 with your dollars for a server that isn't brain-dead. If you stick
1624 with code as shoddy as GroupWise seems to be, you will probably pay
1625 for it with other problems.<p>
1628 <h2><a name="S14">S14. How can I use fetchmail with InterChange?</a></h2>
1630 You can't. At least not if you want to be able to see attachments.
1631 InterChange has a bug similar to the MailMax server; it reports the
1632 message length with attachments but doesn't download them on TOP or
1635 On Jan 9 2001, the people at InfiniteMail sent me mail informing me that
1636 their new 3.61.08 release of InterChange fixes this problem. I don't
1637 have any reports one way or the other yet.<p>
1640 <h2><a name="K1">K1. How can I use fetchmail with SOCKS?</a></h2>
1642 Giuseppe Guerini added a --with-socks option that supports linking
1643 with socks library. If you specify the value of this option as
1644 ``yes'', the configure script will try to find the Rconnect library
1645 and set the makefile up to link it. You can also specify a directory
1646 containing the Rconnect library.<p>
1648 Alan Schmitt has added a similar --with-socks5 option that may work
1649 better if you have a recent version of the SOCKS library.
1652 <h2><a name="S7">S7. How can I use fetchmail with geocities POP3 servers?</a></h2>
1654 Nathan Cutler reports that the the mail.geocities.com POP3 servers
1655 fail to include the first Received line of the message in the send to
1656 fetchmail. This can solve problems if your MUA interprets Received
1657 continuations as body lines and doesn't parse any of the following
1660 Workaround is to use "mda" keyword or "-mda" switch:
1662 mda "sed -e '1s/^\t/Received: /' | formail | /usr/bin/procmail -d <user>"
1664 Replace \t with exactly one tabulation character.
1666 You should also consider using "fetchall" option because Geocities' servers
1667 sometimes think that the first 45 messages have already been read.<P>
1669 Fix: Get an email provider that doesn't suck. The pop-up ads on
1670 Geocities are lame, you should boycott them anyway.<P>
1673 <h2><a name="K2">K2. How can I use fetchmail with IPv6 and IPsec?</a></h2>
1675 To use fetchmail with IPv6, you need a system that supports IPv6, the "Basic
1676 Socket Interface Extensions for IPv6" (RFC 2133).
1677 This currently means that you need to have a BSD/OS or NetBSD system with
1678 the NRL IPv6+IPsec software distribution or a Linux system with a 2.2 or
1679 later kernel and net-tools. It should not be hard to build fetchmail on
1680 other IPv6 implementations if you can port the inet6-apps kit.<P>
1682 To use fetchmail with networking security (read: IPsec), you need a system that
1683 supports IPsec, the API described in the "Network Security API for Sockets"
1684 (draft-metz-net-security-api-01.txt), and the inet6-apps kit. This currently
1685 means that you need to have a BSD/OS or NetBSD system with the NRL IPv6+IPsec
1686 software distribution. A Linux IPsec implementation supporting this API will
1687 probably appear in the coming months.<P>
1689 The NRL IPv6+IPsec software distribution can be obtained from: <a
1690 href="http://web.mit.edu/network/isakmp">http://web.mit.edu/network/isakmp</a>
1693 The inet6-apps kit can be obtained from <a href="http://ftp.ps.pl/pub/linux/IPv6/inet6-apps/">http://ftp.ps.pl/pub/linux/IPv6/inet6-apps/</a>.<P>
1695 More information on using IPv6 with Linux can be obtained from:
1698 <a href="http://www.bieringer.de/linux/IPv6/IPv6-HOWTO/IPv6-HOWTO.html">
1699 http://www.bieringer.de/linux/IPv6/IPv6-HOWTO/IPv6-HOWTO.html</a>
1701 <a href="http://www.ipv6.inner.net/ipv6">http://www.ipv6.inner.net/ipv6</a>
1704 <a href="http://www.inner.net/ipv6">http://www.inner.net/ipv6</a> (via IPv4)
1708 <h2><a name="K3">K3. How can I get fetchmail to work with ssh?</a></h2>
1710 We have five recipes for this.<P>
1712 <h3>Single-User POP3</h3>
1714 First, a lightly edited version of a recipe from Masafumi NAKANE.
1715 This one is easy to set up, but only supports one user at a time.<p>
1717 1. You must have ssh (the ssh client) on the local host and sshd (ssh
1718 server) on the remote mail server. And you have to configure ssh so
1719 you can login to the sshd server host without a password. (Refer to ssh
1720 man page for several authentication methods.)<p>
1722 2. Add something like following to your .fetchmailrc file: <p>
1725 poll mailhost port 1234 via localhost with proto pop3:
1726 preconnect "ssh -l username -f mailhost -L 1234:mailhost:110 sleep 5"
1729 This is an SSH 1.x recipe. According to Mick Papadonis, the
1730 equivalent SSH 2.x recipe is this:<p>
1733 poll localhost port 1234 with proto pop3:
1734 preconnect "ssh -n -S -x -l username -fo mailhost -L 1234:mailhost:110; sleep 5"
1737 The sleep is needed on slower machines to prevent fetchmail from
1738 trying to open the socket before ssh actually makes it ready. Faster
1739 machines may not need it.<p>
1741 (Note that 1234 can be an arbitrary port number. Privileged ports can
1742 be specified only by root.) The effect of this ssh command is to
1743 forward connections made to localhost port 1234 (in above example) to
1746 This configuration will enable secure mail transfer. All the
1747 conversation between fetchmail and remote pop server will be
1750 If sshd is not running on the remote mail server, you can specify an
1751 intermediate host running it. If you do this, however, communication
1752 between the machine running sshd and the POP server will not be encrypted.
1753 And the preconnect line would be like this:<p>
1756 preconnect "ssh -f -L 1234:mailhost:110 sshdhost sleep 20 </dev/null >/dev/null"
1759 You can work this trick with IMAP too, but the port number 110 in the
1760 above would need to become 143. In either case you'll have to specify
1761 a password but the password will not be sent in clear.<p>
1763 There is an explanation of a similar recipe at <a
1764 href="http://sunsite.unc.edu/LDP/HOWTO/mini/Secure-POP+SSH.html">Secure
1765 POP via SSH mini-HOWTO</a>.<P>
1767 <h3>Multi-User POP3</h3>
1769 Second, a recipe from Charlie Brady <cbrady@ind.tansu.com.au>:<p>
1771 Charlie says: "The recipe [from Masafume NAKANE] certainly works, but
1772 the solution I post here is better in a few respects":
1775 <LI>this method will not fail if two or more users attempt to use fetchmail
1777 <LI>you are able to use the full facilities of tcpd to control access
1778 <LI>this method does not depend on the preconnect feature of fetchmail, so
1779 can be used for tunneling of other services as well.
1786 Make sure that the "socket" program is installed on the server
1787 machine. Presently it lives at <a
1788 href="ftp://sunsite.unc.edu/pub/linux/system/network/misc/socket-1.1.tar.gz">
1789 ftp://sunsite.unc.edu/pub/linux/system/network/misc/socket-1.1.tar.gz</a>,
1790 but watch out for a change in version number.<P>
1792 Set up an unprivileged account on your system with a .ssh directory
1793 containing an SSH identity file "identity" with no pass phrase,
1794 "identity.pub" and "known_hosts" containing the host key of your
1795 mailhost. Let's call this account "noddy".
1797 On mailhost, set up no-password access for noddy@yourhost. Add to your
1798 SSH authorized_keys file:
1801 command="socket localhost 110",no-port-forwarding 1024 ......
1804 where "<code>1024 ......</code>" is the content of noddy's identity.pub file.
1806 Create a script /usr/local/bin/ssh.fm and make it executable:
1810 exec ssh -q -C -l your.login.id -e none mailhost socket localhost 110
1813 Add an entry in inetd.conf for whatever port you choose to use - say:
1816 1234 stream tcp nowait noddy /usr/sbin/tcpd /usr/local/bin/ssh.fm
1819 Send a HUP signal to your inetd.
1822 Now just use localhost:1234 to access your POP server.<P>
1824 <h3>Multi-User IMAP</h3>
1826 This one comes comes to us from Joerg Dorchain.
1827 The basic idea is to set up a bidirectional encrypted socket connection:<p>
1830 fetchmail <--> ssh <---> sshd <--> imapd
1831 \---local side--/ \-remote side-/
1834 Use ssh-keygen(1) to set up a special ssh identity with no password
1835 and RSA-only authentication, which executes /usr/sbin/imapd when
1836 authenticated. For security reasons all other commands should be
1837 disabled. (There is some security exposure in using an identity
1838 without a passphrase; it means anyone who can get access to your
1839 account could use it to read your mail).<p>
1841 Running ssh-keygen will generate two files. Have it create the
1842 private key to ~/.ssh/identity-imap. Once you have generated the
1843 corresponding public key, prepend this to the line of key data in it:
1846 command="/usr/sbin/imapd",no-port-forwarding,no-agent-forwarding
1849 This identity data has to be appended to ~/.ssh/authorized_keys on the
1850 remote machine, as usual for RSA authentication. Whenever your ssh
1851 uses this identity, the remote side will run imapd. The imapd will
1852 see that it is not running as root and go into preauthenticated
1855 On the client side, use the <code>plugin</code> keyword to make
1856 fetchmail talk to the stdin of the remote ssh. Here's an examople:
1859 poll mail.dorchain.net
1860 with options proto imap, preauth ssh, plugin fetchmail-imap-wrapper
1863 The wrapper script should look like this:<p>
1867 exec ssh -i $HOME/.ssh/identity-imap $1 /usr/sbin/imapd
1870 <h3>Netcat-based POP or IMAP tunnelling</h3>
1872 Oren Tirosh <oren@mimique.com> sends us a method of using
1873 fetchmail over ssh without port forwarding, using the plugin keyword.<P>
1875 First, set up a poll entry resembling thius one:
1878 poll target.host plugin sshtunnel proto pop3 user foo password *
1881 The important part is the "plugin sshtunnel". Now set up sshtunnel
1885 This is the sshtunnel script:
1887 ssh $1 "nc localhost $2"
1890 Thia method uses netcat to connect to the pop3 port locally on the
1891 target host and create a two-way channel over the ssh connection.<P>
1893 Oren says: "In my experience it is much more reliable than the methods
1894 described in your FAQ. ssh port forwarding often keeps the local port
1895 bound for along timeout and has timing issues requiring tricks like
1896 sleep, etc. I use this method for fetching all the mail for
1899 <h3>Using plugin</h3>
1901 Since 5.4.5, there's been a very simple recipe. Use the following option:
1904 plugin "ssh %h /usr/sbin/rimapd"
1907 You may have to use a different absolute pathname. This option tells
1908 fetchmail that instead of opening a connection on the server's port
1909 143 and doing standard IMAP authentication, fetchmail should ssh to
1910 the server and run rimapd, using the more secure ssh authentication
1911 (as well as getting ssh's encryption).<p>
1914 <h2><a name="K4">K4. What do I have to do to use the IMAP-GSS protocol?</a></h2>
1916 Fetchmail can use RFC1731 GSSAPI authorization to safely identify you
1917 to your IMAP server, as long as you can share Kerberos V credentials
1918 with your mail host and you have a GSSAPI-capable IMAP server.
1919 UW-IMAP (available via FTP at <a
1920 href="ftp://ftp.cac.washington.edu/mail/">ftp.cac.washington.edu</a>)
1921 is the only one I'm aware of and the one I recommend anyway for other
1922 reasons. You'll need version 4.1-FINAL or greater though, and it has
1923 to have GSS support compiled in.<p>
1925 Neither UW-IMAP nor fetchmail compile in support for GSS by default,
1926 since it requires libraries from the Kerberos V distribution
1927 (available via FTP at <a
1928 href="ftp://athena-dist.mit.edu/pub/ATHENA/kerberos">athena-dist.mit.edu</a>).
1929 If you have these, compiling in GSS support is simple: add a
1930 <pre>--with-gssapi=[/path/to/krb5/root]</pre> option to configure. For
1931 instance, I have all of my Kerberos V libraries installed under
1932 /usr/krb5 so I run <pre>configure --with-gssapi=/usr/krb5</pre><p>
1934 Setting up Kerberos V authentication is beyond the scope of this FAQ
1935 (you may find Jim Rome's paper <a
1936 href="http://www.ornl.gov/~jar/HowToKerb.html"> How to Kerberize your
1937 site</a> helpful), but you'll at least need to add a credential for
1938 imap/[mailhost] to the keytab of the mail server (IMAP doesn't just
1939 use the host key). Then you'll need to have your credentials ready on
1940 your machine (cf. kinit).<p>
1942 After that things are very simple. Set your protocol to imap-gss in your
1943 .fetchmailrc, and omit the password, since imap-gss doesn't need one. You
1944 can specify a username if you want, but this is only useful if your mailbox
1945 belongs to a username different from your Kerberos principal. <p>
1947 Now you don't have to worry about your password appearing in cleartext in
1948 your .fetchmailrc, or across the network.<p>
1951 <h2><a name="K5">K5. How can I use fetchmail with SSL?</a></h2>
1953 You'll need to have the <a href="http://www.openssl.org/">OpenSSL</a>
1954 libraries installed. Configure with --with-ssl. If you have the
1955 OpenSSL libraries installed in the default location (/usr/local/ssl)
1956 this will suffice. If you have them installed in a non-default
1957 location, you'll need to specify it as an argument to --with-ssl after
1960 Fetchmail binaries built this way support <code>ssl</code>,
1961 <code>sslkey</code>, and <code>sslcert</code> options that control
1962 SSL encryption. You will need to have an SSL-enabled mailserver
1963 to use these options. See the manual page for details and some words
1964 of care on the limited security provided.<p>
1966 If your open OpenSSL session dies with a message that complains "PRNG
1967 not seeded", update or improve your operating system. This means that
1968 the OpenSSL library on your machine has been unable to locate a source
1969 of random bits from which to seed its random-number generator;
1970 normally these come from the <tt>/dev/urandom</tt>, and this message
1971 probably means your OS doesn't have that device.<P>.
1973 An interactive program could seed the random number generator from
1974 keystroke timings or some other form of user input. Because fetchmail
1975 is primarily designed to run forever as a background daemon, that option
1976 option is not available in this case.<P>
1979 <h2><a name="R1">R1. Fetchmail isn't working, and -v shows `SMTP connect failed' messages.</a></h2>
1981 Fetchmail itself is probably working, but your SMTP port 25 listener
1982 is down or inaccessible.<p>
1984 The first thing to check is if you can telnet to port 25 on your smtp
1985 host (which is normally `localhost' unless you've specified an smtp
1986 option in your .fetchmailrc or on the command line) and get a greeting
1987 line from the listener. If the SMTP host is inaccessible or the listener
1988 is down, fix that first.<p>
1990 In Red Hat Linux 6.9, SMTP is disabled by default. To fix this,
1991 set "DAEMON=yes" in your /etc/sysconfig/sendmail file, then restart
1992 sendmail by running "/sbin/service sendmail restart".<p>
1994 If the listener seems to be up when you test with telnet, the most
1995 benign and typical problem is that the listener had a momentary seizure
1996 due to resource exhaustion while fetchmail was polling it -- process
1997 table full or some other problem that stopped the listener process
1998 from forking. If your SMTP host is not `localhost' or something else
1999 in /etc/hosts, the fetchmail glitch could also have been caused by
2000 transient nameserver failure. <p>
2002 Try running fetchmail -v again; if it succeeds, you had one of these
2003 kinds of transient glitch. You can ignore these hiccups, because a
2004 future fetchmail run will get the mail through. <p>
2006 If the listener tests up, but you have chronic failures trying to
2007 connect to it anyway, your problem is more serious. One way to work
2008 around chronic SMTP connect problems is to use --mda. But this only
2009 attacks the symptom; you may have a DNS or TCP routing problem. You
2010 should really try to figure out what's going on underneath before it
2011 bites you some other way. <p>
2013 We have one report (from toby@eskimo.com) that you can sometimes solve
2014 such problems by doing an <CODE>smtp</CODE> declaration with an IP
2015 address that your routing table maps to something other than the
2016 loopback device (he used ppp0).<p>
2018 We also have a report that this error can be caused by having an
2019 /etc/hosts file that associates your client host name with more than
2022 It's also possible that your DNS configuration isn't
2023 looking at <code>/etc/hosts</code> at all. If you're using libc5,
2024 look at <code>/etc/resolv.conf</code>; it should say something like
2030 so your <code>/etc/hosts</code> file is checked first. If you're
2031 running GNU libc6, check your <code>/etc/nsswitch.conf</code> file. Make
2032 sure it says something like
2038 again, in order to make sure <code>/etc/hosts</code> is seen first.<P>
2040 If you have a hostname set for your machine, and this hostname does
2041 not appear in /etc/hosts, you will be able to telnet to port 25 and
2042 even send a mail with rcpt to: user@host-not-in-/etc/hosts, but
2043 fetchmail can't seem to get in touch with sendmail, no matter what you
2044 set smtpaddress to.<p>
2046 We had another report from a Linux user of fetchmail 2.1 who solved his SMTP
2047 connection problem by removing the reference to -lresolv from his link
2048 line and relinking. Apparently in some older Linux distributions the
2049 libc bind library version works better.<p>
2051 As of 2.2, the configure script has been hacked so the bind library is
2052 linked only if it is actually needed. So under Linux it won't be, and
2053 this particular cause should go away.<p>
2056 <h2><a name="R2">R2. When I try to configure an MDA, fetchmail doesn't work.</a></h2>
2058 (I hear this one from people who have run into the blank-line problem in <a href="#X1">X1</a>.)<p>
2060 Try sending yourself test mail and retrieving it using the
2061 command-line options `<CODE>-k -m cat</CODE>'. This will dump exactly what
2062 fetchmail retrieves to standard output (plus the Received line
2063 fetchmail itself adds to the headers). <p>
2065 If the dump doesn't match what shows up in your mailbox when you
2066 configure an MDA, your MDA is mangling the message. If it doesn't
2067 match what you sent, then fetchmail or something on the server is
2071 <h2><a name="R3">R3. Fetchmail dumps core when given an invalid rc file.</a></h2>
2073 This is usually reported from AIX or Ultrix, but has even been known
2074 to happen on Linuxes without a recent version of <code>flex</code>
2075 installed. The problem appears to be a result of building with an
2076 archaic version of lex.<P>
2078 Workaround: fix the syntax of your .fetchmailrc file.<P>
2080 Fix: build and install the latest version of <a
2081 href="ftp://prep.ai.mit.edu/~ftp/pub/gnu">flex</a> from the Free
2082 Software Foundation. An FSF <a
2083 href="http://www.gnu.ai.mit.edu/order/ftp.html">mirror site</a>
2084 will help you get it faster.<P>
2087 <h2><a name="R4">R4. Fetchmail dumps core in -V mode, but operates normally otherwise.</a></h2>
2089 We've had this reported to us under Linux using libc-5.4.17 and gcc-2.7.2.
2090 It does not occur with libc-5.3.12 or earlier versions.<p>
2092 Workaround: link with GNU malloc rather than the stock C library malloc.<p>
2094 We're told there is some problem with the malloc() code in that
2095 version which makes it fragile in the presence of multiple free()
2096 calls on the same pointer (the malloc arena gets corrupted).
2097 Unfortunately it appears from doing gdb traces that whatever free()
2098 calls producing the problem are being made by the C library itself, not the
2099 fetchmail code (they're all from within fclose, and not an fclose called
2100 directly by fetchmail, either).<p>
2103 <h2><a name="R5">R5. Running fetchmail in daemon mode doesn't work.</a><br></h2>
2105 We have one report from a SunOS 4.1.4 user that trying to run
2106 fetchmail in detached daemon mode doesn't work, but that using the
2107 same options with -N (nodetach) is OK.<P>
2109 If this happens, you have a specific portability problem with the code
2110 in daemon.c that detaches and backgrounds the daemon fetchmail. Tell
2111 me about it so I can try to fix it. As a workaround, you can start
2112 fetchmail with -N and an ampersand to background it. A Sun user
2117 (fetchmail --nodetach <other params> &)
2120 The extra pair of parens is significant --- it makes sure that the process
2121 detaches from the initial shell (one more shell is started and dies
2122 immediately, detaching fetchmail and making it child of PID 1). This is
2123 important when you start fetchmail interactively and than quit
2124 interactive shell. The line above makes sure fetchmail lives after
2127 This should not happen under Linux or any truly POSIX-conformant Unix.<P>
2130 <h2><a name="R6">R6. Fetchmail hangs when used with pppd.</a></h2>
2132 Your problem may be with pppd's `demand' option. We have a report that
2133 fetchmail doesn't play well with it, but works with pppd if `demand'
2134 is turned off. We have no idea why this is.<p>
2137 <h2><a name="R7">R7. Fetchmail randomly dies with socket errors.</a></h2>
2139 Check the MTU value in your PPP interface reported by
2140 <code>/sbin/ifconfig</code>. If it's over 600, change it in your PPP
2141 options file. (<code>/etc/ppp/options</code> on my box). Here are
2142 option values that work:<P>
2150 <h2><a name="R8">R8. Fetchmail running as root stopped working after an OS upgrade</a></h2>
2152 In RH 6.0, the HOME value in the boot-time root environment changed
2153 from /root to / as the result of a change in init. Move your
2154 .fetchmailrc or use a -f option to explicitly point at the file.
2155 (Oddly, a similar problem has been reported from Debian systems.)<P>
2158 <h2><a name="R9">R9. Fetchmail is timing out after fetching certain
2159 messages but before deleting them</a></h2>
2161 There's a TCP/IP stalling problem under Redhat 6.0 (and possibly other
2162 recent Linuxes) that can cause this symptom. Brian Boutel writes:<p>
2165 TCP timestamps are turned on on my Linux boxes (I assume it's now the
2166 default). This uses 12 extra bytes per segment.
2167 When the tcp connection starts, the other end agrees a MSS of 1460,
2168 and then fragments 1460 byte chunks into 1448 and 12, because
2169 is is not allowing for the timestamp.<p>
2171 Then, for reasons I can't explain, it waits a long time (typically 2
2172 minutes) after the ack is sent before sending the next (fragmented)
2173 packet. Turning off tcp timestamps avoids the fragmentation and
2174 restores normal behaviour. To do this, [execute]<p>
2176 echo 0 > /proc/sys/net/ipv4/tcp_timestamps<p>
2178 I'm still unclear about the details of why this is happening. At least
2179 [now] I am now getting good performance and no queue blocking.
2183 <h2><a name="R10">R10. Fetchmail is timing out during message fetches</a></h2>
2185 This is probably a general networking issue. Sending a "RETR" command will
2186 cause the server to start sending large amounts of data, which means
2187 large packets. If your networking layer has a packet-fragmentation
2188 problem, that's where you'll see it.<p>
2191 <h2><a name="D1">D1. I think I've set up fetchmail correctly, but I'm not getting any mail.</a></h2>
2193 Maybe you have a .forward or alias set up that you've forgotten about. You
2194 should probably remove it.<p>
2196 Or maybe you're trying to run fetchmail in multidrop mode as root
2197 without a .fetchmailrc file. This doesn't do what you think it
2198 should; see question <a href="#C1">C1</a>.<p>
2200 Or you may not be connecting to the SMTP listener. Run fetchmail -v
2201 and see <a href="#R1">R1</a>.<p>
2204 <h2><a name="R11">R11. Fetchmail is dying with SIGPIPE.</a></h2>
2206 This probably means you have an <code>mda</code> option. Your MDA is
2207 croaking while being passed a message. Best fix is to remove the
2208 <code>mda</code> option and pass mail to your port 25 SMTP listener.<p>
2211 <h2><a name="D2">D2. All my mail seems to disappear after a dropped connection.</a></h2>
2213 One POP3 daemon used in the Berkeley Unix world that reports itself as
2214 POP3 version 1.004 actually throws the queue away. 1.005 fixed that.
2215 If you're running this one, upgrade immediately. (It also truncates
2216 long lines at column 1024)<P>
2218 Many POP servers, if an interruption occurs, will restore the whole
2219 mail queue after about 10 minutes. Others will restore it right
2220 away. If you have an interruption and don't see it right away, cross
2221 your fingers and wait ten minutes before retrying.<P>
2223 Some servers (such as Microsoft's NTMail) are mis-designed to restore
2224 the entire queue, including messages you have deleted. If you have
2225 one of these and it flakes out on you a lot, try setting a small
2226 <code>--fetchlimit</code> value. This will result in more IP connects
2227 to the server, but will mean it actually executes changes to the queue
2230 Qualcomm's qpopper, used at many BSD Unix sites, is better behaved.
2231 If its connection is dropped, it will first execute all DELE commands as
2232 though you had issued a QUIT (this is a technical violation of
2233 the POP3 RFCs, but a good idea in a world of flaky phone lines). Then it
2234 will re-queue any message that was being downloaded at hangup time.
2235 Still, qpopper may require a noticeable amount of time to do deletions
2236 and clean up its queue. (Fetchmail waits a bit before retrying in
2237 order to avoid a `lock busy' error.)<P>
2240 <h2><a name="D3">D3. Mail that was being fetched when I interrupted my fetchmail seems to have been vanished.</a></h2>
2242 Fetchmail only sends a delete mail request to the server when either
2243 (a) it gets a positive delivery acknowledgment from the SMTP
2244 listener, or (b) it gets an error 571 (the spam-filter error) from the
2245 listener. No interrupt can cause it to lose mail.<p>
2247 However, IMAP2bis has a design problem in that its normal fetch
2248 command marks a message `seen' as soon as the fetch command to get it
2249 is sent down. If for some reason the message isn't actually delivered
2250 (you take a line hit during the download, or your port 25 listener
2251 can't find enough free disk space, or you interrupt the delivery in
2252 mid-message) that `seen' message can lurk invisibly in your server
2255 Workaround: add the `<CODE>fetchall</CODE>' keyword to your fetch options.<p>
2257 Solution: switch to an <a href="http://www.imap.org">IMAP4</a> server.<p>
2260 <h2><a name="M1">M1. I've declared local names, but all my multidrop
2261 mail is going to root anyway.</a></h2>
2263 Somehow your fetchmail is never recognizing the hostname part of
2264 recipient names it parses out of To/Cc/envelope-header lines as
2265 matching the name of the mailserver machine. To check this, run
2266 fetchmail in foreground with -v -v on. You will probably see a lot of
2267 messages with the format ``line rejected, %s is not an alias of the
2268 mailserver'' or ``no address matches; forwarding to %s.'' <p>
2270 These errors usually indicate some kind of DNS configuration problem
2271 either on the server or your client machine. <p>
2273 The easiest workaround is to add a `<CODE>via</CODE>' option (if
2274 necessary) and add enough aka declarations to cover all of your
2275 mailserver's aliases, then say `<CODE>no dns</CODE>'. This will take
2276 DNS out of the picture (though it means mail may be uncollected if
2277 it's sent to an alias of the mailserver that you don't have
2280 It would be better to fix your DNS, however. DNS problems can hurt
2281 you in lots of ways, for example by making your machines
2282 intermittently or permanently unreachable to the rest of the net.<P>
2284 Occasionally these errors indicate the sort of header-parsing problem
2285 described in <a href="#M7">M7</a>.<P>
2288 <h2><a name="M2">M2. I can't seem to get fetchmail to route to a local domain properly.</a></h2>
2290 A lot of people want to use fetchmail as a poor man's internetwork
2291 mail gateway, picking up mail accumulated for a whole domain in a single
2292 server mailbox and then routing based on what's in the To/Cc/Bcc lines.<p>
2294 In general, this is not really a good idea. It would be smarter to
2295 just let the mail sit in the mailserver's queue and use fetchmail's
2296 ETRN or ODMR modes to trigger SMTP sends periodically (of course, this means
2297 you have to poll more frequently than the mailserver's expiration period).
2298 If you can't arrange this, try setting up a UUCP feed.<P>
2300 If neither of these alternatives is available, multidrop mode may do
2301 (though you <em>are</em> going to get hurt by some mailing list
2302 software; see the caveats under THE USE AND ABUSE OF MULTIDROP
2303 MAILBOXES on the man page). If you want to try it, the way to do it
2304 is with the `<CODE>localdomains</CODE>' option.<p>
2306 In general, if you use localdomains you need to make sure of two other
2309 <strong>1. You've actually set up your .fetchmailrc entry to invoke multidrop mode.</strong><p>
2311 Many people set a `<CODE>localdomains</CODE>' list and then forget
2312 that fetchmail wants to see more than one name (or the wildcard `*')
2313 in a `<CODE>here</CODE>' list before it will do multidrop routing.<p>
2315 <strong>2. You may have to set `no envelope'.</strong><p>
2317 Normally, multidrop mode tries to deduce an envelope address from a message
2318 before parsing the To/Cc/Bcc lines (this enables it to avoid losing to mailing
2319 list software that doesn't put a recipient address in the To lines).<p>
2321 Some ways of accumulating a whole domain's messages in a single server
2322 mailbox mean it all ends up with a single envelope address that is
2323 useless for rerouting purposes. You may have to set `<CODE>no
2324 envelope</CODE>' to prevent fetchmail from being bamboozled by this.<p>
2326 Check also answer <a href="#T1">T1</a> on a reliable way to do multidrop
2327 delivery if your ISP (or your mail redirection provider) is using qmail.<p>
2330 <h2><a name="M3">M3. I tried to run a mailing list using multidrop, and I have a mail loop!</a></h2>
2332 This isn't fetchmail's fault. Check your mailing list. If the list
2333 expansion includes yourself or anybody else at your mailserver (that is, not on
2334 the client side) you've created a mail loop. Just chop the host part off any
2335 local addresses in the list.<p>
2337 If you use sendmail, you can check the list expansion with
2338 <CODE>sendmail -bv</CODE>.<p>
2341 <h2><a name="M4">M4. My multidrop fetchmail seems to be having DNS problems.</a></h2>
2343 We have one report from a Linux user (not the same one as in <a
2344 href="#R1">R1</a>!) who solved this problem by removing the reference
2345 to -lresolv from his link line and relinking. Apparently in some
2346 older Linux distributions the libc5 bind library version works
2349 As of 2.2, the configure script has been hacked so the bind library is linked
2350 only if it is actually needed. So under Linux it won't be, and this problem
2354 <h2><a name="M5">M5. I'm seeing long DNS delays before each message is processed.</a></h2>
2356 Use the `<CODE>aka</CODE>' option to pre-declare as many of your
2357 mailserver's DNS names as you can. When an address's host part
2358 matches an aka name, no DNS lookup needs to be done to check it.<p>
2360 If you're sure you've pre-declared all of your mailserver's DNS names,
2361 you can use the `<CODE>no dns</CODE>' option to prevent other hostname
2362 parts from being looked up at all.<p>
2364 Sometimes delays are unavoidable. Some SMTP listeners try to call DNS
2365 on the From-address hostname as a way of checking that the address is valid.<p>
2368 <h2><a name="M6">M6. How do I get multidrop mode to work with majordomo?</a></h2>
2370 In order for sendmail to execute the command strings in the majordomo
2371 alias file, it is necessary for sendmail to think that the mail it
2372 receives via SMTP really is destined for a local user name. A normal
2373 virtual-domain setup results in delivery to the default mailbox,
2374 rather than expansion through majordomo.<P>
2376 Michael <michael@bizsystems.com> gave us a recipe for dealing
2377 with this case that pairs a run control file like this:<P>
2380 poll your.pop3.server proto pop3:
2382 localdomains virtual.localdomain1.com virtual.localdomain2.com ...
2383 user yourISPusername is root * here,
2384 password yourISPpassword fetchall
2387 with a hack on your local sendmail.cf like this:<P>
2390 #############################################
2391 # virtual info, local hack for ruleset 98 #
2392 #############################################
2394 # domains to treat as direct mapped local domain
2396 CVvirtual.localdomain1.com virtual.localdomain2.com ...
2397 ---------------------------
2399 -------------------------
2400 # handle virtual users
2402 R$+ <@ $=V . > $: $1 < @ $j . >
2403 R< @ > $+ < @ $=V . > $: $1 < @ $j . >
2404 R< @ > $+ $: $1
2405 R< error : $- $+ > $* $#error $@ $1 $: $2
2406 R< $+ > $+ < @ $+ > $: $>97 $1
2409 This ruleset just strips virtual domain names off the addresses of incoming
2410 mail. Your sendmail must be 8.8 or newer for this to work. Michael
2414 I use this scheme with 2 virtual domains and the default ISP
2415 user+domain and service about 30 mail accounts + majordomo on my
2416 inside pop3 server with fetchmail and sendmail 8.83
2420 <h2><a name="M7">M7. Multidrop mode isn't parsing envelope addresses from
2421 my Received headers as it should.</a></h2>
2423 It may happen that you're getting what appear to be well-formed
2424 sendmail Received headers, but fetchmail can't seem to extract an
2425 envelope address from them. There can be a couple of reasons for
2428 <h3>Spurious Received lines need to be skipped:</h3>
2430 First, fetchmail might be looking at the wrong Received header.
2431 Normally it looks only on the first one it sees, on the theory that
2432 that one was last added and is going to be the one containing your
2433 mailserver's theory of who the message was addressed to.<P>
2435 Some (unusual) mailserver configurations will generate extra Received
2436 lines which you need to skip. To arrange this, use the optional
2437 skip prefix argument of the `envelope' option; you may need to say
2438 something like `<code>envelope 1 Received</code>' or `<code>envelope 2
2441 <h3>The `by' clause doesn't contain a mailserver alias:</h3>
2443 When fetchmail parses a Received line that looks like
2446 Received: from send103.yahoomail.com (send103.yahoomail.com [205.180.60.92])
2447 by iserv.ttns.net (8.8.5/8.8.5) with SMTP id RAA10088
2448 for <ksturgeon@fbceg.org>; Wed, 9 Sep 1998 17:01:59 -0700
2451 it checks to see if `iserv.ttns.net' is a DNS alias of your mailserver
2452 before accepting `ksturgeon@fbceg.org' as an envelope address. This
2453 check might fail if your DNS were misconfigured, or if you were using `no dns'
2454 and had failed to declare iserv.ttns.net as an alias of your server.<P>
2457 <h2><a name="M8">M8. Users are getting multiple copies of messages.</a></h2>
2459 It's a consequence of multidrop. What's happening is that you have
2460 N users subscribed to the same list. The list software sends N
2461 copies, not knowing they will end up in the same multidrop box. Since
2462 they are both locally addressed to all N users, fetchmail delivers N
2463 copies to each user.<P>
2465 Fetchmail tries to eliminate adjacent duplicate messages in a
2466 multidrop mailbox. However, this logic depends on the message-ID
2467 being identical in both copies. It also depends on the two copies
2468 being adjacent in the server mailbox. The former is usually the case,
2469 but the latter condition sometimes fails in a timing-dependent way if
2470 the server was processing multiple incoming mail streams.
2472 I could eliminate this problem by keeping a list of all message-IDs
2473 received during a poll so far and dropping any message that matches a
2474 seen mail ID. The touble is that this is an O(N**2) operation that
2475 might significantly slow down the retriweval of large mail batches.<P>
2478 <h2><a name="X1">X1. Spurious blank lines are appearing in the headers of fetched mail.</a></h2>
2480 What's probably happening is that the POP/IMAP daemon on your
2481 mailserver is inserting a non-RFC822 header (like X-POP3-Rcpt:) and
2482 something in your delivery path (most likely an old version of the
2483 <em>deliver</em> program, which sendmail often calls to do local delivery) is
2484 failing to recognize it as a header.<p>
2486 This is not fetchmail's problem. The first thing to try is installing
2487 a current version of <em>deliver</em>. If this doesn't work, try to
2488 figure out which other program in your mail path is inserting the
2489 blank line and replace that. If you can't do either of these things,
2490 pick a different MDA (such as procmail) and declare it with the
2491 `<CODE>mda</CODE>' option.<p>
2494 <h2><a name="X2">X2. My mail client can't see a Subject line.</a></h2>
2496 First, see <a href="#X1">X1</a>. This is quite probably the same
2497 problem (X-POP3-Rcpt header or something similar being inserted by
2498 the server and choked on by an old version of <em>deliver</em>).<p>
2500 The O'Reilly sendmail book does warn that IDA sendmail doesn't process
2501 X- headers correctly. If this is your problem, all I can suggest is
2502 replacing IDA sendmail, because it's broken and not RFC822 conformant.<p>
2505 <h2><a name="X3">X3. Messages containing "From" at start of line are being split.</a></h2>
2507 If you know the messages aren't split in your server mailbox, then this
2508 is a problem with your POP/IMAP server, your client-side SMTP listener or
2509 your local delivery agent. Fetchmail cannot split messages.<p>
2511 Some POP server daemons ignore Content-Length headers and split messages on
2512 From lines. We have one report that the 2.1 version of the BSD popper
2513 program (as distributed on Solaris 2.5 and elsewhere) is broken this way.<p>
2515 You can test this. Declare an mda of `cat' and send yourself one
2516 piece of mail containing "From" at start of a line. If you see a
2517 split message, your POP/IMAP server is at fault. Upgrade to a more
2520 Sendmail and other SMTP listeners don't split RFC822 messages either.
2521 What's probably happening is either sendmail's local delivery agent or
2522 your mail reader are not quite RFC822-conformant and are breaking
2523 messages on what it thinks are Unix-style From headers. You can
2524 figure out which by looking at your client-side mailbox with vi or
2525 more. If the message is already split in your mailbox, your local
2526 delivery agent is the problem. If it's not, your mailreader is the
2529 If you can't replace the offending program, take a look at your
2530 sendmail.cf file. There will likely be a line something like<p>
2533 Mlocal, P=/usr/bin/procmail, F=lsDFMShP, S=10, R=20/40, A=procmail -Y -d $u
2536 describing your local delivery agent. Try inserting the `E' option in the
2537 flags part (the F= string). This will make sendmail turn each dangerous
2538 start-of-line From into a >From, preventing programs further downstream
2542 <h2><a name="X4">X4.</a><a name="generic_mangling">My mail is being mangled in a new and different way</a></h2>
2544 The first thing you need to do is pin down what program is doing the
2545 mangling. We don't like getting bug reports about fetchmail that are
2546 actually due to some other program's malfeasance, so please go through
2547 this diagnostic sequence before sending us a complaint.<P>
2549 There are five possible culprits to consider, listed here in the order
2550 they pass your mail:<P>
2553 <li> Programs upstream of your server mailbox.
2554 <li> The POP or IMAP server on your mailserver host.
2555 <li> The fetchmail program itself.
2556 <li> Your local sendmail.
2557 <li> Your LDA (local delivery agent), as called by sendmail or
2558 specified by <code>mda</CODE>.
2561 Often it happens that fetchmail itself is OK, but using it exposes
2562 pre-existing bugs in your downstream software, or your downstream
2563 software has a bad interaction with POP/IMAP. You need to pin down
2564 exactly where the message is being garbled in order to deduce what is
2565 actually going on.<P>
2567 The first thing to do is send yourself a test message, and retrieve it
2568 with a .fetchmailrc entry containing the following (or by running with
2569 the equivalent command-line options):<P>
2572 mda "cat >MBOX" keep fetchall
2575 This will capture what fetchmail gets from the server, except for (a)
2576 the extra Received header line fetchmail prepends, (b) header address
2577 changes due to <code>rewrite</code>, and (c) any end-of-line changes
2578 due to the <code>forcecr</code> and <code>stripcr</code> options.
2579 MBOX will in fact contain what programs downstream of fetchmail
2582 The most common causes of mangling are bugs and misconfigurations in
2583 those downstream programs. If MBOX looks unmangled, you will know
2584 that is what is going on and that it is not fetchmail's problem. Take
2585 a look at the other FAQ items in this section for possible clues about
2586 how to fix your problem.<P>
2588 If MBOX looks mangled, the next thing to do is compare it with your
2589 actual server mailbox (if possible). That's why you specified
2590 <code>keep</code>, so the server copy would not be deleted. If your
2591 server mailbox looks mangled, programs upstream of your server mailbox
2592 are at fault. Unfortunately there is probably little you can do about
2593 this aside from complaining to your site postmaster, and nothing at
2594 all fetchmail can do about it!<P>
2596 More likely you'll find that the server copy looks OK. In that case
2597 either the POP/IMAP server or fetchmail is doing the mangling. To
2598 determine which, you'll need to telnet to the server port and simulate
2599 a fetchmail session yourself. This is not actually hard (both POP3
2600 and IMAP are simple, text-only, line-oriented protocols) but requires
2601 some attention to detail. You should be able to use a fetchmail -v
2602 log as a model for a session, but remember that the "*" in your LOGIN
2603 or PASS command dump has to be replaced with your actual password.<P>
2605 The objective of manually simulating fetchmail is so you can see
2606 exactly what fetchmail sees. If you see a mangled message, then your
2607 server is at fault, and you probably need to complain to your
2608 mailserver administrators. However, we like to know what the broken
2609 servers are so we can warn people away from them. So please send
2610 us a transcript of the session including the mangling <em>and the
2611 server's initial greeting line</em>. Please tell us anything else
2612 you think might be useful about the server, like the server host's
2613 operating system.<P>
2615 If your manual fetchmail simulation shows an unmangled message,
2616 congratulations. You've found an actual fetchmail bug, which is a
2617 pretty rare thing these days. Complain to us and we'll fix it.
2618 Please include the session transcript of your manual fetchmail
2619 simulation along with the other things described in the FAQ entry on
2620 <a href="#G3">reporting bugs</a>.
2623 <h2><a name="X5">X5. Using POP3, retrievals seems to be fetching too much!</a></h2>
2625 This may happen in versions of fetchmail after 4.4.1 and before 4.4.8.
2626 Versions after 4.4.1 use POP3's TOP command rather than RETR, in order
2627 to avoid marking the message seen (leaving it unseen is helpful for
2628 later recovery if you lose your connection in the middle of a
2631 Versions of fetchmail from 4.4.2 through 4.4.7 had a bad interaction
2632 with Eudora qpopper versions 2.3 and later. The TOP bounds check was
2633 fooled by an overflow condition in the TOP argument. Decrementing the
2634 TOP argument in 4.4.7 fixed this.<P>
2636 Fix: Upgrade to a later version of fetchmail.<P>
2638 Workaround: set the <code>fetchall</code> option. Under POP3
2639 this has the side effect of forcing RETR use.<P>
2642 <h2><a name="X6">X6. My mail attachments are being dropped or mangled.</a></h2>
2644 This isn't fetchmail's doing -- fetchmail never drops lines in message
2645 bodies or attachments. It may be your POP server, or it may be the
2646 sender's mail user agent (or a bad combination of both).<p>
2648 The Mail Max POP3 server and the InterChange and Imail IMAP servers
2649 are known to simply drop MIME attachments when uploading messages.
2650 We've had sporadic reports of problems with Microsoft Exchange and
2651 Outlook servers. Windows- and NT-based POP servers seem especially
2652 prone to mangle attachments. If you are running one of these,
2653 replacing your server with a Unix machine is probably the only
2654 effective solution.<p>
2656 We've also had a report that Lotus Notes sometimes trashes the
2657 MIME type of messages. In particular, it seems to modify MIME
2658 headers introducing type application/pdf, mangling the type
2659 to application/octet-stream. It may corrupt other MIME types
2662 The IMAP service of Lotus Domino has a known bug in the way it
2663 generates MIME Content-type headers (observed on Lotus Domino 5.0.2b).
2664 It's a subtle one that doesn't show up when Netscape Messenger and
2665 other clients use a FETCH BODY[] to grab the whole message. When
2666 fetchmail uses FETCH RFC822.HEADER and FETCH RFC822.TEXT to get first
2667 the header and then the body, Domino generates different Boundary tags
2668 for each part, .e.g. one tag is declared in the Content-type header and
2669 another is used to separate the MIME parts in the body. This doesn't
2670 work. (I have heard a rumor that this bug is scheduled to be fixed
2671 in Domino release 6; you can find a workaround at contrib/domino.)<p>
2673 Another rich source of attachment problems is Microsoft Exchange and
2674 Microsoft Outlook. If you see unreadable attachments with a
2675 ContentType of "application/x-tnef", you're having this problem. The
2676 <a href="http://world.std.com/~damned/software.html">TNEF</a> utility
2679 Rob Funk explains: Unfortunately there also remain many mail user
2680 agents that don't write correct MIME messages. One big offender is Sun
2681 MailTool attachments, which are formatted enough like MIME that some
2682 programs could get confused; these are generated by the mailtool and
2683 dtmail programs (the mail programs in Sun's OpenWindows and CDE
2686 One solution to problems related to misformatted MIME attachments is
2687 the <a href="ftp://ftp.uu.se/pub/unix/networking/mail/emil/">emil</a>
2689 href="ftp://ftp.uu.se/pub/unix/networking/mail/emil/TUTORIAL.html">tutorial</a>
2690 file at that site for details on emil. It is useful for
2691 converting character sets, attachment encodings, and attachment
2692 formats. At this writing, emil does not appear to have been
2693 maintained since a patch to version 2.1.0beta9 in late 1997, but it is
2696 One good way of using emil is from within procmail. You can have
2697 procmail look for signs of problematic message formatting, and pipe
2698 those messages through emil to be fixed. emil will not always be able
2699 to fix the problem, in which case the message is unchanged.<p>
2701 A possible rule to be inserted into a .procmailrc file for using emil
2706 * 1^1 ^Content-Type: \/X-sun[^;]*
2707 * 1^1 ^Content-Type: \/application/mac-binhex[^;]*
2708 * 1^1 ^Content-Transfer-Encoding: \/x-binhex[^;]*
2709 * 1^1 ^Content-Transfer-Encoding: \/x-uuencode[^;]*
2711 LOG="Converting $MATCH
2714 | emil -A B -T Q -B BA -C iso-8859-1 -H Q -F MIME \
2715 | gawk '{gsub(/\r\n?/,"\n");print $0}'
2719 The "1^1" in the conditions is a way of specifying to procmail that if
2720 any one of the four listed expressions is found in the message, the
2721 total condition is considered true, and the message gets passed into
2722 emil. These four subconditions check whether the message has a Sun
2723 attachment, a binhex attachment, or a uuencoded attachment; there are
2724 others that could be added to check these things better and to check
2725 other relevant conditions. The "LOG=" line writes a line into the
2726 procmail log; the lone double-quote beginning the following line makes
2727 sure the log entry gets an end-of-line character. The call to gawk
2728 (GNU awk) is for fixing end-of-line conventions, since emil sometimes
2729 leaves those in the format of the originating machine; it could
2730 probably be replaced with a sed subsitution.<p>
2732 The emil call itself tries to ensure that the message uses:
2734 <li> BinHex encoding for any Apple Macintosh-only attachments
2735 <li> Quoted-Printable encoding for text (when necessary)
2736 <li> Base64 Encoding for binary attachments
2737 <li> iso-8859-1 character set for text (unfortunately emil can't yet
2738 convert from windows-1252 to iso-8859-1)
2739 <li> Quoted-Printable encoding for headers
2740 <li> MIME attachment format
2743 Most of these (the primary exceptions being the character set and the
2744 Apple binary format) are as they should be for good internet
2745 interoperability.<p>
2747 Some mail servers (Lotus Domino is a suspect here) mangle
2748 Sun-formatted messages, so the conversion to MIME needs to happen
2749 before such programs see the message. The ideal is to rid the world
2750 of Sun-formatted messages: don't use mailtool for sending attachments
2751 (it doesn't understand MIME anyway, and most of the world doesn't
2752 understand its attachments, so it really shouldn't be used at all),
2753 and make sure dtmail is set to use MIME rather than mailtool's format.<p>
2756 <h2><a name="X7">X7. Some mail attachments are hanging fetchmail.</a></h2>
2758 This isn't fetchmail's problem either; fetchmail doesn't know anything
2759 about mail attachments and doesn't treat them any differently from
2760 plain message data.<P>
2762 The most usual cause of this problem seems to be bugs in your network
2763 transport layer's capability to handle the very large TCP/IP packets
2764 that attachments tend to turn into. You can test this theory by trying to
2765 download the offending message through a webmail account; using HTTP
2766 for the message tends to simulate large-packet stress rather well, and
2767 you will probably find that the messages that seem to be choking
2768 fetchmail will make your HTTP download speed drop to zero.<P>
2770 This problem can be caused by subtle bugs in the packet-reassembly
2771 layer of your TCP/IP stack; these often don't manifest at normal
2772 packet sizes. It may also be caused by malfunctioning path-MTU
2773 discovery on the mailserver. Or, if there's a modem in the link,
2774 it may be because the attachment contains the Hayes mode escape "+++".
2777 <h2><a name="O1">O1. The --logfile option doesn't work if the logfile doesn't exist.</a></h2>
2779 This is a feature, not a bug. It's in line with normal practice for
2780 system daemons and allows you to suppress logging by removing the log,
2781 without hacking potentially fragile startup scripts. To get around
2782 it, just touch(1) the logfile before you run fetchmail (this will have
2783 no effect on the contents of the logfile if it already exists).<P>
2786 <h2><a name="O2">O2. Every time I get a POP or IMAP message the header
2787 is dumped to all my terminal sessions.</a></h2>
2789 Fetchmail uses the local sendmail to perform final delivery, which
2790 Netscape and other clients doesn't do; the announcement of new messages
2791 is done by a daemon that sendmail pokes. There should be a ``biff''
2792 command to control this. Type
2798 to turn it off. If this doesn't work, try the command
2804 which is essentially what <code>biff -n</code> will do. If this
2805 doesn't work, comment out any reference to ``comsat'' in your
2806 /etc/inetd.conf file and restart inetd.<P>
2808 In Slackware Linux distributions, the last line in /etc/profile is
2820 to solve the problem system-wide.<P>
2823 <h2><a name="O3">O3. Does fetchmail reread its rc file every poll cycle?</a></h2>
2825 No, but versions 5.2.2 and later will notice when you modify your rc
2826 file and restart, reading it.
2829 <h2><a name="O4">O4. Why do deleted messages show up again when I take
2830 a line hit while downloading?</a></h2>
2832 Because you're using a POP3 other than Qualcomm qpopper, or an IMAP
2833 with a long expunge interval.<P>
2835 According to the POP3 RFCs, deletes aren't actually performed until
2836 you issue the end-of-session QUIT command. Fetchmail cannot fix this,
2837 because doing it right takes cooperation from the server. There are
2838 two possible remedies:<P>
2840 One is to switch to qpopper (the free POP3 server from Qualcomm,
2841 the Eudora people). The qpopper software violates the POP3 RFCs by
2842 doing an expunge (removing deleted messages) on a line hangup, as well
2843 as on processing a QUIT command.<P>
2845 The other (which we recommend) is to switch to <a
2846 href="http://www.imap.org">IMAP</a>. IMAP has an explicit expunge
2847 command and fetchmail normally uses it to delete messages immediately
2848 after they are downloaded.<P>
2850 If you get very unlucky, you might take a line hit in the window
2851 between the delete and the expunge. If you've set a longer expunge
2852 interval, the window gets wider. This problem should correct itself
2853 the next time you complete a successful query.<P>
2856 <h2><a name="O5">O5. Why is fetched mail being logged with my name, not the real From address?</a></h2>
2858 Because logging is done based on the address indicated by the sending
2859 SMTP's MAIL FROM, and some listeners are picky about that address.<p>
2861 Some SMTP listeners get upset if you try to hand them a MAIL FROM
2862 address naming a different host than the originating site for your
2863 connection. This is a feature, not a bug -- it's supposed to help
2864 prevent people from forging mail with a bogus origin site. (RFC 1123
2865 says you shouldn't do this exclusion...)<p>
2867 Since the originating site of a fetchmail delivery connection is
2868 localhost, this effectively means these picky listeners will barf on
2869 any MAIL FROM address fetchmail hands them with an @ in it!<p>
2871 Versions 2.1 and up try the header From address first and fall back to
2872 the calling-user ID. So if your SMTP listener isn't picky, the log
2876 <h2><a name="O6">O6. I'm seeing long sendmail delays or hangs near the start of each poll cycle.</a></h2>
2878 Sendmail does a hostname lookup when it first starts up, and also each
2879 time it gets a HELO in listener mode.<p>
2881 Your resolver configuration may be causing one of these lookups to
2882 fail and time out. Check <code>/etc/resolv.conf</code> and
2883 <code>/etc/hosts</code> file. Make sure your hostname and
2884 fully-qualified domain name are both in <code>/etc/hosts</code>, and
2885 that hosts is looked at before DNS is queried. You probably also want
2886 your remote mail server(s) to be in the hosts file.<p>
2888 You can suppress the startup-time lookup if need to by reconfiguring
2889 with <code>FEATURE(nodns)</code>.<p>
2891 Configuring your bind library to cache DNS lookups locally may help,
2892 and is a good idea for speeding up other services as well. Switching to
2893 a faster MTA like qmail or exim might help. <p>
2896 <h2><a name="O7">O7. Why doesn't fetchmail deliver mail in date-sorted order?</a></h2>
2898 Because that's not the order the server hands it to fetchmail in.<P>
2900 Fetchmail getting mail from a POP server delivers mail in the order
2901 that your server delivers mail. Fetchmail can't do anything about
2902 this; it's a limitation of the underlying POP protocol.<P>
2904 In theory it might be possible for fetchmail in IMAP mode to sort
2905 messages by date, but this would be in violation of two basics of
2906 fetchmail's design philosophy: (a) to be as simple and transparent a
2907 pipe as possible, and (b) to <em>hide</em>, rather than emphasize, the
2908 differences between the remote-fetch protocols it uses.<P>
2910 Re-ordering messages is a user-agent function, anyway.<P>
2913 <h2><a name="O8">O8. I'm using pppd. Why isn't my monitor option working?</a></h2>
2915 There is a combination of circumstances that can confuse fetchmail.
2916 If you have set up demand dialing with pppd, and pppd has an idle
2917 timeout, and you have lcp-echo-interval set, then the
2918 lcp-echo-interval time must be longer than the pppd idle timeout.
2919 Otherwise it is going keep increasing the packet counters that fetchmail
2920 relies upon, triggering fetchmail into polling after its own delay
2921 interval and thus preventing the pppd link from ever reaching its
2922 inactivity timeout.<p>
2925 <h2><a name="O9">O9. Why does fetchmail keep retrieving the same messages
2926 over and over?</a></h2>
2928 First, check to see that you haven't enabled the <cite>keep</cite>
2929 and <cite>fetchall</cite> option. If you have, turn <cite>keep</cite> off.<p>
2931 There are various forms of lossage involving the POP3 UIDL feature
2932 that can lead to all your old messages being seen again after a line
2933 drop. I have given up trying to fix these, as the UIDL code breaks
2934 worse every time I touch it. The problem is fundamental; maintaining
2935 and garbage-collecting the right kind of client-side state is just
2936 hard. Whoever put UIDLs in RFC1725 and removed LAST should be hung
2937 up by his thumbs and whipped with scorpions. The right answers are
2938 either (a) live with the occasional breakage, (b) switch to IMAP4,
2939 or (c) fix the code yourself and send me a patch. Unless you choose
2940 (c), I don't want to hear about it.<p>
2942 This can also happen when some other mail client is logged in to your
2943 mail server, if it uses a simple exclusive-locking scheme (and many,
2944 especially most POP3 servers, do exactly that). Your fetchmail is
2945 able to retrieve the messages, but because the mailbox is write-locked
2946 by the other instance yours can neither mark messages seen or delete them.
2947 The solution is to either (a) wait for the other client to finish, or (b)
2950 James Stevens <James.Stevens@kyzo.com> writes:<p>
2953 We had a Linux box dialing the Net and collecting mail from an NT POP3
2954 server. Fetchmail was correctly collecting and deleting each e-mail
2955 one by one. However,the dial-up connection was very unreliable and
2956 would often just drop out in the middle of a session.<p>
2958 Interestingly, unless the TCP POP3 connection was terminated normally
2959 (I guess with a POP3 "QUIT" command) NT would then roll back all the
2962 This meant if the first e-mail was very large it might just end up
2963 continuously collecting it, basically jamming the queue. Or, if the
2964 queue became very full itmight never get a long enough phone
2965 connection to retrieve the entire mailbox, and NT would roll back any
2966 deletes, so it would end up collecting (and delivering) the first few
2967 e-mails again and again. As the POP3 mailbox became fuller and fuller
2968 the chances of getting a connection long enough to collect theentire
2969 mailbox became smaller and smaller.<p>
2971 Our solution was to make fetchmail only collect a few (say 5 or 10)
2972 e-mails at atime, thus trying to ensure that the POP3 connection is
2973 terminated correctly.
2976 Unfortunately, this is exactly the way POP3 servers are supposed
2977 to behave on a line drop, according to the RFCs. I recommend
2978 switching to IMAP and using a short expunge interval.<p>
2981 <table width="100%" cellpadding=0><tr>
2982 <td width="30%">Back to <a href="index.html">Fetchmail Home Page</a>
2983 <td width="30%" align=center>To <a href="/~esr/sitemap.html">Site Map</a>
2984 <td width="30%" align=right>$Date: 2001/02/12 01:14:27 $
2987 <P><ADDRESS>Eric S. Raymond <A HREF="mailto:esr@thyrsus.com"><esr@snark.thyrsus.com></A></ADDRESS>
2992 compile-command: "(cd ~/WWW; upload fetchmail/fetchmail-FAQ.html)"