- dovecot - dovecot.org

[Dovecot] LDA null logging patch
by Joseph Tam 20 Apr '11

20 Apr '11

I don't really need logging of successful local deliveries since the MTA provides enough data. I can't get rid of LDA delivery logs by setting deliver_log_format = All this does is produce an empty log message e.g. Apr 16 01:22:05 dserver dovecot: lda(user): Timo, would you consider inclusion of this patch which will skip logging deliveries if deliver_log_format is empty: --- src/lib-lda/mail-deliver.c~ Sat Apr 16 01:27:23 2011 +++ src/lib-lda/mail-deliver.c Sat Apr 16 01:29:05 2011 @@ -100,2 +100,5 @@ + if (!(ctx->set->deliver_log_format) || !(*ctx->set->deliver_log_format)) + return; + va_start(args, fmt); Joseph Tam <jtam.home(a)gmail.com>

2 1

[Dovecot] Debug new emails?
by Beto Moreno 20 Apr '11

20 Apr '11

Hi guys. I would like to debug all the new emails arrive to the users mailbox, I have some issue with 1 users that claim that is losing some important emails. The issue that I are not logging to much info in dovecot, I had read the mail_log plugin wiki is not difficult to setup. I want to see went the users receive the new emails, because postfix have the log that say" "sent" from the envelop to the issue user, but I the user say that he didn't receive the email, check his emails and is true I didn't see those 2 emails that postfix send to the user mailbox. I check my spam server and is not blocking that envelop, the spam is not causing this. I'm using Centos 5.4 Postfix 2.3.x Dovecot 1.0.x,but reading mail_log plugin say that we need 1.2.x, can I still use this plugin or I have other options? Thanks!!!

3 3

[Dovecot] expire-tool skipping mailbox
by Alex Reinhart 20 Apr '11

20 Apr '11

Hello, I have dovecot set up with expire-tool to nuke Junk messages greater than 30 days old. On a server with many user accounts, it seems to work on most of them. However, on my user account, expire-tool skips my Junk folder entirely, despite it containing messages from January of this year. Here is the expire-tool --test output: Info: alex(a)scienceforums.net/Junk: timestamp 1284027481 (Thu Sep 9 05:18:01 2010) -> 1304023056 (Thu Apr 28 15:37:36 2011) Then it skips to the next mailbox. (Note that I intentionally set the timestamp back to force it to consider my mailbox.) I don't understand why it doesn't want to expunge my 2700 junk messages when it will expunge several thousand in other accounts happily. The expire plugin has been enabled for probably over a year, so all of the messages should be tracked. My only hypothesis is that it has something to do with variables in the mail_location; however, mail_location should be overridden by a user_query, though that contains variables too. Would this account for the issue? It seems unlikely, since other users on the server work just fine. dovecot -n output for the interested: # 1.2.9: /etc/dovecot/dovecot.conf # OS: Linux 2.6.32-24-server x86_64 Ubuntu 10.04.2 LTS ext3 base_dir: /var/run/dovecot/ log_timestamp: %Y-%m-%d %H:%M:%S protocols: managesieve imaps pop3s listen(default): *:143 listen(imap): *:143 listen(pop3): *:110 listen(managesieve): * ssl_listen(default): *:993 ssl_listen(imap): *:993 ssl_listen(pop3): *:995 ssl_listen(managesieve): login_dir: /var/run/dovecot//login login_executable(default): /usr/lib/dovecot/imap-login login_executable(imap): /usr/lib/dovecot/imap-login login_executable(pop3): /usr/lib/dovecot/pop3-login login_executable(managesieve): /usr/lib/dovecot/managesieve-login first_valid_uid: 150 last_valid_uid: 150 mail_privileged_group: mail mail_location: maildir:/var/vmail/%d/%u mbox_write_locks: fcntl dotlock mail_executable(default): /usr/lib/dovecot/imap mail_executable(imap): /usr/lib/dovecot/imap mail_executable(pop3): /usr/lib/dovecot/pop3 mail_executable(managesieve): /usr/lib/dovecot/managesieve mail_plugins(default): quota imap_quota expire mail_plugins(imap): quota imap_quota expire mail_plugins(pop3): quota expire mail_plugins(managesieve): mail_plugin_dir(default): /usr/lib/dovecot/modules/imap mail_plugin_dir(imap): /usr/lib/dovecot/modules/imap mail_plugin_dir(pop3): /usr/lib/dovecot/modules/pop3 mail_plugin_dir(managesieve): /usr/lib/dovecot/modules/managesieve lda: mail_plugins: quota sieve expire postmaster_address: postmaster(a)xyloid.org sendmail_path: /usr/lib/sendmail auth_socket_path: /var/run/dovecot/auth-master sieve_global_path: /etc/dovecot/sieve.global auth default: user: nobody passdb: driver: sql args: /etc/dovecot/dovecot-sql.conf userdb: driver: sql args: /etc/dovecot/dovecot-sql.conf socket: type: listen client: path: /var/spool/postfix/private/auth mode: 432 user: postfix group: postfix master: path: /var/run/dovecot/auth-master mode: 432 user: vmail group: mail plugin: quota: dict:user::proxy::quotadict quota_rule: *:storage=1G quota_rule2: Trash:storage=10%% quota_rule3: Junk:ignore sieve: /var/vmail/%d/%n/dovecot.sieve expire: Trash 7 Trash/* 7 Junk 30 expire_dict: proxy::expire dict: quotadict: mysql:/etc/dovecot/dovecot-dict-quota.conf expire: mysql:/etc/dovecot/dovecot-dict-expire.conf And dovecot-dict-expire.conf: connect = host=localhost dbname=mail user=mail password=[password] map { pattern = shared/expire/$user/$mailbox table = expires value_field = expire_stamp fields { username = $user mailbox = $mailbox } } Thanks. - Alex Reinhart

2 2

[Dovecot] sieve-test ignores precedence and other for vacation
by Burckhard Schmidt 20 Apr '11

20 Apr '11

Hello, I'm using pigeonhole-0.2.3, it works fine. I tried to verify some build in checks of vacation concerning precedence or auto-submitted header lines. It looks like sieve-test ignores header lines like Auto-Submitted: auto-submitted Precedence: bulk Return-Path: MAILER-DAEMON@... if I have a vacation script like require ["vacation", "regex", "fileinto", "variables", "envelope"]; if allof (header :regex ["to"] ["address1", "address2"], header :matches "to" "*") { vacation :days 7 : ... "text"; } The email to test contains a line like Precedence: bulk or Return-Path: <MAILER-DAEMON@...> or Auto-Submitted: yes sieve-test -D -C script.sieve mail I get Performed actions: * send vacation message: => seconds : 604800 ... text of vacation message END MESSAGE Implicit keep: * store message in folder: INBOX sieve-test(root): Info: final result: success -- Regards --- Burckhard Schmidt

2 1

[Dovecot] sieve logging
by Carsten Laun-De Lellis 20 Apr '11

20 Apr '11

Hi all I have installed dovecot 2.0.8 and sieve. Everything works fine, but I wonder how can I log sieve actions? Any help? Mit freundlichem Gruß Carsten Laun-De Lellis Dipl.-Ing. Elektrotechnik Certified Information Systems Auditor (CISA) Hauptstrasse 13 D-67705 Trippstadt Phone: +49 (6306) 992140 Mobile: +49 (151) 27530865 email: <mailto:carsten.delellis@delellis.net> carsten.delellis(a)delellis.net

2 1

[Dovecot] Question about Outlook workarounds
by Michael Durket 19 Apr '11

19 Apr '11

Do the Outlook workarounds in Dovecot 1.x still apply to current versions of Outlook? Or has Microsoft fixed Outlook such that workarounds are no longer needed in Dovecot (and if so, after what version of Outlook can I remove the workaround settings in my dovecot.conf file)?

6 5

[Dovecot] Outlook double attachments
by Ladislav Pašek 19 Apr '11

19 Apr '11

Hello, does anybody have experience with delivery of double attachments into Outlook? I suppose that this would be a problem of Outlook 2010 but to be sure I'd like to ask here as we are using Dovecot 2.0.11 Best Regards Ladislav

2 1

Re: [Dovecot] Solaris hardware crypto engines
by Chris Hoogendyk 19 Apr '11

19 Apr '11

The attachment made the message too long for the forum. So I canceled the original. If you want the Sun article, you can google 05071387.pdf. The original went through to Martin. On 4/18/11 5:02 PM, Chris Hoogendyk wrote: > On 4/18/11 6:30 AM, Martin Preen wrote: >> Hello, >> I tried to find out how about to use the hardware crypto engines under >> Solaris (Sun Fire T2000). It seems, that its not just a compilation issue: >> >>> For operations that are to be offloaded, it is necessary to restrict use to subset >> > of OpenSSL functions (the EVP_ functions) and explicitly indicate the use of the PKCS11 >> > engine; something like the following works for bulk ciphers (the process for RSA is similar): >>> >>> ENGINE *e; >>> ENGINE_load_builtin_engines(); >>> e = ENGINE_by_id("pkcs11"); >>> ENGINE_set_default_ciphers(e); >>> EVP_CIPHER_CTX_init (&ctx); >>> EVP_EncryptInit (&ctx, EVP_des_cbc (), key, iv); >>> EVP_EncryptUpdate (.....); >> >> Since I'm not familiar with SSL programming, I wonder if it's possible to >> modify the dovecot 1.2.x source code. Maybe someone has already tried this >> or can give me hint. > > You don't want to do that. > > The objective is for the ssl engine to be low level and basically invisible to layers of > application above that use it. So, build openssl properly and it just works for everything else. > > I'm not at work today, so I don't have reference to my notes. Ah, just found an email I sent to my > colleagues -- copied it below. -- --------------- Chris Hoogendyk - O__ ---- Systems Administrator c/ /'_ --- Biology& Geology Departments (*) \(*) -- 140 Morrill Science Center ~~~~~~~~~~ - University of Massachusetts, Amherst <hoogendyk(a)bio.umass.edu> --------------- Erdös 4 -------- Original Message -------- Subject: building apps with ssl Date: Thu, 18 Mar 2010 11:30:42 -0400 From: Chris Hoogendyk<hoogendyk(a)bio.umass.edu> cool. I've got things working now so that I can build applications with the openssl built by Sun to tap into the cryptographic accelerators on the T2. It turns out that the /opt/coolstack stuff related to openssl is actually symlinks to /usr/sfw. So, in the Makefile, if I set the openssl directory to /usr/sfw, rather than the default /usr/local/openssl, it finds the includes, the libraries, and the binaries that are needed. Then I specifically point the certs directory (in the Makefile) to /usr/local/openssl/certs for local control. Then, for the example of uw-imap, everything builds just fine. This should simply apply to anything we need to build with openssl on any of the T5220s. One side note is that this was a custom Makefile. Applications that use ./configure would presumably find the openssl through the paths provided in /usr/local/src/env. In the process of figuring that out, I also found that one of the most recent Solaris 10 kernel patches that we have on imladris specifically addresses awareness of the crypto accelerators and included patches to the openssl related items in /usr/sfw. So, I'm pretty confident that everything is alright. It's interesting too that in the paper presented at ACM by Sun last year they said that a couple of classes of attacks on crypto processes were negated by running it directly on the CPU. One of them had to do with timing. I should forward that paper to you as well. OK, just attached it. -- --------------- Chris Hoogendyk - O__ ---- Systems Administrator c/ /'_ --- Biology& Geology Departments (*) \(*) -- 140 Morrill Science Center ~~~~~~~~~~ - University of Massachusetts, Amherst <hoogendyk(a)bio.umass.edu> --------------- Erdös 4

1 0

[Dovecot] Solaris hardware crypto engines
by Martin Preen 18 Apr '11

18 Apr '11

Hello, I tried to find out how about to use the hardware crypto engines under Solaris (Sun Fire T2000). It seems, that its not just a compilation issue: > For operations that are to be offloaded, it is necessary to restrict use to subset > of OpenSSL functions (the EVP_ functions) and explicitly indicate the use of the PKCS11 > engine; something like the following works for bulk ciphers (the process for RSA is similar): > > ENGINE *e; > ENGINE_load_builtin_engines(); > e = ENGINE_by_id("pkcs11"); > ENGINE_set_default_ciphers(e); > EVP_CIPHER_CTX_init (&ctx); > EVP_EncryptInit (&ctx, EVP_des_cbc (), key, iv); > EVP_EncryptUpdate (.....); Since I'm not familiar with SSL programming, I wonder if it's possible to modify the dovecot 1.2.x source code. Maybe someone has already tried this or can give me hint. Regards. Martin

2 1

Re: [Dovecot] 2.0.12 ipwd.c does not compile under solaris
by Joseph Tam 18 Apr '11

18 Apr '11

> From: Roger Fujii <rmf(a)lookhere.com> > > the new src/lib/ipwd.c introduced in 2.0.12 no longer compiles on > solaris. Under solaris, > the *_r routines doesn't pass in result as the last parameter and return > errno - it just returns result. > obviously a job for configure..... This was solved several Emails back about forcing the use of POSIX threading models (a workaround path is provided): http://www.dovecot.org/list/dovecot/2011-April/058602.html Joseph Tam <jtam.home(a)gmail.com>

1 0