- dovecot - dovecot.org

[Dovecot] sqlite patch
by Jakob Hirsch 22 Jan '06

22 Jan '06

Hi, http://plonk.de/sw/dovecot/dovecot.sqlite.4.diff It uses sqlite3 and is not heavily tested, but it works quite fine here. Have fun!

2 2

[Dovecot] Dovecot-LDA & sieve
by Björn Ketelaars 22 Jan '06

22 Jan '06

I’m trying to make use of sieve by means of the Dovecot LDA Wiki - http://wiki.dovecot.org/LDA. Looking at maillog, mail is being delivered by means of deliver to the right user: Jan 22 16:05:31 gateway sm-mta[17258]: STARTTLS=server, relay=[10.0.0.100], version=TLSv1/SSLv3, verify=NO, cipher=DHE-RSA-AES256-SHA, bits=256/256 Jan 22 16:05:32 gateway sm-mta[17258]: k0MF5VJ2017258: from=<bjorn(a)hydroxide.nl>, size=354, class=0, nrcpts=1, msgid=<43D39F31.3080502(a)hydroxide.nl>, proto=ESMTP, daemon=MTA, relay=[10.0.0.100] Jan 22 16:05:32 gateway sm-mta[32229]: k0MF5VJ2017258: to=| "/usr/local/libexec/dovecot/deliver", ctladdr=<test(a)hydroxide.nl> (1000/1000), delay=00:00:00, xdelay=00:00:00, mailer=prog, pri=30633, dsn=2.0.0, stat=Sent Unfortunately .dovecot.sieve is left alone and thus nothing is being sorted. Any suggestions maybe?

1 0

[Dovecot] Resync dovecot index after manual moving mails in maildir
by Danilo Reinhardt 22 Jan '06

22 Jan '06

Hello, i'm new to dovecot and this list. Dovecot works very well after all! I now have a small problem, maybe there is a solution. For automatically spam learning i've created a maildir folder called "spam learn". Every hour spamassassin scans this folder and learn from the mails or files at this folder. After this learn i move the files to trash. If at this moment a client is connected via imap to this account, the mua (in my case thunderbird) doesnt update its folder view after reselecting the folder. Is this a mua problem or is this a index "problem" in dovecot (i know, it isnt a real problem or bug). If the second is the truth, is there any possibility to tell dovecot to rescan the maildir? Thanks in advance. Bye from Germay Danilo

2 4

[Dovecot] %d interpolation problem fixed
by William Ross 22 Jan '06

22 Jan '06

On 6 Dec 2005, at 17:21, Timo Sirainen wrote: > On Tue, 2005-12-06 at 17:11 +0000, William Ross wrote: >> On 6 Dec 2005, at 16:51, Timo Sirainen wrote: >> >>> On Fri, 2005-10-28 at 17:52 +0200, Riccardo Bini wrote: >>>>> I just wrote this: http://wiki.dovecot.org/Quota >>>> Well... but you have to correct %n with vpopmail backend. %u in the >>>> default_mail_env (like in the log) is always the user name without >>>> the >>>> domain.... >>> >>> That's a bug in Dovecot's vpopmail code. I don't know why it doesn't >>> work. Would be nice if someone using vpopmail fixed that.. >> >> point me in the right direction and I'll have a go. > > src/auth/passdb-vpopmail.c > > If user logs in as user@domain, then they both should be in > request->user string. %d is taken from there. I don't see request- > >user > being modified in that file either, so I think it should have worked.. I finally got round to working on this. I'm a perl programmer, and it has all been very educational, so please bear with me :) I was stuck because the domain name was not interpolated into the Maildir path[0], when using the configuration recommended here: http://wiki.dovecot.org/ QuestionsAndAnswers#head-153aa253aff40951a05cf238da4245b1cc1a4665 It turns out that the answer is really simple. You don't need to supply a default_mail_env parameter with vpopmail and dovecot v1.x. Indeed, you shouldn't: the path to the maildir is retrieved from vpopmail and if you specify default_mail_env then it will be used in place of that correct value. Also, it will be used wrongly, as there is also an interpolation bug, and %d isn't replaced correctly. The code in passdb-vpopmail and userdb-vpopmail works perfectly, by the way: the login email address[1] is parsed correctly, the vpopmail hookup is correct, and the vqpasswd struct that is returned by vpopmail_lookup_vqp contains all the right values. eg (these are debugging-for-dummies lines that I've scattered everywhere): > dovecot: Jan 17 22:51:42 Info: auth(default): * > vpopmail_verify_plain: now vpop_user = sysadmin, vpop_domain = > spanner.org, request->user = sysadmin(a)spanner.org, vpw->pw_dir = / > home/vpopmail/domains/spanner.org/sysadmin in the callback method I got this tab-separated auth_reply_stream: > dovecot: Jan 18 00:30:01 Info: auth(default): userdb_callback str: > USER 2 sysadmin uid=7797 gid=2109 home=/home/ > vpopmail/domains/spanner.org/sysadmin but I still saw these lines later: > dovecot: Jan 18 01:31:46 Info: imap(sysadmin): maildir: root=/home/ > vpopmail/domains//sysadmin/Maildir, index=/home/vpopmail/domains// > sysadmin/Maildir, control=, inbox= I became suspicious, and tested with nonsense values in default_mail_env: sure enough i got the same thing. During the userdb stage we have the correct path to the maildir, but later on we see a different value derived from the nonsense configuration I've just entered. I guess that we're getting the initial value from vpopmail and try the same thing but with no default_mail_env entry at all. It works perfectly. I say some new words I have learned from my wife. I don't know where the failure lies, but I can tell you that the 'user' value passed to get_var_expand_table (which comes from getenv ("USER")) contains only the first part of the email address. That's probably not the intention, because get_var_expand_table then tries to split it on @. As it happens the home-directory value is passed to get_var_expand_table as the expansion for %h, so this works too, in a completely redundant way: default_mail_env = maildir:%h I've tested this with an new tarball of 1.0.beta1 and it works fine - with the configuration change - so if this all sounds sane to you then I'll update the wiki. best, will [0] under vpopmail the maildir for the address brian(a)domain.com usually lives somewhere like: /home/vpopmail/domains/domain.com/brian/, [1] in order to disambiguate usernames, vpopmail has you log in with your full email address as the user id. dovecot would therefore need to split the email address and interpolate the fragments into the maildir path to get to the right place. however, the vpopmail password file also specifies the maildir for each user, which i guess is what dovecot is using now.

2 5

[Dovecot] new configure-option --disable-index
by Marten Lehmann 22 Jan '06

22 Jan '06

Hello, I experimented with dovecot the whole evening yesterday and the mailserver worked fine since there haven't been that much users connecting to it. I'm using dovecot 1.0.alpha5 as pop3 and pop3s server. The maildirs are mounted through NFS and since there is no nfslockd running I disabled mmap in the config and told dovecot to use dotlock. Also I experienced, that dovecot was much faster when using one login-process per connection instead of sharing them throughout several connections. Sharings seems to be a greater performance problem (bottleneck) than additional fork()s. However, under normal load at business time the server load increased to a value that made it impossible to download emails without timeouts, so we switched back to tcpserver/qmail-pop3d for pop3 and only kept dovecot for doing the pop3s part. Dovecot is really fine, but it still isn't NFS ready. Especially when using dovecot as pop3 only this index-stuff must be disengageable. What is the use of generating a complex index file for hundreds of messages, that will be deleted in the next step after downloading them? But the index is also a problem when using IMAP over NFS. Access to the files is too slow, so the connections keep hanging around and the i/o waiting value is raised to a maximum which slows down the whole machine. It simply doesn't make sense to build and update an index every time the maildir is accessed (which is very cpu and timeconsuming), while search operations (the only moment when the index is actually used and not just heavily updated) don't affect even 1% of all IMAP calls (and such a command doesn't even exist for pop3). I would be much obliged if you could create an option called "--disable-index" in the configure-script or something like "disable-index = yes" in dovecot.conf that makes dovecot ignoring existing index files and doesn't care for related operations, so NFS should work much faster and IMAP search would simply parse the files directly instead of looking in the index. I would really appriciate this very much and I'm absolutely sure, that a lot of people beside me are waiting for this. Regards Marten

7 22

[Dovecot] core files
by jared 22 Jan '06

22 Jan '06

Hey all: Was playing around with kill -11 on dovecot processes to see which ones would dump core and which ones wouldn't. If ulimit is correct dovecot and dovecot-auth will dump. If you perform the recommended (Set mail_drop_priv_before_exec = yes in config file) step, imap will also dump core. But I'm unable to force imap-login to dump core. I'm curious if this is possible without the linux kernel patch that's recommended on http://dovecot.org/bugreport.html. (BTW tried the patch on my 2.4.18 kernel but the patch had all kinds of errors.) Thanks, Jared DeMott

2 1

[Dovecot] Assertation failed when trying to login twice in one session
by Steffen Kaiser 22 Jan '06

22 Jan '06

Hello, this command: echo "user testu pass q" ; sleep 5; echo "user testu pass test quit" | nc localhost pop3 results in this assert: dovecot: Jan 20 10:16:09 Error: auth(default_with_listener): file auth-client-connection.c: line 33 (auth_client_send): assertion failed: (conn->refcount > 1) dovecot: Jan 20 10:16:09 Error: child 18656 (auth) killed with signal 6 without listener: dovecot: Jan 20 10:24:07 Error: auth(default): file auth-client-connection.c: line 33 (auth_client_send): assertion failed: (conn->refcount > 1) dovecot: Jan 20 10:24:07 Error: child 18733 (auth) killed with signal 6 The password of user 'testu' is 'test' :) The same with IMAP: 1 login q q 1 NO Authentication failed. 2 login testu test 2 OK Logged in. 3 SELECT INBOX * FLAGS (\Answered \Flagged \Deleted \Seen \Draft) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft \*)] Flags permitted. * 0 EXISTS * 0 RECENT * OK [UIDVALIDITY 1137748810] UIDs valid * OK [UIDNEXT 1] Predicted next UID 3 OK [READ-WRITE] Select completed. The connection is OK, though. dovecot: Jan 20 10:19:26 Info: imap(testu): maildir: root=/home/testu/MailDir, index=, control=/var/spool/dovecot/30001, inbox= dovecot: Jan 20 10:19:27 Error: auth(default): file auth-client-connection.c: line 33 (auth_client_send): assertion failed: (conn->refcount > 1) dovecot: Jan 20 10:19:27 Error: child 18710 (auth) killed with signal 6 Bye, -- Steffen Kaiser

2 2

[Dovecot] dovecot 1.0beta1 making base_dir world-writable
by Brad 22 Jan '06

22 Jan '06

I can also confirm this, on my own mail server since upgrading to beta1. Dovecot always makes the base_dir world-writable on startup. ----- Forwarded message from Oleg Safiullin <form(a)pdp-11.org.ru> ----- Date: Fri, 20 Jan 2006 07:09:17 +0600 From: Oleg Safiullin <form(a)pdp-11.org.ru> To: brad(a)openbsd.org Subject: dovecot 1.0beta1 Hi. After update dovecot to 1.0beta1 some strange thing happens -- during startup it says "fixing permissions for /var/dovecot to make it world-readable", but actually it makes it world-writable. Please look at it. WBR //form ----- End forwarded message -----

2 1

[Dovecot] Simultaneous POP login - inbox corrumption
by John Woodell 22 Jan '06

22 Jan '06

When two POP clients connect to the same account, within a second or two, the inbox can get corrumpted. Here is a sample of the first few lines of the corrumpted inboxes... # 1-- # # # From... # # > # # From.. # 4504 # # From... # 7-- # # # From... # 4504 # # From... # # > # # From... # 868-- # # From...

2 1

[Dovecot] 1.0 beta 1 feature request: passwd-file vs. prefetch
by Matthias Andree 22 Jan '06

22 Jan '06

Greetings, it appears as though 1.0.beta1 doesn't support "userdb prefetch {}" with "passdb passwd-file { arg=/etc/mailpasswd }" even if the latter contains all necessary information. Is this planned for a future version or would the benefits be too small to justify that? Another idea: can userdb prefetch {} fall back to retrying the passdb entries to see if they come up with useful information? -- Matthias Andree

2 3