- dovecot - dovecot.org

[Dovecot] log_path and info_log_path - ownership of parent directory
by Phil Howard 07 May '10

07 May '10

The Ubuntu package didn't create the directory for these. What ownership/permissions are needed for it? Does it need to be owned by the dovecot user?

2 1

[Dovecot] disable plaintext auth ... only for some addresses
by Phil Howard 06 May '10

06 May '10

I'd like to disable plaintext authentication (e.g. only allow authentication that does STARTTLS or connects on SSL/TLS only ports) only for certain (most) IP addresses. I want to exempt a few addresses (users coming over known VPNs). Fortunately, all this is coming in over a firewall (Sonicwall) in which I can NAT traffic by IP address to go to specific port numbers. So, if I can establish a different disable_plaintext_auth policy by port number (for extra port numbers I'll choose later), that would let me accomplish this. If I cannot do this, then my only alternative is making the SSL/TLS only ports the only ones open to the internet, and use the non-SSL/TLS ports only for the VPNs (with disable_plaintext_auth = no). But I read somewhere that this is discouraged. What say ye?

1 0

[Dovecot] adding users to passwd-file
by Phil Howard 06 May '10

06 May '10

Is there a tool equivalent to the system "passwd" command (or maybe "adduser" or "useradd") that can support a passwd-file by setting a password, encrypting it with the salted MD5 scheme? The system "passwd" command doesn't have an option to "do it to this alternate file instead of /etc/shadow".

4 5

[Dovecot] dovecot and pigeonhole
by Vasiliy G Tolstov 06 May '10

06 May '10

Hello. Today i'm compile dovecot and pigeonhole from hg repo. After start dovecot i get errors in log like this: May 6 12:04:15 selfip dovecot: imap: Error: dlopen(/usr/lib/dovecot/lib90_sieve_plugin.so) failed: /usr/lib/dovecot/libdovecot-sieve.so.0: undefined symbol: mail_deliver_get_log_var_expand_table Google sasys, that this functions added in dovecto 2010-04-26, why a get this error when load sieve plugin? Thnak You. -- Vasiliy G Tolstov <v.tolstov(a)selfip.ru> Selfip.Ru

2 2

[Dovecot] !include
by Andre Hübner 06 May '10

06 May '10

dovecot 1.2.11 Hello, i do some tests with: !include /etc/dovecot/conf.d/*.conf dovecot fails to start if /etc/dovecot/conf.d/ is empty. Is this a expected behaviour? !include_try works without error but seems to me to be more usual when explicit, complete filename is given. is a little bit strange in contrast to behaviour of other servertypes which support a include-directive. Thanks, Andre

1 0

Re: [Dovecot] Dovecot Won't start on SLITAZ Linux
by Ludovic LEMARINEL 06 May '10

06 May '10

2010/5/6 Ludovic LEMARINEL <l.lemarinel(a)gmail.com> > 2010/5/6 Tom Hendrikx <tom(a)whyscream.net> > > On 06/05/10 13:40, Ludovic LEMARINEL wrote: >> > 2010/5/6 Charles Marcus <CMarcus(a)media-brokers.com> >> > >> >> On 2010-05-06 4:51 AM, Ludovic LEMARINEL wrote: >> >>> I've just compiled dovecot under Slitaz, but I can't get it to >> >>> run... >> >>> >> >>> It always die with this message in log : >> >>> >> >>> 2010-05-06 12:35:12 dovecot: Info: Dovecot v1.2.11 starting up >> >>> 2010-05-06 12:35:12 dovecot: Fatal: setrlimit(RLIMIT_DATA, 256): >> >>> Operation not permitted >> >> >> >> Google is your friend. Simply highlighting the above error: >> >> >> >> Fatal: setrlimit(RLIMIT_DATA, 256): Operation not permitted >> >> >> >> Provides the answer: >> >> >> >> http://tinyurl.com/25rmjod >> >> >> >> > >> > Thanks for your answer, but no, this page does not provide the answer. >> > The error message this guy gets is not the same as me >> >> Your original post says it does. If you receive another error message, >> please add it to your problem report. Also, please try to start dovecot >> from the commandline (as suggested in the url from Charles), and tell us >> about its behaviour, output and logging. >> >> > , and I don't have any >> > "/etc/security/limits.conf" file, so I can't change this setting . >> >> Dovecot tells you only what setting to to adjust. Where you put your >> settings is largely dependant on your OS or distro implementation. >> >> >> Ow, and please do not top-post. >> >> Regards, >> Tom >> > > > I've tried to run dovecot from the command line : it says what was the last > error (see my first post), and then nothing more. > When I do "ps", I can't see any process running, and the log says dovecot > finished... > > No more message than this :( > I'm going to try with 2.0 beta, and if no luck, I'll try to go down to 1.1. > Some news from my problem : I've tried 2.0 beta, 1.1, and both didn't worked. I went down to v1.0.15, and it works... Still haven't found what's wrong with newer versions...

1 0

[Dovecot] 2.0 beta4 (5d76f5b13883): Home directory not recognized
by Thomas Leuxner 06 May '10

06 May '10

Hi, latest Mercurial seems to have problems with the Config Module unloads: May 5 19:18:16 spectre dovecot: master: Dovecot v2.0.beta4 (5d76f5b13883) starting up May 5 19:18:34 spectre dovecot: managesieve: Error: fd_read() partial input (54/88) May 5 19:18:34 spectre dovecot: imap-login: Fatal: Error reading configuration: read(/var/run/dovecot/config) failed: EOF May 5 19:18:34 spectre dovecot: master: Error: service(imap-login): command startup failed, throttling Restarting IMAP/POP3 mail server: dovecotError: user root: Initialization failed: userdb didn't return a home directory, but plugin setting quota used it (%h): dict:user::file:%h/mdbox/dovecot-quota Fatal: Invalid user settings. Refer to server log for more information. $ tail /etc/dovecot/conf.d/90-plugin.conf quota = dict:user::file:%h/mdbox/dovecot-quota quota_rule = *:storage=1GB quota_rule2 = Trash:storage=10%% acl = vfile:/var/vmail/conf.d/%d/acls:cache_secs=300 } Regards Thomas

2 2

[Dovecot] migrating maildir from 1.1.7 to 1.2.11
by Romer Ventura 05 May '10

05 May '10

Hello, I have a SLES server running dovecot 1.1.7 and postfix which happens to be the production server. We are dumping all of the paid OSes so i have configured a Debian Lenny server to run the new mail server. The debian server has everything installed and it seems to be working fine. However, i need to get all the mailboxes from the SLES server, will there be any problem if i just copy the "vmail" folder to the new server...? Thanks ------ Romer Ventura

2 1

[Dovecot] Expire plugin
by Thomas M Goerger 05 May '10

05 May '10

Hi, I'm testing out the dovecot expire plugin, and not seeing anything go into the database I've got set up. It's a mysql db. I've gone in and verified that the db is set up correctly. The socket is set up so that deliver should be able to talk to the db, but I'm not seeing anything in the expires table when I look. Is there anywhere that expire can log to so I can see what might be happening? It doesn't look like it logs to the dovecot log. Thanks! ************************************************************************* * Tom Goerger - Email/Unix System Administrator * * * * University of Minnesota Email: tmg(a)umn.edu * * Operations, Infrastructure and Architecture Phone: 4-5804 * * Internet Services Office: 626J WBOB * * * *************************************************************************

2 6

[Dovecot] Plugin to handle message saved and handle move from tmp to cur/new
by Alex Baule 05 May '10

05 May '10

Hello everyone.... I need to finish my plugin, but i depends from some stuffs in the save/move process that i can't find how its work. In new zlib plugin (2.0) , the message is saved compressed now, but i dont find how is done the process to write the file to tmp, and move it to the cur or new directory (inside .Sent). I need to hook this function, to get access to some variables, like Flags and filename. There is a way to do this ? Tks.

1 0

[Dovecot] [Bug] Case sensitive usernames and variables
by Неворотин Вадим 05 May '10

05 May '10

Today I've found a very irritating bug: I use mail_location = maildir:/var/mail/%u/ in my dovecot.conf. And if user type Username, username, USERname etc at login dovecot create a different Maildir's for one real user. But LDA works corretly and save all messages to lowercase username mailbox. No problem, add L: mail_location = maildir:/var/mail/%Lu/ Then all works good. But I has also such namespace configuration: namespace shared { separator = / prefix = &BCAEMARBBEEESwQ7BDoEOA-/%%u/ location = maildir:/var/mail/%%u:INDEX=/var/mail/%u/shared/%%u subscriptions = no list = children } And here I can't replace %u with %Lu, because dovecot write to dovecot.log: IMAP(malamut): Error: Namespace '&BCAEMARBBEEESwQ7BDoEOA-/%Lu/': shared: Shared namespace prefix contains unknown variables So user has only one mailbox (with lovercase name), but a lot of different indexes for shared mailboxes, eg /var/mail/username/shared, /var/mail/USERname/shared etc. How can I tune dovecot to use always lowercase username? Dovecot 1.2.10

2 2

[Dovecot] 2.0 FTBFS since sqlpool changes when --with-sql=plugin is used
by Michal Hlavinka 05 May '10

05 May '10

Hi, yesterday lib-sql got driver-sqlpool with some changes in Makefile.am, but these changes in makefile works only for --with-sql=yes. When --with-sql=plugin is used build fails with: libtool --tag=CC --mode=link gcc -std=gnu99 -O2 -g -pipe -Wall -Wp,- D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 - m64 -mtune=generic -Wall -W -Wmissing-prototypes -Wmissing-declarations - Wpointer-arith -Wchar-subscripts -Wformat=2 -Wbad-function-cast -Wstrict- aliasing=2 -o checkpassword-reply checkpassword-reply.o ../../src/lib- dovecot/libdovecot.la -lrt libtool: link: gcc -std=gnu99 -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 - fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic - Wall -W -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar- subscripts -Wformat=2 -Wbad-function-cast -Wstrict-aliasing=2 -o .libs/auth auth.o auth-cache.o auth-client-connection.o auth-master-connection.o mech- otp-skey-common.o mech-plain-common.o auth-penalty.o auth-request.o auth- request-handler.o auth-settings.o auth-stream.o auth-worker-client.o auth- worker-server.o db-checkpassword.o db-sql.o db-passwd-file.o main.o mech.o mech-anonymous.o mech-plain.o mech-login.o mech-cram-md5.o mech-digest-md5.o mech-external.o mech-gssapi.o mech-ntlm.o mech-otp.o mech-skey.o mech-rpa.o mech-apop.o mech-winbind.o passdb.o passdb-blocking.o passdb-bsdauth.o passdb- cache.o passdb-checkpassword.o passdb-passwd.o passdb-passwd-file.o passdb- pam.o passdb-shadow.o passdb-sia.o passdb-vpopmail.o passdb-sql.o userdb.o userdb-blocking.o userdb-checkpassword.o userdb-nss.o userdb-passwd.o userdb- passwd-file.o userdb-prefetch.o userdb-static.o userdb-vpopmail.o userdb-sql.o db-ldap.o passdb-ldap.o userdb-ldap.o -Wl,--export-dynamic libpassword.a ../lib-ntlm/libntlm.a ../lib-otp/libotp.a ../../src/lib-sql/.libs/libsql.a ../../src/lib-dovecot/.libs/libdovecot.so -lcrypt -lpam -ldl -lrt -Wl,-rpath - Wl,/usr/lib64/dovecot libtool: link: gcc -std=gnu99 -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 - fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic - Wall -W -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar- subscripts -Wformat=2 -Wbad-function-cast -Wstrict-aliasing=2 -o .libs/checkpassword-reply checkpassword-reply.o ../../src/lib- dovecot/.libs/libdovecot.so -ldl -lrt -Wl,-rpath -Wl,/usr/lib64/dovecot ../../src/lib-sql/.libs/libsql.a(sql-api.o): In function `sql_init': /home/mhlavink/cvsf/dovecot/devel/dovecot-2.0.beta4/src/lib-sql/sql-api.c:70: undefined reference to `driver_sqlpool_init' Regards, Michal Hlavinka

1 1

[Dovecot] Access filesystem (NFS) as some user
by Mihamina Rakotomandimby 05 May '10

05 May '10

Manao ahoana, Hello, Bonjour, My Dovecot server has to access the users Maildirs. Maildirs are on a NFS mount that only a user (vmail/500) may read and write. No one else. How to configure Dovecot so that at least disk read-write are performed as a "vmail" user (that I have created with the right attributes)? Is the workaround of changing dovecot's system UID/GID a not-so-bad one? (I am trying to see if it works, at least...) Misaotra, Thanks, Merci. -- Architecte Informatique chez Blueline/Gulfsat: Administration Systeme, Recherche & Developpement +261 34 29 155 34 / +261 33 11 207 36

2 3

Re: [Dovecot] Dovecot deliver LDA problem
by Noel Butler 05 May '10

05 May '10

bugger sent that direct, please send that output to the list :) On Wed, 2010-05-05 at 09:04 +1000, Noel Butler wrote: > Alex, > > Please show the output of dovecot -n > > > On Tue, 2010-05-04 at 15:23 -0700, Alex wrote: > > > Thanks for the hint. > > > > I enabled 'socket listen' it in dovecot.conf as you pointed it out. > > > > Now dovecot.err log has these error messages: > > > > May 04 14:57:19 deliver(user1(a)some_domain.com): Error: User > > user1(a)some_domain.com is missing UID (set mail_uid) > > May 04 14:57:19 deliver(user2(a)some_domain.com): Error: User > > user2(a)some_domain.com is missing UID (set mail_uid) > > May 04 14:57:25 deliver(user3(a)some_domain.com): Error: User > > user3(a)some_domain.com is missing UID (set mail_uid) > > > > > > I tried to set necessary mail_uid in dovecot-sql.conf in user_query: > > > > user_query = SELECT maildir as mail, 51 as userdb_uid, 51 as > > userdb_gid, > > 51 as mail_uid FROM dovecot_email_maps WHERE email = '%u' > > > >

2 2

[Dovecot] Dovecot deliver LDA problem
by Alex 05 May '10

05 May '10

Hello folks, I was trying to set up Dovecot deliver LDA instead of Postfix default virtual LDA In Postfix main.cf I specified: virtual_transport = dovecot dovecot_destination_recipient_limit = 1 In Postfix master.cf I entered: dovecot unix - n n - - pipe user=mailer flags=DRhu argv=/usr/local/mail/libexec/dovecot/deliver -f ${sender} -d ${recipient} In Dovecot dovecot.conf I have the following entries: passdb sql { args = /etc/dovecot/dovecot-sql.conf } userdb prefetch { } userdb sql { args = /etc/dovecot/dovecot-sql.conf } In Dovecot dovecot-sql.conf I have this: user_query = SELECT maildir as mail FROM dovecot_email_maps WHERE email = '%u' password_query = SELECT user, password, userdb_uid, userdb_gid, userdb_home, userdb_mail FROM dovecot_user_maps WHERE user = '%n' After reloading postfix and restarting dovecot I'm getting this type of warnings in dovecot.err log: May 04 10:43:10 deliver(some_address(a)some_domain.com): Error: userdb lookup: connect(/var/run/dovecot//auth-master) failed: No such file or directory Could somebody please help me to resolve this mistery. What am I missing ether in configuration file or somewhere else? Many thanks in advance, Alex

2 2

[Dovecot] CDB Authentication database
by Luca Corti 04 May '10

04 May '10

Hello, I'd be interested in seeing a CDB based backend for the userdb/passdb lookups. Is anyone working on this? Is there any documentation (other than the wiki Design section and the code itself) on implementing database backends? thanks Luca

5 5

[Dovecot] Dovecot antispam plugin with virtual users
by Philipp 04 May '10

04 May '10

Hi, I'm running dovecot 1.2.11 with the antispam plugin 1.2+20090702-1+b1 installed from the debian repos (squeeze) and dspam 3.9.0+git20100416-1. Unfortunately this doesn't seem to work well with my virtual users, stored on a local openldap server: While sending mail manually to dspam does work, there's some trouble when trying to use the antispam plugin. I tried using both sendmail (exim4 in my case) and configuring antispam_mail_sendmail = /usr/bin/dspam to point to dspam, with similar results: exim: [15042] Failed to get user name for uid 5001 dspam: dspam[18961]: Unable to determine the runtime user both errors are reproducible by running exim/dspam with sudo -u \#5001 with 5001 being the virtual uid used of the user that is logged in through imap. Of course the uid 5001 is invalid for the rest of the system, so it can't run anything, but as far as I can tell the plugin uses this uid to call the respective programs :\ Now I was wondering if it was possible to keep the plugin from using the imap-uid and instead specifying one elsewhere? I guess this would be of fairly great interest for everyone using virtual users, be it ldap or anything else. Or is there a different way to do this? Do I maybe just miss setting a variable somewhere? Philipp

2 1

[Dovecot] dovecot user
by Timo Sirainen 04 May '10

04 May '10

Any thoughts on this: The primary use for "dovecot" user has been for login processes. But people keep misunderstanding this and try to use dovecot for accessing mails. For years I've been wondering about renaming this user to something else like dovelogin, but it never really seemed practical. So now with v2.0 there are a bunch of new processes, and for example anvil and dict are now run as dovecot user by default. But it's not really good that login processes can just go and kill those processes. And even worse, if drop_priv_before_exec=yes they could ptrace these processes. So I think we need two Dovecot users for v2.0: 1. Completely untrusted user for login processes. 2. Slightly more trusted internal Dovecot user. So "dovecot" could be reused for 2. And it would no longer be a mortal sin to use dovecot user for owning mail files. For 1. there would be a new user. I'd use "dovelogin", but apparently tools still don't much like usernames that are longer than 8 characters. Like ps could show numeric uid instead of 9 character long username. So .. any suggestions? "dovlogin" could be one possibility I guess. It would be nice if the name somehow reminded of login processes, but maybe something else could be used too, like: dovenil, dovenull, dovezero, dovenone, dovevoid, doveint, dovedown, dovein, dove0

17 33

[Dovecot] Dovecot SQL user_query
by Alex 04 May '10

04 May '10

Hello folkz, Is it possible to setup in dovecot-sql.conf file to run user_query twice if its first run didn't yield any values. And likewise don't proceed with the second run if the first user_query execution produced the result. I have to query two tables which specify there email messages shall be delivered via Dovecot deliver LDA depending on the incoming email addresses. Please don't question me about this setup. I inherited this mess from the previous Network Admin. First query should be ran like this: user_query = SELECT concat('maildir:/data/mail/',maildir) as mail FROM folders WHERE email = '%u' If the first query doesn't yield any result the second shall be ran like this: user_query = SELECT maildir as home FROM users WHERE email = '%u' Please advise if such user_query setup is possible at all. Thanks in advance, Alex

5 4

[Dovecot] exit codes for 'sievec'
by Jerry 04 May '10

04 May '10

As far as I can tell, 'sievec' always exits with '0' whether there is an error or not. If that is correct, I believe it should be modified to exit with some other code if the conversion and writing of the new file are not completed successfully. I have written a few script files and a 'make' file specifically for sievec. The inability to get an accurate exit code is problematic for me. -- Jerry Dovecot.user(a)seibercom.net Disclaimer: off-list followups get on-list replies or get ignored. Please do not ignore the Reply-To header. __________________________________________________________________ Fortune's Fictitious Country Song Title of the Week: "How Can I Miss You if You Won't Go Away?"

2 2

[Dovecot] zlib Plugin Dovecot 2.0 - ostream
by Alex Baule 03 May '10

03 May '10

Hello Everyone... In zlib save mail have it: struct ostream *output; There is a way to get the file descriptor from output used in ostream zlib plugin ? to get in istream is i_stream_get_fd, but don't have o_stream_get_fd......

2 6

[Dovecot] Dovecot Vs Procmail LDA
by Wilberth Pérez 03 May '10

03 May '10

Hi everyone: Anyone knows what are the differences (or advantages and disadvantages) between Dovecot LDA with postfix Vs. Procmail LDA with postfix ? -- -------------------------------------------------------- LCC Wilberth de Jesús Pérez Segura CCSA- Administración de Servicios y Seguridad de las TI Correo: wilberth.perez(a)uady.mx Universidad Autónoma de Yucatán Secretaría General Coordinación Administrativa de Tecnologías de Información RIUADY C-59 x Av. Itzáes (999)923-74-28 Ext. 1117 Mérida, Yucatán, México 2010 --------------------------------------------------------

1 0

[Dovecot] [SUGGESTION] Managesieve: Transparent home dir creation
by Felix Leimbach 03 May '10

03 May '10

Hi all, I'm using dovecot 2.0-beta4's sieve plugin and it works great but here is an idea for improvement: When creating the first sieve script for an account whose home dir does not yet exist then dovecot *does* create that homedir but the sieve client times out. Tried Thunderbird's sieve addon as well as horde's ingo. The log says: dovecot: managesieve-login: Login: user=<me(a)example.com>, method=PLAIN, rip=127.0.0.1, lip=127.0.0.1, secured dovecot: managesieve(me(a)example.com): sieve-storage: sieve_storage_verify_dir(/mnt/data/vmail/dovecot-homedirs/example.com/me//s… failed: File exists dovecot: managesieve(me(a)example.com): Fatal: Failed to open Sieve storage. Because the homedir was created (great!) when I retry the same action then all works perfectly well. So why not make that homedir creation transparent so that the clients' first try succeeds instead of the second try? Cheers, Felix

3 3

[Dovecot] Migrating domain with diferent default_pass_scheme
by emiliojoel 03 May '10

03 May '10

Hi I have now my main mail server using dovecot with diferents domains and for user auth using mysql with default_pass_scheme = PLAIN, Now I am migrating a new domain from other server with dovecot using mysql with default_pass_scheme = MD5. My question is: Is possible have a different default_pass_scheme depending of the domain? could I use dovecot with domains using mysql auth with default_pass_scheme = PLAIN and other with default_pass_scheme = MD5? any sugestion? thanks Emilio -- View this message in context: http://old.nabble.com/Migrating-domain-with-diferent-default_pass_scheme-tp… Sent from the Dovecot mailing list archive at Nabble.com.

2 1

[Dovecot] dovecot server down, dovecot deliver keeps mail?
by mimosinnet＠gmail.com 03 May '10

03 May '10

My dovecot server was down las friday, and it seems that mails have been lost. This is what I have got in my logs since the server has been down: May 3 03:35:34 generatech dovecot: deliver(mail(a)domain.org): userdb lookup: connect(/var/run/dovecot/auth-master) failed: No such file or directory I guess that because the server was down dovecot transport could not deliver the mail. Nevertheless, I am wondering if dovecot deliver agent keeps undelivered mail somewhere. Thanks! I have been using this howto for the configuration: http://en.gentoo-wiki.com/wiki/Mail_server_using_Postfix_and_Dovecot This is dovecot 1.2.6, and my configuration: # dovecot -n # 1.2.6: /etc/dovecot/dovecot.conf # OS: Linux 2.6.28-hardened-r9 x86_64 Gentoo Base System release 2.0.1 rootfs listen: *:143 ssl_listen: *:52003 ssl_cert_file: /etc/ssl/dovecot/dovecot.pem ssl_key_file: /etc/ssl/dovecot/dovecot.key login_dir: /var/run/dovecot/login login_executable: /usr/libexec/dovecot/imap-login first_valid_uid: 8 last_valid_uid: 8 first_valid_gid: 12 last_valid_gid: 12 mail_uid: 8 mail_gid: 12 mail_location: maildir:/var/mail/%d/%n/Maildir/:INDEX=/var/mail/%d/%n/indexes mail_plugins: quota imap_quota lda: postmaster_address: postmaster(a)generatech.org mail_plugins: quota sieve auth default: mechanisms: plain login user: nobody passdb: driver: sql args: /etc/dovecot/dovecot-sql.conf userdb: driver: prefetch 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: mail group: mail plugin: sieve: /var/mail/%d/%n/.dovecot.sieve sieve_dir: ~/var/mail/%d/%n/sieve -- Ningún Lugar Activisme Cultural per a la Transformació Social Cultural Activism for Social Transformation ningun_lugar(a)riereta.net http://ningunlugar.org/ http://generatech.ningunlugar.org/ http://video.ningunlugar.org

2 1

[Dovecot] courier-imap and dovecot in migration process
by No body ist Perfect 02 May '10

02 May '10

Hello! I plan a migration from courier-imap to dovecot. at now run courier-imap as imaps server only and with perdition as redirector. can i run courier-imap and dovecot at same time on the same server(listing on diferent ports of course), i mean with shared home directory tree (Mailbox dir) ? Maybe access conflict or some problems ? Thanks a lot

1 0

[Dovecot] What causes dovecot to rebuild dovecot-uidlist?
by Pat 02 May '10

02 May '10

Hello, In what situations will Dovecot 1.2.11 rebuild its uidlist? I am using the maildir storage format, and I am also modifying the uidlist manually (so it's possible I could jack* things up). I have found some related code in maildir-uidlist.c: maildir_uidlist_next. 1) Blank filename or UID 0 2) Any UID not being greater than the proceeding UID 3) Any UID being greater than ~4 billion. 4) Any UID being greater than the suggested next uid. 6) If a filename contains a / 7) General header line sanity checking. Are there any other sections of the code that do sanity checks on dovecot-uidlist (for Maildir type storage)? And optionally do you agree that maildir_uidlist_set_corrupted is the best place to watch to see if I have mangled the uidlist? Thanks in advance, Pat * I insert into dovecot-uidlist manually when I deliver new messages, or discover a message that is not in uidlist for some reason. For compatability reasons I maintain a completely separate cache. If there is a better way to keep the uidlist up to date when the user isn't necessarily accessing the account via dovecot imap, I'm open to suggestions. However my situation is fairly unique I think.

2 2

[Dovecot] My dovecot does not store Thunderbird flags
by Julien Buratto 01 May '10

01 May '10

Hi all, my Dovecot 1.1.20 does not store flags (I mean, Thunderbird Labels like ToDo, Important and so on..) When logging in via telnet, I get: . examine INBOX * FLAGS (\Answered \Flagged \Deleted \Seen \Draft NonJunk $MDNSent Junk $Forwarded NotJunk receipt-handled $Label7 $label1 $label2 $label3 $label4 $label5 expired_2w expired_2m delete rimandata forwarded) * OK [PERMANENTFLAGS ()] Read-only mailbox. As per Thunderbird mozillazine page about tags (http://kb.mozillazine.org/Tags#IMAP) that's what the issue is. PERMANENTFLAGS should allow /* while in my server it's empty. Could someone point me how can I change this behaviour or some hints ? Thank you -- Julien Buratto PS: In case someone asks for it > dovecot -n # 1.1.20: /usr/local/etc/dovecot.conf # OS: Linux 2.6.9-67.0.4.EL i686 CentOS release 4.7 (Final) log_path: /var/log/maillog info_log_path: /var/log/maillog protocols: imaps imap pop3 pop3s disable_plaintext_auth: no login_dir: /usr/local/var/run/dovecot/login login_executable(default): /usr/local/libexec/dovecot/imap-login login_executable(imap): /usr/local/libexec/dovecot/imap-login login_executable(pop3): /usr/local/libexec/dovecot/pop3-login mail_location: mbox:~/mail/:INBOX=mbox mbox_lazy_writes: no mail_executable(default): /usr/local/libexec/dovecot/imap mail_executable(imap): /usr/local/libexec/dovecot/imap mail_executable(pop3): /usr/local/libexec/dovecot/pop3 mail_plugin_dir(default): /usr/local/lib/dovecot/imap mail_plugin_dir(imap): /usr/local/lib/dovecot/imap mail_plugin_dir(pop3): /usr/local/lib/dovecot/pop3 pop3_no_flag_updates(default): no pop3_no_flag_updates(imap): no pop3_no_flag_updates(pop3): yes lda: postmaster_address: postmaster(a)example.com auth default: cache_size: 1024 verbose: yes passdb: driver: pam args: session=yes cache_key=%u%s dovecot userdb: driver: passwd plugin: acl: vfile:/etc/dovecot-acls

1 2

[Dovecot] UK Blackberry BIS and Dovecot - 2-way sync
by Lucio Godoy 01 May '10

01 May '10

Hi All; I have search the net and i cannot find any answers or fixes for this issue. I have a RHEL5 server running Dovecot Version 1.2.11 all worked well until i have diceded to subscribe to Vodafone's BIS service (BB Bold 9700). Internally I use outlook, and it does 2-way sync without any issues, but the blackberry only syncs new email and it does not sync deleted emails when I delete them in Outlook. Is there any special setting to allow such synchronisation? My server is secured by a pix firewall with ports 993, and 995 open for email sync (Secure IMAP and Secure Pop), i have setup the blackberry to use port 993. Many thanks for your help. Lucio =========================================================== The Home CCTV Security System made easy! http://www.kaduco.com/our-list-products-services/cctv-security-camera-syste… _________________________________________________________________ http://clk.atdmt.com/UKM/go/197222280/direct/01/ We want to hear all your funny, exciting and crazy Hotmail stories. Tell us now

1 0

[Dovecot] zlib Plugin Dovecot 2.0 - ostream
by Alex Baule 30 Apr '10

30 Apr '10

Hello Everyone. There is possible to rename the email saved by a plugin, like zlib ? i explain.... i need to add 1 line in email header, because of that, i want to rewrite the filename to update the S and W flag. Tks !

1 0

[Dovecot] 1.0.15 --> 1.2.11 --> got too little data
by Andre Hübner 30 Apr '10

30 Apr '10

Hello, did updates from 1.0.15 to 1.2.11 and is in most cases succesful. Sometimes dovecot leaves errormessage in Log: IMAP(username): FETCH [HEADER] for mailbox INBOX UID 5106 got too little data (copying): 516 vs 528 Until now only Inbox was effected and no other folders. I delete cachesfiles in ~mail/.imap to solve this but is there a way to do this automatical (self healing)? Dovecot could recreate complete caches if there was a difference in calculation. Thanks, Andre

2 6

[Dovecot] convert old configuration file with doveconf : Unknown setting
by de Lafond Guillaume 30 Apr '10

30 Apr '10

Hi, It happens when I try to convert this config (with the dovecot 2.0 beta4 release) : test-mail:/tmp# dovecot -n -c ./dovecot.old.conf # 1.2.11: ./dovecot.old.conf # OS: Linux 2.6.26-2-686-bigmem i686 Debian 5.0.4 log_path: /var/log/dovecot.log info_log_path: /var/log/dovecot_info.log protocols: imap listen: 127.0.0.1:30143 disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable: /usr/lib/dovecot/imap-login login_user: aquaray login_greeting: Mail - test v2 login_process_per_connection: no login_process_size: 96 login_processes_count: 20 login_max_processes_count: 196 max_mail_processes: 10024 mail_max_userip_connections: 10000 verbose_proctitle: yes first_valid_uid: 89 mail_location: maildir:~/:INDEX=/var/mail/dovecot_indexes/%1d/%d/%n fsync_disable: yes maildir_very_dirty_syncs: yes mbox_write_locks: fcntl dotlock mail_drop_priv_before_exec: yes mail_plugins: quota imap_quota imap_client_workarounds: outlook-idle delay-newmail namespace: type: private separator: / inbox: yes list: yes subscriptions: yes namespace: type: private separator: / prefix: INBOX/ hidden: yes list: no subscriptions: yes lda: postmaster_address: postmaster(a)test-dovecotv2.aquaray.com hostname: test-dovecotv2.aquaray.com log_path: /var/log/dovecot-deliver.log info_log_path: /var/log/dovecot-deliver.log auth_socket_path: /var/run/dovecot-socket-auth/auth-master auth default-auth-master: mechanisms: plain login apop cram-md5 master_user_separator: * passdb: driver: sql args: /etc/dovecot/dovecot-sql.conf userdb: driver: sql args: /etc/dovecot/dovecot-sql.conf socket: type: listen master: path: /var/run/dovecot-socket-auth/auth-master mode: 384 user: aquaray auth default: mechanisms: plain login apop cram-md5 cache_size: 2048 cache_ttl: 2000 master_user_separator: * count: 10 passdb: driver: sql args: /etc/dovecot/dovecot-sql-master.conf pass: yes master: yes passdb: driver: sql args: /etc/dovecot/dovecot-sql.conf userdb: driver: sql args: /etc/dovecot/dovecot-sql.conf plugin: quota: maildir:User quota test-mail:/tmp# /usr/local/dovecot2/bin/doveconf -n -c ./ dovecot.old.conf # 2.0.beta4 (6891d20929ca): ./dovecot.old.conf doveconf: Warning: Obsolete setting in ./dovecot.old.conf:26: login_dir has been removed doveconf: Warning: Obsolete setting in ./dovecot.old.conf:29: login_chroot has been replaced by service { chroot } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 29: Unknown setting: chroot doveconf: Warning: Obsolete setting in ./dovecot.old.conf:32: login_user has been replaced by service { user } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 32: Unknown setting: user doveconf: Warning: Obsolete setting in ./dovecot.old.conf:36: login_process_size has been replaced by service { vsz_limit } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 36: Unknown setting: vsz_limit doveconf: Warning: Obsolete setting in ./dovecot.old.conf:39: login_process_per_connection has been replaced by service { service_count } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 39: Unknown setting: service_count doveconf: Warning: Obsolete setting in ./dovecot.old.conf:42: login_processes_count has been replaced by service { process_min_avail } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 42: Unknown setting: process_min_avail doveconf: Warning: Obsolete setting in ./dovecot.old.conf:45: login_max_processes_count has been replaced by service { process_limit } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 45: Unknown setting: process_limit doveconf: Warning: Obsolete setting in ./dovecot.old.conf:48: login_max_connections has been replaced by service { client_limit } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 48: Unknown setting: client_limit doveconf: Warning: Obsolete setting in ./dovecot.old.conf:75: mail_drop_priv_before_exec has been replaced by service { drop_priv_before_exec } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 75: Unknown setting: drop_priv_before_exec doveconf: Warning: Obsolete setting in ./dovecot.old.conf:88: max_mail_processes has been replaced by service { process_limit } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 88: Unknown setting: process_limit doveconf: Warning: Obsolete setting in ./dovecot.old.conf:91: mail_process_size has been replaced by service { vsz_limit } doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 91: Unknown setting: vsz_limit It does not understand the old 1.1 setting "ssl_disable". test-mail:/tmp# /usr/local/dovecot2/bin/doveconf -n -c ./ dovecot.old.conf # 2.0.beta4 (6891d20929ca): ./dovecot.old.conf doveconf: Fatal: Error in configuration file ./dovecot.old.conf line 21: Unknown setting: ssl_disable ----------------------------------------------------------- After removing all fatal error (comment all settings that are not understood), it works. If needed, I could test against the another other version (latest hg ?). Regards, -- Guillaume de Lafond Aqua Ray SARL

2 3

[Dovecot] Quota maildirsize and dbox
by luben karavelov 30 Apr '10

30 Apr '10

Hello, Is there any reason that maildirsize Quota store should not work in dbox folders or any other folder based mailbox stores? I have patched a test version here and it seems to work (added "dbox" check in src/plugins/quota/quota-maildir.c line 785). If there could be any complications or unexpected consequences of this setup? Thanks in advance luben

3 2

[Dovecot] Expire plugin / expire-tool redesign for doveadm
by Timo Sirainen 30 Apr '10

30 Apr '10

Expire plugin / expire-tool seems annoyingly inflexible currently, so I was thinking about a more generic redesign: The expire plugin keeps track of the oldest message in configured mailboxes. Its only configuration is the list of tracked mailboxes. There's no configuration like what the expire timeouts are or anything. There will be a new doveadm command: doveadm expunge [-u user | -A] <mailbox> <search query> So when you want to expunge all mails from Trash older than 1 week for all users, you say: doveadm expunge -A Trash savedbefore 1w This command works even without expire plugin. To optimize it to avoid looking into all users' Trash mailbox, there's a new expire doveadm plugin, which can use the expire dict to filter out users who don't have anything to expunge. This also supports another feature that the plugin can optimize: doveadm move -A INBOX Archive/2009/INBOX since 2009-01-01 before 2010-01-01 And since people have had problems waiting for expire dict to fill, there could be also a new parameter that does the filling immediately. TODO: - dbox altmove feature will be removed from expire plugin. it may need another plugin, or some other way to configure expire plugin for it. it's anyway a special case. - When using multiple mailboxes it would be more optimal to handle all mailboxes for a user at once, rather than using separate doveadm commands. Maybe the command syntax needs some more thinking to support this. Different mailboxes could have different rules though..

3 6

[Dovecot] Released Pigeonhole Sieve v0.1.16 for Dovecot v1.2.11
by Stephan Bosch 30 Apr '10

30 Apr '10

Hello Dovecot users, Pigeonhole development has slowed down significantly in the last few months due to the fact that I got swamped in work. The plan is now to move active development of Pigeonhole to Dovecot v2.0, meaning that this will be the last release for v1.2 that contains significant development changes and new features. However, if a certain new feature is simple enough and there is a demand for it, I may backport it from v2.0 and release it along with any bugfixes that may still arise. The main thing that this new release adds is full support for the spamtest, spamtestplus and virustest extensions. These extensions implement a uniform means of testing the content spam/virus status headers in mail messages: http://ietfreport.isoc.org/idref/rfc5235/ Not many people have tested these new features yet and therefore these are currently considered experimental. Test these thoroughly before you start using them! Documentation is available in doc/spamtest-virustest.txt Changelog Sieve v0.1.16: * Finished implementation of spamtest, spamtestplus and virustest extensions. These are not enabled by default and need to be activated with the sieve_extensions setting. Documentation available in doc/spamtest-virustest.txt + Vacation extension: the from address of the generated reply is now by default equal to whatever known recipient alias matched the headers of the message. If it is one of the aliases specified with :addresses, it is used in stead of the envelope recipient address that was used before. + Restructured and optimized the lexical scanner. + Added --with-docs configure option to allow disabling installation of documentation. - Accidentally omitted 'extern' in two declarations of global variables in header files, causing compile failures on certain systems. - Deprecated imapflags extension: fixed implicit assignment of flags. Turns out this never really worked, but the effect of this bug was obscured by the removeflag bug fixed in the previous release. - Fixed various memset argument mixups in enotify extension. This caused warnings on certain systems, but luckily no adverse effects at runtime. The releases are available as follows: Sieve: http://www.rename-it.nl/dovecot/1.2/dovecot-1.2-sieve-0.1.16.tar.gz http://www.rename-it.nl/dovecot/1.2/dovecot-1.2-sieve-0.1.16.tar.gz.sig Refer to http://pigeonhole.dovecot.org for more information. Have fun testing the new releases and don't hesitate to notify me when there are problems. Regards, -- Stephan Bosch stephan(a)rename-it.nl

3 6

[Dovecot] Again: Occasional problem with POP3-Download
by Andre Tann 30 Apr '10

30 Apr '10

Hello everyone, in february I posted the mail below[1]. I can now reproduce the problem. When filling up a mailbox with postfix' smtp-source with dozens of identical mails, then it is never possible for me to fetch one single mail out of the mailbox. The process of fetching mail via POP3 always fails. It makes no difference which mail client I use. I have tested KMail, Outlook and Thunderbird. Can anyone confirm this? Is dovecot for some reason unable to deliver such identical mails? Can it be the case that this is a problem with too similar message ids? ~/Maildir/cur$ cat * | grep -i message-id Message-Id: <1024.0003.0008@alphamobil> Message-Id: <1024.0003.0009@alphamobil> Message-Id: <1024.0003.0006@alphamobil> Message-Id: <1024.0003.0007@alphamobil> Message-Id: <1024.0003.0005@alphamobil> Message-Id: <1024.0003.0004@alphamobil> Message-Id: <1024.0003.0003@alphamobil> Message-Id: <1024.0003.0002@alphamobil> Message-Id: <1024.0003.0001@alphamobil> Message-Id: <1024.0003.0000@alphamobil> Thanks and greetings from Munich! AT [1] Hello everyone, I have two mailservers [1] running dovecot. Mail store format is maildir. Every once in a while I have the following problem: When a User downloads his mail via POP3, the client says something like "the server has closed the connection unexpectedly, and gives this reason: [...blahblah...]". One can now cancel the download, and restart it, and the problem shows up again and again. The first few mails get downloaded, and then the process stops. Strange to say the "blahblah"-portion is a piece of text out of the email that doesn't want to get downloaded. Then I remove this single message out of the Maildir/cur-directory, and everything is fine again. It is easy to figure out which mail to delete because it's the first one that is not market as seen [...] filename...:2,S otherfilename...;2, <= this file must be deleted [...] For me it looks as if dovecot had a problem with some byte sequence in this mail, and for some reason it gives an error message to the client which is taken out of the mail. On the SuSE-Machine, before dovecot there was a courier-imapd running, which never showed this kind of problem, so it should not be a client problem. Hopefully my description is clear enough. Does anyone have an idea what's going on here? Thanks. Andre [1] One machine is a openSuSE 11.2 with # rpm -qa | grep dovecot dovecot12-backend-sqlite-1.2.9-0.1.1.i586 dovecot12-backend-mysql-1.2.9-0.1.1.i586 dovecot12-backend-pgsql-1.2.9-0.1.1.i586 dovecot12-1.2.9-0.1.1.i586 The other machine is a Debian Lenny with # dpkg -l | grep dovecot dovecot-common 1:1.0.15-2.3+lenny1 dovecot-imapd 1:1.0.15-2.3+lenny1 dovecot-pop3d 1:1.0.15-2.3+lenny1 -- Andre Tann

3 3

[Dovecot] Right way to Dovecot Load Balance
by Proskurin Kirill 30 Apr '10

30 Apr '10

Hello. I building LB&HA Email cluster based on Exim+Dovecot+LDAP. Now I thinking what the most right way to LB Dovecot(2 nodes with shared storage). HA-Proxy? Dovecot Proxy? Nginx? Else? -- Best regards, Proskurin Kirill

2 1

[Dovecot] Question about auth multiple configuration
by Fabrice MATHIEU 30 Apr '10

30 Apr '10

Hello, My mail system is build with postfix, dovecot and roundcube. In first time users can view and "manage" their mail only on the webmail. So this one (webmail) use IMAP (no tls/ssl at all) authentication to give access to users maildir. This connection is made on the 'loopback' interface and use PLAIN method. This works fine (configuration below without ssl parameters). Now I want to see and send my mail with a MUA (thunderbird). As my system is on an Internet provider I want to add more security "solutions". I use smtps with postfix and SASL/auth by dovecot socket mechanism to send mail. I use imaps with dovecot for managing my maildir. I have added ssl parameters to the configuration file (see below) and I required certificate from the client (to avoid man in the middle attack and to get access only for client with user certificate). To get this "ssl_require_client_cert" is set to yes. Ok, now imaps works perfectly. But since a certificate is required webmail authentication (localhost) and SASL (postfix auth trough dovecot socket mechanism) don't work. Webmail => dovecot: imap-login: Disconnected (cert required, client didn't start TLS): method=PLAIN, rip=127.0.0.1, lip=127.0.0.1, secured Postfix => postfix/smtpd[71640]: warning: TOTO[X.A.B.C]: SASL PLAIN authentication failed: Client didn't present valid SSL certificate That's normal. But 127.0.0.1 client(network) is considered by dovecot as secure, so won't the auth possible without certificate ? Can't we make two auth policy to make secure (client crt require) for public IP/client and less "secure" (without crt client) for local process (postfix) and local newtwork(127.0.0.1) for roundcube ? I see section "auth default { ... }" and is used by ... default ! But can we make an other one to make this two particular authentication on the same "instance" ? Thank you. And then my configuration : operating system => FreeBSD 7.1 dovecot --version => 1.2.8 dovecot -n => # 1.2.8: /usr/local/etc/dovecot.conf # OS: FreeBSD 7.1-RELEASE i386 ufs protocols: imap listen: 127.0.0.1 ssl_listen: X.Y.Z.T ssl: yes ssl_ca_file: /usr/local/etc/dovecot/CAclient.pem ssl_cert_file: /usr/local/etc/dovecot/imaps.pem ssl_key_file: /usr/local/etc/dovecot/imaps.key ssl_verify_client_cert: yes login_dir: /var/run/dovecot/login login_executable: /usr/local/libexec/dovecot/imap-login verbose_proctitle: yes first_valid_uid: 1000 first_valid_gid: 1000 mail_privileged_group: mail mail_location: maildir:/jails/mails/%d/%n imap_client_workarounds: delay-newmail netscape-eoh tb-extra-mailbox-sep lda: postmaster_address: postmaster@SOMETHING hostname: SOMETHING_ELSE sendmail_path: /usr/sbin/sendmail auth default: mechanisms: plain login username_format: %Lu passdb: driver: passwd-file args: username_format=%n /jails/mails/dov_pass/%d/usr_pas userdb: driver: passwd-file args: username_format=%n /jails/mails/dov_pass/%d/usr_pas ssl_require_client_cert = yes ssl_username_from_cert = no socket: type: listen client: path: /var/spool/postfix/private/auth mode: 432 user: postfix group: wheel master: path: /var/run/dovecot/auth-master mode: 384 Best regards, _________________________________________________________________ Consultez gratuitement vos emails Orange, Gmail, Free, ... directement dans HOTMAIL ! http://www.windowslive.fr/hotmail/agregation/

3 8

[Dovecot] Question about dovecot imap proxy
by Monika Janek 29 Apr '10

29 Apr '10

Hi everyone! I have dovecot (1.2.11) on one our external mail servers acting as a proxy. The client (ifor now, my iphone) connects fine via ssl to the external mailserver but I can't seem to get a secure connection now to the internal destination imap server (between external mail server and internal imap server, it's going through port 143). Running tcpdump, I can clearly see my password being transmitted on our internal network. I read the addition to the documentation about dovecot proxy but I'm not sure where to add these variables: In v1.2.rc4+ the connections to destination server can be TLS/SSL encrypted by returning: * ssl=yes: Use SSL and require a valid verified remote certificate. *WARNING: Unless used carefully, this is an insecure setting!* Currently host must be an IP address, so this setting accepts any certificate signed by a trusted CA. The host name isn't checked in any way against the certificate's CN. The only way to use this securely is to only use and allow your own private CA's certs, anything else is exploitable by a man-in-the-middle attack. * ssl=any-cert: Use SSL, but don't require a valid remote certificate. * starttls: Use STARTTLS command instead of doing SSL handshake immediately after connected. Can anyone point me in the right direction? Thanks so much! :) -- Monika Janek Systems Administrator, Side Effects Software Toronto, Ontario Canada 416-504-9876 x207 www.sidefx.com

2 8

[Dovecot] Assertion failed in mail-search-build.c
by Mark Moseley 29 Apr '10

29 Apr '10

With Friday's and today's mercurial repo, I'm getting this: Apr 26 12:55:31 master: Info: Dovecot v2.0.beta4 starting up (core dumps disabled) Apr 26 12:58:34 imap-login: Info: Login: user=<mail@box>, method=PLAIN, rip=10.1.1.206, lip=192.168.152.1 Apr 26 12:58:34 imap(mail@box): Panic: file mail-search-build.c: line 59 (mail_search_build_key_int): assertion failed: (sarg->value.subargs != NULL) Apr 26 12:58:34 imap(mail@box): Error: Raw backtrace: /usr/lib/dovecot/libdovecot.so.0 [0xb76a3ad1] -> /usr/lib/dovecot/libdovecot.so.0 [0xb76a3b4f] -> /usr/lib/dovecot/libdovecot.so.0(i_error+0) [0xb76a3e08] -> /usr/lib/dovecot/libdovecot-storage.so.0 [0xb76f859e] -> /usr/lib/dovecot/libdovecot-storage.so.0 [0xb76f860f] -> /usr/lib/dovecot/libdovecot-storage.so.0(mail_search_build+0x82) [0xb76f86f2] -> dovecot/imap [mail@box 10.1.1.206 SEARCH](imap_search_args_build+0x48) [0x805d8a8] -> dovecot/imap [mail@box 10.1.1.206 SEARCH](cmd_search+0xbb) [0x805342b] -> dovecot/imap [mail@box 10.1.1.206 SEARCH] [0x80564dc] -> dovecot/imap [mail@box 10.1.1.206 SEARCH] [0x8056579] -> dovecot/imap [mail@box 10.1.1.206 SEARCH](client_handle_input+0x2d) [0x80566ed] -> dovecot/imap [mail@box 10.1.1.206 SEARCH](client_input+0x5f) [0x805710f] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run+0xf5) [0xb76b0115] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_run+0x30) [0xb76af160] -> /usr/lib/dovecot/libdovecot.so.0(master_service_run+0x2a) [0xb769ca5a] -> dovecot/imap [mail@box 10.1.1.206 SEARCH](main+0x2b5) [0x805fb95] -> /lib/libc.so.6(__libc_start_main+0xe5) [0xb752e455] -> dovecot/imap [mail@box 10.1.1.206 SEARCH] [0x804e951] Apr 26 12:58:34 master: Error: service(imap): child 13220 killed with signal 6 (core dumps disabled) Login is working fine, just blowing up on that assert by doing a search. The search is pretty basic, in python: server.search( None, "(ALL)" ) Testing with a previous version (from a couple of days earlier -- I don't have the intervening), I don't hit that assertion. The last changeset on mail-search-build.c in the working copy is 11146:3649ad8c9f9a. Let me know if there's anything you'd like me to try.

2 1

[Dovecot] pop3-login: auth failure -- due to number as first character in login name
by dovecot mail list account 29 Apr '10

29 Apr '10

Seem to have found a error in version 1.2.6 regarding usernames having a number as the first character. Customers with a number as the first character in thier user name can not login via pop3 clients. However they can login through our webmail interface. If we remove the number from the username then the customer can login. # 1.2.6: /etc/dovecot.conf # OS: Linux 2.6.31.12-desktop-3mnb x86_64 Mandriva Linux 2010.0 protocols: pop3 ssl_listen: * ssl_cert_file: /etc/pki/tls/certs/dovecot.pem ssl_key_file: /etc/pki/tls/private/dovecot.pem disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable: /usr/lib64/dovecot/pop3-login mail_privileged_group: mail mail_location: mbox:~/mail:INBOX=/home/mail/%u mail_executable: /usr/lib64/dovecot/pop3 mail_plugin_dir: /usr/lib64/dovecot/modules/pop3 pop3_uidl_format: %08Xv%08Xu lda: postmaster_address: postmaster(a)example.com auth default: mechanisms: plain login verbose: yes passdb: driver: pam userdb: driver: passwd socket: type: listen client: path: /var/spool/postfix/private/auth mode: 432 user: postfix group: postfix plugin: quota_warning: storage=95%% /usr/local/bin/quota-warning.sh 95 quota_warning2: storage=80%% /usr/local/bin/quota-warning.sh 80

2 1

[Dovecot] Problems with Mailbox is in inconsistent state, please relogin.
by Daniel Machado Grilo 29 Apr '10

29 Apr '10

Dear Dovecot users, I'm a sysadmin in a company that recently migrated the IMAP from courier-imap to dovecot :-) *urray* Everything with the perl migrating scripts (version 1.0) went with OKs and it was a lot of fun (and pray) to move 400Gb of mails and migrate.. and everything is working even better :-) The problem is for some outside users that use webmails, like: - Squirrel - Roundcube - Horde They claim that the connection is constantly dropping.. You might say now: "Webmail's fault!!!" as I said in the beginning :) But the IMAP logs shows SEVERAL times (almost every 10mins ?!): Apr 28 21:10:48 olive dovecot: IMAP(rcoelho): Maildir /cluster/mail/mailboxes/rcoelho sync: UID inserted in the middle of mailbox (192736 > 192720, file = msg.LDW0C) Apr 28 21:10:48 olive dovecot: IMAP(rcoelho): Disconnected: Mailbox is in inconsistent state, please relogin. Apr 28 19:14:11 olive dovecot: IMAP(pcorreia): Maildir /cluster/mail/mailboxes/pcorreia sync: UID < next_uid (7874 < 7875, file = msg.5BNVC) Apr 28 19:14:12 olive dovecot: IMAP(pcorreia): Disconnected: Mailbox is in inconsistent state, please relogin. and so on. I read somewhere on the web that configuring the mbox_lazy_write=no should work .. it did not :-) I have CentOS 5.4 in our mailserver, so the version of dovecot is not the latest (it is 1.0.7).. But of course, this is production environment and I don't want to update if its not needed :-) So my questions are: - Is this a known bug for this version and I have to upgrade ? - Is this a problem in my configs for dovecot ? - Can this be other problem than dovecot ? -- so probably i should be able to add more debug anoying messages to our logs ? - Probably something went wrong in the migration for these maildirs? So now what can we do to fix this ? Our mail is working and, at the moment, I found 3 users with this error.. They have about 10Gb of mail each, I don't think this is a problem for dovecot.. only for our storages.. :-) Tks for your patience, Best Regards!

4 3

[Dovecot] Maildircrypt
by brainstorm 29 Apr '10

29 Apr '10

Hi guys, I was just wondering if this idea makes sense overall or if there are major impediments (indexing performance ?): http://blogs.nopcode.org/brainstorm/2010/04/28/maildircrypt/ Thanks ! Roman

2 1

[Dovecot] Dovecot headers at start of text returned to KMail
by Barry Scott 29 Apr '10

29 Apr '10

The attachment show Kmail (kdepim-4.4.2-1.fc12.x86_64) accessing an email from dovecot over imap. From the wireshark-imap-trace.txt you can see at line 77 that the headers do not end cleanly "X-)". And at line 84 there is text from the headers that is not part of the body "-Spambayes-Trained: ham". The email being access by dovecot is in: 1272011889.M384349P30156.f12barry.office.onelan.co.uk,S=6494,W=6606:2,S Because of this bug kmail fails to show the message cleanly on screen. # 1.2.11: /etc/dovecot.conf # OS: Linux 2.6.32.11-99.fc12.x86_64 x86_64 Fedora release 12 (Constantine) login_dir: /var/run/dovecot/login login_executable(default): /usr/libexec/dovecot/imap-login login_executable(imap): /usr/libexec/dovecot/imap-login login_executable(pop3): /usr/libexec/dovecot/pop3-login mail_executable(default): /usr/libexec/dovecot/imap mail_executable(imap): /usr/libexec/dovecot/imap mail_executable(pop3): /usr/libexec/dovecot/pop3 mail_plugin_dir(default): /usr/lib64/dovecot/imap mail_plugin_dir(imap): /usr/lib64/dovecot/imap mail_plugin_dir(pop3): /usr/lib64/dovecot/pop3 auth default: passdb: driver: pam userdb: driver: passwd Barry

2 4

[Dovecot] How can I add change_password
by Yavuz Maşlak 29 Apr '10

29 Apr '10

Hello I use dovecot, postfix, mysql and squirrelmail as a mail server. it works well. Meanwhile I am novice at postfix. As you know, there is a change_password plugin on squirrelmail. I wish to add change_password plugin on squirrelmail. I couldn't find how to integrate postfix and change_password. How can I add it ?

2 1

[Dovecot] How can I integrate change_password plugin, dovecot and postfix
by Yavuz Maşlak 29 Apr '10

29 Apr '10

Hello I use dovecot, postfix, mysql and squirrelmail as a mail server. it works well. Meanwhile I am novice at postfix. As you know, there is a change_password plugin on squirrelmail. I wish to add change_password plugin on squirrelmail. I couldn't find how to integrate postfix and change_password. How can I add it ?

3 3

Re: [Dovecot] returned error 83 (Out of memory)
by J Contad 28 Apr '10

28 Apr '10

FYI, sorted out this issue by taking out a specific email manually. Not sure if it helps, but the email had these attributes: - X-Mailer is PHPMailer [version 1.73] - References: header is quite large, and contains email addresses external to the sender and I's conversation. Removing the email fixed the issue. -- John

3 3

Re: [Dovecot] returned error 83 (Out of memory)
by J Contad 28 Apr '10

28 Apr '10

I tried deleting the file, but got pretty much the same issues. Any other hints on how else to go about it? Cheers! John On Thu, 22 Apr 2010, J Contad wrote: > The folder (which is shared) has about 8 thousand emails in it. I'm not sure > if it's pertinent, but the dovecot.index.cache file is about 8MB. Did you tried to delete the dovecot.index.cache file, maybe it has been corrupted? Regards, - -- Steffen Kaiser -- John

1 0

[Dovecot] somebody using Net::Sieve (Perl) with dovecot 1.2.11?
by Heiko Schlittermann 28 Apr '10

28 Apr '10

Hello, I've just (hopefully) fixed some issue with Net::Sieve (0.05) to cooperate with a current (1.2.11) dovecot managesieve server and STARTTLS. Does somebody use this Perl library? If yes, did it work for you (STARTTLS triggers the problem)? If not, which other (Perl) library do you use? Best regards from Dresden/Germany Viele Grüße aus Dresden Heiko Schlittermann -- SCHLITTERMANN.de ---------------------------- internet & unix support - Heiko Schlittermann HS12-RIPE ----------------------------------------- gnupg encrypted messages are welcome - key ID: 48D0359B --------------- gnupg fingerprint: 3061 CFBF 2D88 F034 E8D2 7E92 EE4E AC98 48D0 359B -

2 1

[Dovecot] zlib plugin dovecot 2.0 - write email
by Alex Baule 26 Apr '10

26 Apr '10

Hello Everyone. There is possible to rename the email writed by a plugin, like zlib ? i explain.... i need to add 1 line in email header, because of that, i want to rewrite the filename to update the S and W flag. Tks !

1 0

[Dovecot] Anyone successfully using expire plugin?
by Thomas M Goerger 23 Apr '10

23 Apr '10

Anyone out there using expire? At this point, I'm not sure that it's even running. It was touching the plugin libraries yesterday, but doesn't appear to be doing so today. The plugin directory wasn't explictly set, so I've done so now, but it's still not touching these. Anyone have a cf file I could take a look at, or any ideas as to what might be happening? Thanks! ************************************************************************* * Tom Goerger - Email/Unix System Administrator * * * * University of Minnesota Email: tmg(a)umn.edu * * Operations, Infrastructure and Architecture Phone: 4-5804 * * Internet Services Office: 626J WBOB * * * *************************************************************************

2 1

[Dovecot] How can I make dovecot aware of user's email aliases?
by Ian P. Christian 23 Apr '10

23 Apr '10

Sieve will not reply to mail it doesn't think was sent directly to a user. I use an SQL auth backend, and a user might login as user1(a)domain.com. What method can I use to inform dovecot that user2(a)domain.com is a forward to user1(a)domain.com, such that the vacation auto response will fire when user2(a)domain.com is emailed (therefore, the mail ends up in user1(a)domain.com's mailbox). -- Blog: http://pookey.co.uk/blog Follow me on twitter: http://twitter.com/ipchristian

1 0

[Dovecot] best choice of user database file to work with postfix?
by Phil Howard 23 Apr '10

23 Apr '10

I'm setting up a Postfix and Dovecot combination. What I want to do is have a user database that (1) is not running from some engine (so not LDAP or SQL or such) ... and (2) is completely disassociated from system users (e.g. most email users are not in /etc/passwd and most /etc/passwd users are not email users). Ideal would be a one-file solution, which can be managed by text editing or simple command line tools. But what I want is ONE file that both Postfix (for valid recipients) and Dovecot (for user login authentication) can use together. An alternative is some way to get Postfix to go through Dovecot to query for users (at the time of mail arriving on SMTP so it doesn't queue anything that would later be rejected). This is a smallish setup on one server, with probably a max of 50 to 100 users and 50 or so role account mailboxes over the next year or two. Any recommendations?

18 50

[Dovecot] have I been added to the mailing list
by ttplayer 23 Apr '10

23 Apr '10

Hi, have I been added to the mailing list? Thank you.

1 0

[Dovecot] Quota monitoring
by Vlamsdoem 23 Apr '10

23 Apr '10

Hello, I would like to monitor users quotas in Dovecot but I don't find a way to do it for every user. Acutally I can log in with user & password to get the users quota but wat I need is a way to monitor the quota of every user with e.g an admin user which would have an acces to the quota of each mailbox. Is there a way to this? Thanks for your help.

6 11

[Dovecot] message to imap users when connecting ?
by Frank Bonnet 22 Apr '10

22 Apr '10

Hello Is it possible that a IMAP client (thunderbirs in my case) popup a message box containing a short message when a user connect to the server ? Is there such feature in Dovecot ? Thanks

3 2

[Dovecot] imap-zlib patch to fix runaway imap process
by Mike Abbott 22 Apr '10

22 Apr '10

Here is a patch to fix the problem where an imap process runs away when imap-zlib is in use. The backtrace showed: io_loop_handler_run -> stream_send_io -> client_output -> o_stream_flush -> o_stream_zlib_flush -> o_stream_zlib_send_flush When o_stream_zlib_flush returns 0, stream_send_io reinstates the IO_WRITE event, causing the infinite loop. diff -ur dovecot-2.0.beta4/src/plugins/zlib/ostream-bzlib.c dovecot-2.0.beta4+fix/src/plugins/zlib/ostream-bzlib.c --- dovecot-2.0.beta4/src/plugins/zlib/ostream-bzlib.c 2010-02-12 20:22:07.000000000 -0600 +++ dovecot-2.0.beta4+fix/src/plugins/zlib/ostream-bzlib.c 2010-04-20 19:15:29.000000000 -0500 @@ -122,15 +122,15 @@ static int o_stream_bzlib_flush(struct ostream_private *stream) { struct bzlib_ostream *zstream = (struct bzlib_ostream *)stream; + int ret; if (o_stream_bzlib_send_flush(zstream) < 0) return -1; - if (o_stream_flush(zstream->output) < 0) { + ret = o_stream_flush(zstream->output); + if (ret < 0) zstream_copy_error(zstream); - return -1; - } - return 0; + return ret; } static ssize_t diff -ur dovecot-2.0.beta4/src/plugins/zlib/ostream-zlib.c dovecot-2.0.beta4+fix/src/plugins/zlib/ostream-zlib.c --- dovecot-2.0.beta4/src/plugins/zlib/ostream-zlib.c 2010-02-18 22:01:56.000000000 -0600 +++ dovecot-2.0.beta4+fix/src/plugins/zlib/ostream-zlib.c 2010-04-20 19:15:29.000000000 -0500 @@ -184,15 +184,15 @@ static int o_stream_zlib_flush(struct ostream_private *stream) { struct zlib_ostream *zstream = (struct zlib_ostream *)stream; + int ret; if (o_stream_zlib_send_flush(zstream) < 0) return -1; - if (o_stream_flush(zstream->output) < 0) { + ret = o_stream_flush(zstream->output); + if (ret < 0) zstream_copy_error(zstream); - return -1; - } - return 0; + return ret; } static ssize_t

2 1

[Dovecot] Quota Issue
by Peter Fraser 21 Apr '10

21 Apr '10

Hi All I'm using dovecot version 1.2.4_1 and I noticed something strange. I checked the list but so far I haven't found anything that exactly fits my setup. I'm using maildirs and ldap as a backend to authenticate. A user went over quota, realized it and proceeded to delete a lot of email. Even now, after about 15 minutes, the quota indicator is still showing that he is over quota even though the users mailbox size is down by over 200MB. Why isn't Dovecot showing the change, and is there a way around this? Thanks for any help.

2 3

[Dovecot] corrupted index cache file, physcial size, in migration to snow leopard
by Norman Packard 21 Apr '10

21 Apr '10

Hi, I am in the middle of migrating a server from Apple's leopard to snow leopard. Having big problems with migrating mail to dovecot. The version is: # dovecotd --version 1.1.17apple0.5 and I use the migration script provide by Apple, /usr/libexec/dovecot/migrate_mail_data.pl The current problems concern migrating imap accounts. The old data typically shows up to mail clients as only partially migrated, with some issues there, like folders change from being listed on the server to being listed locally, but the biggest problem is that many many messages and folders have disappeared from view. But on the server, they seem to be there, in their correctly migrated place. The log files are full of messages like the ones below, which I'm guessing has something to do with the problem. The number of files with this problem is huge. Are there any scripts to rebuild the index cache files with correct physical sizes? Thanks in advance for any help! Best, Norman Apr 21 08:05:47 newhot dovecot[31163]: IMAP(*): Corrupted index cache file /Volumes/xserveHD2/var/spool/imap/dovecot/mail/A1F36563-F978-4F99- A1FF-EC5D223F18D6/dovecot.index.cache: Corrupted physical size for uid=8142: 4325 != 4214 Apr 21 08:05:50 newhot dovecot[31163]: IMAP(*): Corrupted index cache file /Volumes/xserveHD2/var/spool/imap/dovecot/mail/A1F36563-F978-4F99- A1FF-EC5D223F18D6/dovecot.index.cache: Corrupted physical size for uid=8145: 4625 != 4510 Apr 21 08:05:53 newhot dovecot[31163]: IMAP(*): Corrupted index cache file /Volumes/xserveHD2/var/spool/imap/dovecot/mail/A1F36563-F978-4F99- A1FF-EC5D223F18D6/dovecot.index.cache: Corrupted physical size for uid=8148: 4037 != 3956 Apr 21 08:06:04 newhot dovecot[31163]: IMAP(*): Corrupted index cache file /Volumes/xserveHD2/var/spool/imap/dovecot/mail/A1F36563-F978-4F99- A1FF-EC5D223F18D6/dovecot.index.cache: Corrupted physical size for uid=8158: 4673 != 4571

1 0

[Dovecot] Question about listescape plugin
by Luca Di Vizio 21 Apr '10

21 Apr '10

Hi all, with the listescape plugin enabled we a stange result. ** listescape enabled ** 2 list "" % * LIST (\HasChildren) "/" "INBOX" * LIST (\HasChildren) "/" "INBOX/" ** listescape disabled ** 2 list "" % * LIST (\HasChildren) "/" "INBOX" The prefix for this namespace is "INBOX/". Is correct to have "INBOX/" as result of LIST? Info about the system: - Centos 5.3/x86_64. - Dovecot 1.2.11 (I tried also with 2.0.beta4, same results). mail_nfs_storage: yes mail_nfs_index: yes mail_plugins(imap): listescape quota imap_quota namespace: type: private separator: / inbox: yes list: yes subscriptions: yes plugin: quota: maildir Thanks, Luca

1 0

[Dovecot] POP3: Restrict Connections per timeframe
by Michael Kliewe 21 Apr '10

21 Apr '10

Hi everybody, is it possible to limit the amount of POP3 connections/logins per 15 minutes? Many hoster do that, but I cannot find a configuration setting for that. I can only limit the amount of concurrent connections per IP, but not connections per time. Michael

1 0

[Dovecot] dovecot2 beta and gcc3(?) - compile problem
by Andre Hübner 21 Apr '10

21 Apr '10

Hello, try to compile current dovecot2 beta on older machine with gcc 3.3 Compiling results in error: libtool: link: gcc -std=gnu99 -g -O2 -Wall -W -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wformat=2 -Wbad-function-cast -o .libs/imap-login client.o client-authenticate.o imap-login-settings.o imap-proxy.o ../../src/login-common/.libs/libdovecot-login.so ../../src/lib-dovecot/.libs/libdovecot.so -lrt -Wl,-rpath -Wl,/usr/lib/dovecot ../../src/login-common/.libs/libdovecot-login.so: undefined reference to `ERR_peek_last_error' collect2: ld returned 1 exit status make[3]: *** [imap-login] Error 1 make[3]: Leaving directory `/usr/src/dovecot-1.1.rc12/dovecot-2.0.beta4/src/imap-login' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/usr/src/dovecot-1.1.rc12/dovecot-2.0.beta4/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/usr/src/dovecot-1.1.rc12/dovecot-2.0.beta4' make: *** [all] Error 2 Is there something missed on my side or problem with older gcc? dovecot 1.2.11 compiles just fine with same configure-line also there are some warnings like this: libtool: compile: gcc -DHAVE_CONFIG_H -I. -I../../../.. -I../../../../src/lib -I../../../../src/lib-settings -I../../../../src/lib-mail -I../../../../src/lib-imap -I../../../../src/lib-index -I../../../../src/lib-storage -I../../../../src/lib-storage/index -std=gnu99 -g -O2 -Wall -W -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wformat=2 -Wbad-function-cast -MT dbox-sync-rebuild.lo -MD -MP -MF .deps/dbox-sync-rebuild.Tpo -c dbox-sync-rebuild.c -fPIC -DPIC -o .libs/dbox-sync-rebuild.o In file included from dbox-sync-rebuild.c:5: ../../../../src/lib-index/mail-index-modseq.h:34: warning: parameter has incomplete type ../../../../src/lib-index/mail-index-modseq.h:52: warning: parameter has incomplete type or libtool: compile: gcc -DHAVE_CONFIG_H -I. -I../../../.. -I../../../../src/lib -I../../../../src/lib-settings -I../../../../src/lib-mail -I../../../../src/lib-imap -I../../../../src/lib-index -I../../../../src/lib-storage -I../../../../src/lib-storage/index -I../../../../src/lib-storage/index/dbox-common -std=gnu99 -g -O2 -Wall -W -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wformat=2 -Wbad-function-cast -MT mdbox-file.lo -MD -MP -MF .deps/mdbox-file.Tpo -c mdbox-file.c -fPIC -DPIC -o .libs/mdbox-file.o In file included from mdbox-storage.h:5, from mdbox-file.c:16: ../../../../src/lib-storage/mailbox-list-private.h:158: warning: parameter has incomplete type Thanks, Andre

2 3

[Dovecot] Question about listescape
by Luca Di Vizio 21 Apr '10

21 Apr '10

Hi all, with the listescape plugin enabled we a stange result. ** listescape enabled ** 2 list "" % * LIST (\HasChildren) "/" "INBOX" * LIST (\HasChildren) "/" "INBOX/" ** listescape disabled ** 2 list "" % * LIST (\HasChildren) "/" "INBOX" The prefix for this namespace is "INBOX/". Is correct to have "INBOX/" as result of LIST? Info about the system: - Centos 5.3/x86_64. - Dovecot 1.2.11 (I tried also with 2.0.beta4, same results). mail_nfs_storage: yes mail_nfs_index: yes mail_plugins(imap): listescape quota imap_quota namespace: type: private separator: / inbox: yes list: yes subscriptions: yes plugin: quota: maildir Thanks, Luca

1 0

[Dovecot] Checkpassword/prefetch/master_user and problems
by Emerson Pinter 20 Apr '10

20 Apr '10

Hi. I'm trying to use checkpassword for simple auth and masteruser auth. I have two programs, one called checkpassword-master (for masteruser lookup) and another called checkpassword for normal passdb and userdb lookup). All works fine for non-masteruser authentication (in this case dovecot makes a single call to checkpassword binary). But if a master-user authenticates, dovecot execute checkpassword-master binary, and then executes checkpassword binary. Checkpassword binary receive the MASTER_USER env and do all the checks correctly, when checkpassword finish, dovecot logs this message "auth(default): checkpassword: sighandler called for unknown child" and authentication fails... I'm trying checkpassword-reply and fd4 too, on 1.2.11 and 2.0b4, no success. Below is my config: auth default { mechanisms = plain passdb checkpassword { args = /srv/dovecot/checkpassword } passdb checkpassword { args = /srv/dovecot/checkpassword-master master=yes pass=yes } userdb prefetch { } } Anybody can help me ? Tks. -- Emerson Pinter Esta mensagem pode conter informações confidenciais, privilegiadas ou privadas. Caso não seja o destinatário, favor apagá-la e notificar o remetente. Saiba que o uso impróprio das informações existentes é estritamente proibido, sendo tratado conforme as normas da empresa e a legislação em vigor.

1 1

Re: [Dovecot] setgid failed - Not owner
by Zilon X 20 Apr '10

20 Apr '10

Changed home dir of users, in my configuration they are /var/postfix/spool and /var/postfix/dovecot/run, but with no effect. I still got the "fatal setgid failed: Not owner" ( Fatal: setgid(12345(postfix)) failed with euid=12346(dovecot), gid=54322(dovecot), egid=54322(dovecot): Not owner ) 2010/4/20 Andreas Schulze <andreas.schulze(a)datev.de>: > Am 19.04.2010 16:49 schrieb Zilon X: >> postfix:x:12345:12345::/dev/null:/bin/false >> dovecot:x:12346:54322::/dev/null:/bin/false > > I would never use /dev/null as homedir. > postfix home is /var/spool/postfix, dovecot likes /var/run/dovecot. > > -- > Andreas Schulze > Internetdienste | P532 > > DATEV eG > 90329 Nürnberg | Telefon +49 911 319-0 | Telefax +49 911 319-3196 > E-Mail info @datev.de | Internet www.datev.de > Sitz: 90429 Nürnberg, Paumgartnerstr. 6-14 | Registergericht Nürnberg, GenReg Nr.70 > Vorstand > Prof. Dieter Kempf (Vorsitzender) > Dipl.-Kfm. Wolfgang Stegmann (stellvertretender Vorsitzender) > Dipl.-Kfm. Michael Leistenschneider > Jörg Rabe v. Pappenheim > Dipl.-Vw. Eckhard Schwarzer > Vorsitzender des Aufsichtsrates: Reinhard Verholen > >

2 1

[Dovecot] zero sized messages in Maildir, corrupt filenames ?
by Rodolphe Buret 20 Apr '10

20 Apr '10

Hi, I'm starting to see more and more duplicates in my cur directory. they seem to be a copy of an incoming mail (messages share the same date) but they appears in the client as a message without subject and without sender or body. Their filename starts with an underscore and the 4th letter is a %. The rest of the filename is random (but it also includes my hostname). Which program could create such a filename ? If it's in the cur directory, it can only be dovecot, right ? Example of 2 probably related messages in the cur folder: 1785 Apr 20 09:18 msg.zMeU:2,d 0 Apr 20 09:18 _J9C%jVVzLB.{myhostname}:2, In addition, it seems that since we have started to use thunderbird 3, indexes are often corrupt, mailboxes needs reconstruction and so on. Could it be an incompatibility with my version of dovecot ? I've disabled my thunderbird filters that were moving the incoming mail automatically, no difference. I use no antivirus (linux imap client mainly). I can (probably) upgrade to dovecot-1.2.11-3_108 but I'd like to be sure that I won't create more problems before doing it. The homedirs are NFS mounted. dovecot 1.1.7-0_84 squirrelmail 1.4.19 postfix 2.6.5-1 Here is my dovecot -n # 1.1.7: /etc/dovecot.conf # OS: Linux 2.6.18-128.1.1.el5xen x86_64 Red Hat Enterprise Linux Server release 5.3 (Tikanga) log_path: /var/log/dovecot.log protocols: imap imaps listen: * login_dir: /var/run/dovecot/login login_executable: /usr/libexec/dovecot/imap-login mail_location: maildir:~/Maildir mmap_disable: yes mail_nfs_storage: yes mail_nfs_index: yes lock_method: dotlock auth default: passdb: driver: pam userdb: driver: passwd plugin: quota: Thanks for your time, -- Rodolphe Buret Ecole Polytechnique Fédérale de Lausanne System Administrator EPFL STI-IEL-ESL Tel: +41-21-693-0922 ELG 136 (bâtiment ELG) Fax: +41-21-693-0909 Station 11 Email: rodolphe.buret(a)epfl.ch CH-1015 Lausanne Switzerland

2 1

[Dovecot] Dovecot-sieve vacation proposal, use address in addresses as recipient
by David Berard 20 Apr '10

20 Apr '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Dovecot list, In my setup I have several email addresses that uses same account. When I set a vacation the answer is not necessarily send from the destination address. I propose to use the addresses found in addresses (if one match) as From address in vacation reply. I've made a little patch who do that [1]. Best regards. [1] http://polymorf.fr/files/nfrance/cmd-vacation.patch - -- David BERARD - ------------------------------------------------- contact(at)davidberard.fr GPG|PGP KeyId 0xC8533354 GPG|PGP Key http://davidberard.fr/C8533354.gpgkey - ------------------------------------------------- * No electrons were harmed in * * the transmission of this email * -----BEGIN PGP SIGNATURE----- iEYEARECAAYFAku+fUoACgkQOL7fhchTM1TJmACfdQcfgrSqNbjvWTEXrTc7Twly M70An2Nzb2MFby9ARHM1Ys93wGzuRYHy =yaLh -----END PGP SIGNATURE-----

4 4

[Dovecot] setgid failed - Not owner
by Zilon X 19 Apr '10

19 Apr '10

Hello ALL, I got errors on a new installation, maybe I forgot something. The system is Solaris, Dovecot 1.2.10, Postfix 2.7.0, MySQL 5.1.25. Can send mail with dovecot imap, but can`t retrieve the messages. Strange, I can save and retrieve draft messages ok, files are always with postfix:postfix ownership. Sorry if this is newbie question, but can`t find a solution. Syslog info: Apr 19 10:10:17 server postfix/qmgr[10654]: [ID 197553 mail.info] B66EF3FEE7: from=<user@domain>, size=547, nrcpt=1 (queue active) Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] auth(default): master in: USER 1 user@domain service=deliver Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] auth-worker(default): sql(user@domain): SELECT '/var/postfix/domains/domain/user' as home, 'maildir:/var/postfix/domains/domain/user' as mail, 12345 AS uid, 12345 AS gid, concat('dirsize:storage=', quota) AS quota FROM mailbox WHERE username = 'user@domain' AND active = '1' Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] auth(default): master out: USER 1 user@domain home=/var/postfix/domains/domain/user mail=maildir:/var/postfix/domains/domain/user uid=12345 gid=12345 quota=dirsize:storage=10240000 Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] deliver(user@domain): auth input: home=/var/postfix/domains/domain/user Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] deliver(user@domain): auth input: mail=maildir:/var/postfix/domains/domain/user Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] deliver(user@domain): auth input: uid=12345 Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] deliver(user@domain): auth input: gid=12345 Apr 19 10:10:17 server dovecot: [ID 583609 mail.info] deliver(user@domain): auth input: quota=dirsize:storage=10240000 >> Apr 19 10:10:17 server dovecot: [ID 583609 mail.crit] deliver(user@domain): Fatal: setgid(12345(postfix)) failed with euid=12346(dovecot), gid=54322(dovecot), egid=54322(dovecot): Not owner Apr 19 10:10:17 server postfix/pipe[23327]: [ID 197553 mail.info] B66EF3FEE7: to=<user@domain>, relay=dovecot, delay=582, delays=582/0.18/0/0.11, dsn=4.3.0, status=deferred (temporary failure) _____________________________________________________ Configuration at dovecot.conf: login_dir = /var/postfix/dovecot/run/login login_chroot = yes login_user = dovecot mail_privileged_group = postfix mail_full_filesystem_access = yes passdb sql { args = /var/postfix/dovecot/etc/dovecot-sql.conf } userdb sql { args = /var/postfix/dovecot/etc/dovecot-sql.conf } user = nobody socket listen { master { path = /var/postfix/dovecot/run/auth-master mode = 0660 user = postfix group = postfix } client { path = /var/postfix/spool/private/auth mode = 0660 user = postfix group = postfix } } _____________________________________________________ Configuration at main.cf: mail_owner = postfix default_privs = nobody virtual_transport = dovecot dovecot_destination_recipient_limit = 1 smtpd_sasl_type = dovecot _____________________________________________________ Configuration at master.cf: dovecot unix - n n - - pipe flags=DRhu user=dovecot:dovecot argv=/var/postfix/dovecot/libexec/dovecot/deliver -d $(recipient) _____________________________________________________ Entries at passwd: postfix:x:12345:12345::/dev/null:/bin/false dovecot:x:12346:54322::/dev/null:/bin/false Entries at group: postfix::12345: postdrop::54321: dovecot::54322: _____________________________________________________ Thank you for any help!

1 0

[Dovecot] permission denied, no logs
by Shane Brooks 19 Apr '10

19 Apr '10

Hi, We have a sendmail/dovecot/squirrelmail installation that was working ok except for deleted files showing in the list with a [d] beside it. In trying to fix that issue, I apparently made a new error appear in the left navigation area: ERROR: Could not complete request. Query: CREATE "mail/Sent" Reason Given: [NOPERM] Permission denied There are no errors in dovecot, but from what i've read squirrelmail passes these folder requests to the imap server for processing. Can anyone help me verify what my permissions should be? Here is my config: # 1.2.9: /etc/dovecot/dovecot.conf # OS: Linux 2.6.24-19-xen x86_64 CentOS release 5.3 (Final) log_path: /var/log/dovecot/runtime.log protocols: imap imaps pop3 pop3s disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable(default): /usr/libexec/dovecot/imap-login login_executable(imap): /usr/libexec/dovecot/imap-login login_executable(pop3): /usr/libexec/dovecot/pop3-login mail_privileged_group: mail mail_uid: mail mail_gid: mail mail_location: mbox:~:INBOX=/var/mail/%u mail_executable(default): /usr/libexec/dovecot/imap mail_executable(imap): /usr/libexec/dovecot/imap mail_executable(pop3): /usr/libexec/dovecot/pop3 mail_plugin_dir(default): /usr/lib/dovecot/imap mail_plugin_dir(imap): /usr/lib/dovecot/imap mail_plugin_dir(pop3): /usr/lib/dovecot/pop3 lda: postmaster_address: postmaster(a)example.com auth default: cache_size: 10 debug: yes passdb: driver: pam args: cache_key=%u dovecot userdb: driver: passwd Thank you very much! Shane brooks IT Manager 940 West Oakland Avenue Suite A8 Oakland, Florida 34787 407-654-6076 Office 407-654-6063 Fax

2 1

[Dovecot] sieve rule problem after upgrade
by Benny Pedersen 19 Apr '10

19 Apr '10

after i updated from 1.2.6 to 1.2.11 in gentoo this sieve rule stopped working :( # X-Cron-Env if exists "X-Cron-Env" { addflag "\\Seen"; fileinto "Trash"; removeflag "\\Seen"; stop; } any hints why ?, this rule is generated with horde ingo so it might be problem in ingo ? -- xpoint

2 2

[Dovecot] dovecot-1.2-sieve sending vacation replies regardless of recipient (patch)
by Mark Washenberger 19 Apr '10

19 Apr '10

I want a way to send VAR messages without having to specify up front all of the addresses for which such VARs should be sent. I realize this goes against rfc 3834, but it seemed reasonable to add a tag to the vacation command to skip the [to, cc, bcc, ...] recipients check. The patch I made is attached. The new usage with this patch is: vacation [":days" number] [":subject" string] [":x_any_address"] [":from" string] [":addresses" string-list] [":mime"] [":handle" string] <reason: string> What do you think of this patch? Is there a better way to accomplish the same thing without writing a patch? Is there a better way to write this patch to make it a cleaner feature or easier to maintain? Could this ever be part of the mainline? Mark Washenberger Rackspace Hosting Software Developer, Email & Apps mark.washenberger(a)rackspace.com

2 2

[Dovecot] sieve plugin - reject sends MDN. Can it 550 reject ?
by Ken A 19 Apr '10

19 Apr '10

I'm using dovecot-1.2-sieve-0.1.15 with dovecot-1.2.11. I have a reject test filter that looks like this: require "reject"; if header :contains "subject" "reject" { reject "bye"; } Is there any way to get a "reject" filter to send back a 550 reject as detailed in http://ietfreport.isoc.org/idref/rfc5429/#section-2.5 rather than a MDN? The MDN can result in blowback spam. :-( Thanks, Ken -- Ken Anderson Pacific Internet - http://www.pacific.net

3 3

[Dovecot] siever_dir default
by Antonio Casado Rodríguez 19 Apr '10

19 Apr '10

Hi all I have this: # The path to the directory where the personal Sieve scripts are stored. For # ManageSieve this is where the uploaded scripts are stored. #sieve_dir=~/sieve but the correct is # The path to the directory where the personal Sieve scripts are stored. For # ManageSieve this is where the uploaded scripts are stored. # Default value: {mail_location}/sieve #sieve_dir=~/sieve

2 1

[Dovecot] PostgreSQL driver supporting [round-robin] load balancing and redundancy [LONG]
by Rainer Weikusat 18 Apr '10

18 Apr '10

One of the things my employer uses dovecot for is as mail download server for an 'e-mail purification service' (AV/ anti-spam) for smartphones. The service itself presently runs on a rented server somewhere in the UK and the corresponding 'web service' front-end and user account/ mail account database resides on a server in Germany. The UK dovecot server uses the PostgreSQL server on the German machine for user authentication. The latter is reachable using two entirely different 'internet paths' and there was an outage of several hours on one of them a couple of weeks ago. This prompted some frantic network reconfiguration efforts in order to get the abovementioned service going again and resulted in the conviction that - ideally - the dovecot server should be capable of using connections to multiple PostgreSQL servers (or a single server reachable via several IPs) simultaneoulsy, distributing requests among them, and should be capable of detecting a possible problem on one of the db server connections and use the still functioning ones to continue operations. After having spent about eight hours reading through the code of existing pgsql driver, I concluded that modifying that in order to achieve the goals outlined above would require some major brain surgery with a high chance of causing a lot of collateral damage in the course of that and that writing a new driver providing the sought-after feature set was the better idea. Which is the point of this e-mail. Example Usage ------------- Multiple connections can be specified in an ordinary connect-string by separating the necessary parameters with a ;;;-sequence. For the server I was writing about above, this looks like: connect = host=1.1.1.1 dbname=mailgate user=mailgate_user password=secret sslmode=require ;;; \ host=2.2.2.2 dbname=mailgate user=mailgate_user password=secret sslmode=require Leading and trailing whitespace in a connect sub-string is ignored. Basic Operation --------------- After having split the connect string, the driver creates a connection structure (struct multi_pgsql_pgc) for each connection specification contained in that and initiates an asynchronous connect for each. As the actual connections become ready, they are put onto a (linked list) (FIFO) queue of 'connections which are ready'. When a query is submitted to the driver, the first connection structure is removed from the ready list and used to execute the query. After this has finished, the connection is added to the end of the 'connections which are ready' list. If no connection is ready at the time a query was submitted, the query is put on a FIFO queue of queries which will be executed as the resources to actually do so become availabe. Each query has a lifetime of 60 seconds (arbitrary). A query which couldn't be started before his lifetime was over will be aborted with a timeout error. A connection which is currently being used to execute a query will be considered dead if no forward progress of any kind has happened for 20 seconds (also arbitrary). If this happens (or if some kind of I/O error occurs), the existing connection will be closed and an asychronous connect on a newly created one will be started. Assuming the lifetime of the query isn't yet over, it will either be started again on another available connection or pushed back onto the front of the query queue, causing it to be started as soon as a connection becomes available. Attemtps to reestablish the failed connection will continue (with 5 seconds of delay between each) in the background until success. Design Overview --------------- struct multi_pgsql_pgdb & routines working on that provide both the 'simple' query interface and transaction support. struct multi_pgsql_pgc represents a database connection. The result interface is implemented as struct multi_pgsql_result & assorted subroutines. Transactions are provided by a struct multi_pgsql_transaction_context which contains a linked-list of transaction subqueries (struct multi_pgsql_transaction_subquery). A 'class hierarchy' of 'query classes' is used to provide the different kinds of 'queries' supported by the driver (simple queries, simple synchronous queries, transaction queries, synchronous transaction queries). The complete hierarchy itself looks like this: multi_pgsql_query / \ / \ multi_pgsql_simple_query multi_pgsql_transaction_query / \ / \ / \ / \ multi_pgsql_user_query multi_pgsql_sync_query \ / \ / \ multi_pgsql_async_transaction_query multi_pgsql_sync_transaction_query Notable Difference to the pgsql Driver -------------------------------------- sql_exec This interface is not implemented. It is (as far as I can tell) not used anywhere except inside the sqllite driver. It cannot be used for queries which return information because there is no way to access this information and it can neither generally be used for queries which cause side effects since it cannot be determined if execution was successful (the only possible use I can think of is actually a rollback after a failed transaction). sql_result_get_field_value_binary Not implemented. Appears to be unused and the way I understand the libpq programming documentation, the only way to retrieve a binary result to some query is to use PQsendQueryParams or PQsendQueryPrepare and request that some columns are returned in binary. Nothing like this is supported in the present, abstract SQL database interface. synchronous query execution Because background connection reestablishment might still be going on after a synchronous query has been executed, it is necessary to switch any timeouts/ io operations registered to the ioloop used for synchronous query execution back to the main ioloop. The code implements this by using io_loop_set_current to temporarily restore the original ioloop, unregistering and registering any timeouts/ io-requests possibly associated with one of the existing database connections, switching back to the synchronous execution ioloop afterwards and then destroying that via io_loop_destroy. This works but is probably not exactly how the existing interfaces were supposed to be used. error messages for synchronously committed transaction These are put onto the data stack by the commit_s subroutine which will hopefully (judging from the dict-code, this should work, too) be sufficient to enable them to survive until someone higher up in the callchain wants to look at them. Remarks on the Patch -------------------- Created against dovecot-1-2-cf3fe573a560. Applies cleanly to 1.2.11. Applies to 2.0.beta4 with a rejection because of a feature which doesn't exist anymore. The result can be compiled but it hasn't been tested. ---- Index: dovecot/configure.in =================================================================== RCS file: /sysdata/cvs/dovecot/configure.in,v retrieving revision 1.1.1.2 retrieving revision 1.1.1.2.6.9 diff -u -r1.1.1.2 -r1.1.1.2.6.9 --- dovecot/configure.in 15 Mar 2010 18:18:14 -0000 1.1.1.2 +++ dovecot/configure.in 5 Apr 2010 16:16:03 -0000 1.1.1.2.6.9 @@ -159,6 +159,11 @@ [ --with-sqlite Build with SQLite3 driver support], TEST_WITH(sqlite, $withval), want_sqlite=no) + +AC_ARG_WITH(multi-pgsql, +[ --with-multi-pgsql Build with PostgreSQL driver supporting multiple connections], + TEST_WITH(multi-pgsql, $withval), + want_multi_pgsql=no) AC_ARG_WITH(lucene, [ --with-lucene Build with CLucene full text search support], @@ -1969,7 +1974,12 @@ LIBS=$old_LIBS fi -if test $want_pgsql != no; then +want_pgsql_driver=no +test "$want_pgsql" != no && want_pgsql_driver=yes +test "$want_multi_pgsql" != no && want_pgsql_driver=yes + + +if test $want_pgsql_driver != no; then AC_CHECK_PROG(PG_CONFIG, pg_config, YES, NO) if test $PG_CONFIG = NO; then # based on code from PHP @@ -2014,19 +2024,20 @@ fi PGSQL_LIBS="$PGSQL_LIBS -lpq" AC_DEFINE(HAVE_PGSQL,, Build with PostgreSQL support) - found_sql_drivers="$found_sql_drivers pgsql" + test "$want_pgsql" != no && found_sql_drivers="$found_sql_drivers pgsql" + test "$want_multi_pgsql" != no && found_sql_drivers="$found_sql_drivers multi_pgsql" if test "$all_sql_drivers" = "yes"; then - sql_drivers="$sql_drivers pgsql" + sql_drivers="$sql_drivers pgsql pgsql_timeout" fi ], [ - if test $want_pgsql = yes; then + if test $want_pgsql_driver = yes; then AC_ERROR([Can't build with PostgreSQL support: libpq-fe.h not found]) fi ]) CPPFLAGS=$old_CPPFLAGS ], [ - if test $want_pgsql = yes; then + if test $want_pgsql_driver = yes; then AC_ERROR([Can't build with PostgreSQL support: libpq not found]) fi ]) @@ -2323,6 +2334,7 @@ build_pgsql=no build_mysql=no build_sqlite=no +build_multi_pgsql=no for driver in $sql_drivers; do if test "$driver" = "pgsql"; then AC_DEFINE(BUILD_PGSQL,, Built-in PostgreSQL support) @@ -2333,7 +2345,10 @@ elif test "$driver" = "sqlite"; then AC_DEFINE(BUILD_SQLITE,, Built-in SQLite support) build_sqlite=yes - fi + elif test "$driver" = "multi_pgsql"; then + AC_DEFINE(BUILD_MULTI_PGSQL,, Built-in PostgreSQL support with multiple connection support) + build_multi_pgsql=yes + fi done if test $build_pgsql = no; then not_sql_drivers="$not_sql_drivers pgsql" @@ -2344,11 +2359,15 @@ if test $build_sqlite = no; then not_sql_drivers="$not_sql_drivers sqlite" fi +if test $build_multi_pgsql = no; then + not_sql_drivers="$not_sql_drivers multi_pgsql" +fi AC_SUBST(sql_drivers) AM_CONDITIONAL(BUILD_PGSQL, test "$build_pgsql" = "yes") AM_CONDITIONAL(BUILD_MYSQL, test "$build_mysql" = "yes") AM_CONDITIONAL(BUILD_SQLITE, test "$build_sqlite" = "yes") +AM_CONDITIONAL(BUILD_MULTI_PGSQL, test "$build_multi_pgsql" = "yes") AM_CONDITIONAL(SQL_PLUGINS, test "$want_sql" = "plugin") dnl ** Index: dovecot/src/lib-sql/Makefile.am =================================================================== RCS file: /sysdata/cvs/dovecot/src/lib-sql/Makefile.am,v retrieving revision 1.1.1.1 retrieving revision 1.1.1.1.6.3 diff -u -r1.1.1.1 -r1.1.1.1.6.3 --- dovecot/src/lib-sql/Makefile.am 28 Dec 2009 13:52:04 -0000 1.1.1.1 +++ dovecot/src/lib-sql/Makefile.am 5 Apr 2010 15:05:54 -0000 1.1.1.1.6.3 @@ -15,11 +15,16 @@ SQLITE_LIB = libdriver_sqlite.la SQL_DRIVER_PLUGINS += sqlite endif +if BUILD_MULTI_PGSQL +MULTI_PGSQL_LIB = libdriver_multi_pgsql.la +SQL_DRIVER_PLUGINS += multi_pgsql +endif sql_module_LTLIBRARIES = \ $(MYSQL_LIB) \ $(PGSQL_LIB) \ - $(SQLITE_LIB) + $(SQLITE_LIB) \ + $(MULTI_PGSQL_LIB) sql_moduledir = $(moduledir)/sql endif @@ -39,7 +44,8 @@ driver_sources = \ driver-mysql.c \ driver-pgsql.c \ - driver-sqlite.c + driver-sqlite.c \ + driver-multi-pgsql.c endif libsql_a_SOURCES = \ @@ -62,6 +68,11 @@ libdriver_sqlite_la_LIBADD = $(SQLITE_LIBS) libdriver_sqlite_la_CPPFLAGS = -I$(top_srcdir)/src/lib $(SQLITE_CFLAGS) libdriver_sqlite_la_SOURCES = driver-sqlite.c + +libdriver_multi_pgsql_la_LDFLAGS = -module -avoid-version +libdriver_multi_pgsql_la_LIBADD = $(MULTI_PGSQL_LIBS) +libdriver_multi_pgsql_la_CPPFLAGS = -I$(top_srcdir)/src/lib $(MULTI_PGSQL_CFLAGS) +libdriver_multi_pgsql_la_SOURCES = driver-multi-pgsql.c endif headers = \ Index: dovecot/src/lib-sql/driver-multi-pgsql.c =================================================================== RCS file: dovecot/src/lib-sql/driver-multi-pgsql.c diff -N dovecot/src/lib-sql/driver-multi-pgsql.c --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ dovecot/src/lib-sql/driver-multi-pgsql.c 14 Apr 2010 15:31:48 -0000 1.1.2.134 @@ -0,0 +1,1763 @@ +/* + Copyright (c) MAD Partners, Ltd 2010 (rweikusat(a)madpartnerltd.com) + Portions Copyright (c) 2004-2010 Dovecot authors, see the included COPYING file +*/ + +/* includes */ +#include "lib.h" +#include "array.h" +#include "ioloop.h" +#include "ioloop-internal.h" /* kind of dirty, but it should be fine.. */ +#include "sql-api-private.h" + +#ifdef BUILD_MULTI_PGSQL +#include <ctype.h> +#include <stdlib.h> +#include <time.h> +#include <libpq-fe.h> + +/* macros */ +/* #define DEBUG */ +#ifdef DEBUG +# define dprintf(args) i_info args +#else +# define dprintf(args) +#endif + +#define DEINIT_ERROR "db driver deinit" +#define TIMEOUT_ERROR "query timed out" + +/* constants */ +enum { + MULTI_PGSQL_POOL = 512, + MULTI_PGSQL_XACT_POOL = 1024, + + TIMEOUT_UNIT = 1000, + + RECONNECT_DELAY = 5 * TIMEOUT_UNIT, + USER_QUERY_LIFETIME = 60, + + PGC_IO_TIMEOUT = 20, + MIN_QUERY_TIMEOUT = 10 +}; + +/** general helper routines */ +static char *kill_pg_errmsg_newline(char *s) +{ + char *r; + unsigned c; + + r = s; + while ((c = *r) && c != '\n') ++r; + if (c) *r = 0; + + return s; +} + +/** error result */ +struct multi_pgsql_error_result { + struct sql_result api; + char *msg; +}; + +static void error_result_free(struct sql_result *r) +{ + struct multi_pgsql_error_result *result; + + result = (void *)r; + dprintf(("%s: %p", __func__, result)); + + i_free(result->msg); + i_free(result); +} + +static void error_result_nop(struct sql_result *r) +{ + dprintf(("%s: %p", __func__, r)); +} + +static int error_result_next_row(struct sql_result *r) +{ + dprintf(("%s: %p", __func__, r)); + return -1; +} + +static char const *error_result_get_error(struct sql_result *r) +{ + struct multi_pgsql_error_result *result; + + result = (void *)r; + dprintf(("%s: %p", __func__, result)); + + return result->msg; +} + +static void init_error_result(struct multi_pgsql_error_result *result, + char const *msg) +{ + struct sql_result_vfuncs *v; + + memset(&result->api, 0, sizeof(result->api)); + + v = &result->api.v; + v->free = error_result_nop; + v->next_row = error_result_next_row; + v->get_error = error_result_get_error; + + result->msg = (char *)msg; + + dprintf(("%s: %p", __func__, result)); +} + +/** multi_pgsql_result */ +struct multi_pgsql_result { + struct sql_result api; + struct multi_pgsql_pgc *pgc; + + PGresult *pgr; + unsigned row, n_rows; + char const **values, *errmsg; +}; + +/*** methods/ subroutines */ +static void multi_pgsql_result_free(struct sql_result *r) +{ + struct multi_pgsql_result *result; + + result = (void *)r; + + PQclear(result->pgr); + if (result->values) i_free(result->values); + i_free(result); +} + +static void multi_pgsql_result_nop(struct sql_result *r) +{ + (void)r; +} + + +static int multi_pgsql_result_next_row(struct sql_result *r) +{ + struct multi_pgsql_result *result; + unsigned row; + + dprintf(("%s: %p", __func__, r)); + + result = (void *)r; + row = result->row; + if (++row < result->n_rows) return result->row = row; + return 0; +} + +static int multi_pgsql_result_first_row(struct sql_result *r) +{ + struct multi_pgsql_result *result; + PGresult *pgr; + ExecStatusType status; + unsigned n_rows; + + result = (void *)r; + pgr = result->pgr; + + status = PQresultStatus(pgr); + dprintf(("%s: %p: result status %s(%d)", + __func__, r, PQresStatus(status), status)); + + switch (status) { + case PGRES_COMMAND_OK: + case PGRES_COPY_OUT: + case PGRES_COPY_IN: + n_rows = 0; + break; + + case PGRES_TUPLES_OK: + n_rows = PQntuples(pgr); + + dprintf(("%s: %p: %u rows", __func__, result, n_rows)); + break; + + default: + return -1; + } + + result->n_rows = n_rows; + result->api.v.next_row = multi_pgsql_result_next_row; + return n_rows; +} + +static unsigned multi_pgsql_result_get_fields_count(struct sql_result *r) +{ + struct multi_pgsql_result *result; + unsigned count; + + result = (void *)r; + count = PQnfields(result->pgr); + + dprintf(("%s: %p: %u", __func__, result, count)); + return count; +} + +static char const *multi_pgsql_result_get_field_name(struct sql_result *r, + unsigned i) +{ + struct multi_pgsql_result *result; + char const *name; + + result = (void *)r; + name = PQfname(result->pgr, i); + + dprintf(("%s: %p: %u: '%s'", __func__, result, i, name)); + return name; +} + +static int multi_pgsql_result_find_field(struct sql_result *r, + char const *fname) +{ + struct multi_pgsql_result *result; + int i; + + result = (void *)r; + i = PQfnumber(result->pgr, fname); + + if (i == -1) { + i_info("%s: %p: no field named '%s'", + __func__, result, fname); + return -1; + } + + dprintf(("%s: %p: '%s' <-> %d", + __func__, result, fname, i)); + return i; +} + +static char const *multi_pgsql_result_get_field_value(struct sql_result *r, + unsigned i) +{ + struct multi_pgsql_result *result; + char const *value; + unsigned row; + + result = (void *)r; + row = result->row; + value = PQgetvalue(result->pgr, row, i); + if (!*value && PQgetisnull(result->pgr, row, i)) value = NULL; + + dprintf(("%s: %p: %u,%u: '%s'", + __func__, result, row, i, value ? value : "NULL")); + return value; +} + +static unsigned char const * +multi_pgsql_result_get_field_value_binary(struct sql_result *r, + unsigned i, size_t *size) +{ + (void)i; + i_info("%s: %p: not implemented", __func__, r); + + *size = 0; + return NULL; +} + +static char const *multi_pgsql_result_find_field_value(struct sql_result *r, + char const *fname) +{ + int i; + + dprintf(("%s: %p", __func__, r)); + + i = r->v.find_field(r, fname); + return i != -1 ? r->v.get_field_value(r, i) : NULL; +} + +static char const * const *multi_pgsql_result_get_values(struct sql_result *r) +{ + struct multi_pgsql_result *result; + char const **values; + unsigned n; + + dprintf(("%s: %p", __func__, r)); + result = (void *)r; + + values = result->values; + if (values) return values; + + n = r->v.get_fields_count(r); + values = result->values = i_new(char const *, n); + do { + --n; + values[n] = r->v.get_field_value(r, n); + } while (n); + + return values; +} + +static char const * +multi_pgsql_result_get_error(struct sql_result *r) +{ + struct multi_pgsql_result *result; + char const *errmsg; + + result = (void *)r; + + errmsg = result->errmsg; + if (errmsg) return errmsg; + + errmsg = result->errmsg = + kill_pg_errmsg_newline(PQresultErrorMessage(result->pgr)); + return errmsg; +} + +/*** vtable/ init */ +static struct sql_result multi_pgsql_result = { + MEMBER(v) { + multi_pgsql_result_nop, + multi_pgsql_result_first_row, + multi_pgsql_result_get_fields_count, + multi_pgsql_result_get_field_name, + multi_pgsql_result_find_field, + multi_pgsql_result_get_field_value, + multi_pgsql_result_get_field_value_binary, + multi_pgsql_result_find_field_value, + multi_pgsql_result_get_values, + multi_pgsql_result_get_error + } +}; + +static void init_multi_pgsql_result(struct multi_pgsql_result *result, + PGresult *pgr) +{ + result->api = multi_pgsql_result; + result->pgr = pgr; + result->row = 0; + result->values = NULL; +} + +/** query classes */ +struct multi_pgsql_query { + struct multi_pgsql_query *p; + struct multi_pgsql_query_vtable *vtable; +}; + +struct multi_pgsql_query_vtable { + void (*start)(struct multi_pgsql_query *, struct multi_pgsql_pgc *); + void (*result)(struct multi_pgsql_query *, struct multi_pgsql_pgc *, + PGresult **); + void (*abort)(struct multi_pgsql_query *, struct sql_result *); + void (*dtor)(struct multi_pgsql_query *); + time_t (*eol)(struct multi_pgsql_query *); + void (*set_sync_ioloop)(struct multi_pgsql_query *, struct ioloop *); + char const *(*get_query)(struct multi_pgsql_query *); +}; + +static void start_new_query_on_pgc(struct multi_pgsql_query *, struct multi_pgsql_pgc *); +static void start_query_on_pgc(struct multi_pgsql_pgc *); +static void done_with_query(struct multi_pgsql_pgc *); + +/*** method invocation wrappers */ +#define the_qry(q) ((struct multi_pgsql_query *)q) + +static inline void start_query(void *q, struct multi_pgsql_pgc *pgc) +{ + dprintf(("%s: %p", __func__, q)); + the_qry(q)->vtable->start(the_qry(q), pgc); +} + +static inline void query_result(void *q, struct multi_pgsql_pgc *pgc, + PGresult **pgr) +{ + dprintf(("%s: %p", __func__, q)); + the_qry(q)->vtable->result(the_qry(q), pgc, pgr); +} + +static inline void abort_query(void *q, struct sql_result *r) +{ + dprintf(("%s: %p", __func__, q)); + + the_qry(q)->vtable->abort(the_qry(q), r); + the_qry(q)->vtable->dtor(the_qry(q)); +} + +static inline void destroy_query(void *q) +{ + dprintf(("%s: %p", __func__, q)); + the_qry(q)->vtable->dtor(the_qry(q)); +} + +static inline time_t query_eol(void *q) +{ + dprintf(("%s: %p", __func__, q)); + return the_qry(q)->vtable->eol(the_qry(q)); +} + +static inline void set_query_sync_ioloop(void *q, struct ioloop *ioloop) +{ + dprintf(("%s: %p", __func__, q)); + the_qry(q)->vtable->set_sync_ioloop(the_qry(q), ioloop); +} + +static inline char const *get_query(void *q) +{ + return + the_qry(q)->vtable->get_query(the_qry(q)); +} + +#undef the_qry + +/*** query */ +static void init_multi_pgsql_query(struct multi_pgsql_query *qry, + struct multi_pgsql_query_vtable *vtable) +{ + qry->p = NULL; + qry->vtable = vtable; +} + +/*** simple_query */ +struct multi_pgsql_simple_query { + struct multi_pgsql_query super; + + time_t eol_at; + char *query; +}; + +static void start_simple_query(struct multi_pgsql_query *q, + struct multi_pgsql_pgc *pgc) +{ + start_new_query_on_pgc(q, pgc); +} + +static time_t simple_query_eol(struct multi_pgsql_query *q) +{ + struct multi_pgsql_simple_query *qry; + + qry = (void *)q; + dprintf(("%s: %p", __func__, qry)); + return qry->eol_at; +} + +static char const *simple_query_get_query(struct multi_pgsql_query *q) +{ + return ((struct multi_pgsql_simple_query *)q)->query; +} + +static void init_simple_query(struct multi_pgsql_simple_query *qry, + struct multi_pgsql_query_vtable *vtable, + unsigned query_lifetime, char const *query) +{ + init_multi_pgsql_query(&qry->super, vtable); + + qry->query = (void *)query; + qry->eol_at = time(NULL) + query_lifetime; +} + +/*** user_query */ +struct multi_pgsql_user_query { + struct multi_pgsql_simple_query super; + + sql_query_callback_t *cb; + void *ctx; +}; + +static void abort_user_query(struct multi_pgsql_query *q, + struct sql_result *r) +{ + struct multi_pgsql_user_query *qry; + + qry = (void *)q; + + T_BEGIN { + qry->cb(r, qry->ctx); + } T_END; +} + +static void user_query_result(struct multi_pgsql_query *q, + struct multi_pgsql_pgc *pgc, + PGresult **pgr) +{ + struct multi_pgsql_result result; + struct multi_pgsql_user_query *qry; + + qry = (void *)q; + init_multi_pgsql_result(&result, *pgr); + + T_BEGIN { + qry->cb(&result.api, qry->ctx); + } T_END; + + if (result.values) i_free(result.values); + + done_with_query(pgc); +} + +static void user_query_dtor(struct multi_pgsql_query *q) +{ + struct multi_pgsql_user_query *qry; + + qry = (void *)q; + + i_free(qry->super.query); + i_free(qry); +} + +/**** vtable/ init */ +static struct multi_pgsql_query_vtable user_query_vtable = { + start_simple_query, + user_query_result, + abort_user_query, + user_query_dtor, + simple_query_eol, + NULL, + simple_query_get_query +}; + +static struct multi_pgsql_query *create_user_query(char const *query, + sql_query_callback_t *cb, void *ctx) +{ + struct multi_pgsql_user_query *qry; + + qry = i_new(struct multi_pgsql_user_query, 1); + + init_simple_query(&qry->super, &user_query_vtable, + USER_QUERY_LIFETIME, i_strdup(query)); + + qry->cb = cb; + qry->ctx = ctx; + return (void *)qry; +} + +/*** sync_query */ +struct multi_pgsql_sync_query { + struct multi_pgsql_simple_query super; + + struct ioloop *ioloop; + PGresult *pgr; + char *errmsg; +}; + +static void sync_query_result(struct multi_pgsql_query *q, + struct multi_pgsql_pgc *pgc, + PGresult **pgr) +{ + struct multi_pgsql_sync_query *qry; + + qry = (void *)q; + + qry->errmsg = NULL; + qry->pgr = *pgr; + *pgr = NULL; + + done_with_query(pgc); +} + +static void abort_sync_query(struct multi_pgsql_query *q, + struct sql_result *r) +{ + struct multi_pgsql_sync_query *qry; + + qry = (void *)q; + + qry->pgr = NULL; + qry->errmsg = i_strdup(r->v.get_error(r)); +} + +static void sync_query_dtor(struct multi_pgsql_query *q) +{ + struct multi_pgsql_sync_query *qry; + + qry = (void *)q; + io_loop_stop(qry->ioloop); +} + +static void set_sync_query_sync_ioloop(struct multi_pgsql_query *q, + struct ioloop *ioloop) +{ + struct multi_pgsql_sync_query *qry; + + qry = (void *)q; + qry->ioloop = ioloop; +} + +static struct sql_result * +result_from_sync_query(struct multi_pgsql_sync_query *qry) +{ + struct sql_result *result; + PGresult *pgr; + + pgr = qry->pgr; + + if (!pgr) { + result = (void *)i_new(struct multi_pgsql_error_result, 1); + init_error_result((void *)result, qry->errmsg); + result->v.free = error_result_free; + + return result; + } + + result = (void *)i_new(struct multi_pgsql_result, 1); + init_multi_pgsql_result((void *)result, pgr); + result->v.free = multi_pgsql_result_free; + + return result; +} + +/**** vtable/ init */ +static struct multi_pgsql_query_vtable sync_query_vtable = { + start_simple_query, + sync_query_result, + abort_sync_query, + sync_query_dtor, + simple_query_eol, + set_sync_query_sync_ioloop, + simple_query_get_query +}; + +static inline void init_sync_query(struct multi_pgsql_sync_query *qry, + char const *query) +{ + init_simple_query(&qry->super, &sync_query_vtable, + USER_QUERY_LIFETIME, query); +} + +/*** transaction query */ +struct multi_pgsql_transaction_subquery { + struct multi_pgsql_transaction_subquery *p; + + char *query; + unsigned *affected_rows; +}; + +struct multi_pgsql_transaction_context { + struct sql_transaction_context super; + struct multi_pgsql_transaction_subquery *first, **link_to; + pool_t pool; +}; + +struct multi_pgsql_transaction_query { + struct multi_pgsql_query super; + struct multi_pgsql_transaction_context *x_ctx; + + struct multi_pgsql_transaction_subquery *sub; + time_t cur_eol; + + /* + If set, points to the PGresult corresponding + with the subquery which caused the transaction + to fail. + */ + PGresult *pgr; +}; + +struct multi_pgsql_transaction_query_vtable { + struct multi_pgsql_query_vtable super; + void (*xact_done)(struct multi_pgsql_transaction_query *, char const *); +}; + +static inline void transaction_query_xact_done(void *q, char const *errmsg) +{ + struct multi_pgsql_transaction_query *qry; + struct multi_pgsql_transaction_query_vtable *vtable; + + dprintf(("%s: %p", __func__, q)); + + qry = q; + vtable = (void *)qry->super.vtable; + vtable->xact_done(q, errmsg); +} + + +static void start_transaction_query(struct multi_pgsql_query *q, + struct multi_pgsql_pgc *pgc) +{ + struct multi_pgsql_transaction_query *qry; + + qry = (void *)q; + qry->sub = qry->x_ctx->first; + start_new_query_on_pgc(q, pgc); +} + +static void transaction_query_result(struct multi_pgsql_query *q, + struct multi_pgsql_pgc *pgc, + PGresult **ppgr) +{ + struct multi_pgsql_transaction_query *qry; + struct multi_pgsql_transaction_subquery *sub; + PGresult *pgr; + char const *errmsg; + ExecStatusType status; + + qry = (void *)q; + + pgr = qry->pgr; + if (pgr) { + errmsg = kill_pg_errmsg_newline(PQresultErrorMessage(pgr)); + + query_done: + transaction_query_xact_done(qry, errmsg); + done_with_query(pgc); + return; + } + + sub = qry->sub; + + pgr = *ppgr; + status = PQresultStatus(pgr); + dprintf(("%s: %p: status %s(%d)", + __func__, qry, PQresStatus(status), status)); + + switch (status) { + case PGRES_COMMAND_OK: + case PGRES_TUPLES_OK: + case PGRES_COPY_OUT: + case PGRES_COPY_IN: + if (sub->affected_rows) + *sub->affected_rows = atoi(PQcmdTuples(pgr)); + + sub = sub->p; + if (!sub) { + errmsg = NULL; + goto query_done; + } + + qry->sub = sub; + qry->cur_eol = time(NULL) + USER_QUERY_LIFETIME; + break; + + default: + qry->pgr = pgr; + *ppgr = NULL; + + qry->cur_eol = 0; /* don't restart failed transactions */ + sub->query = "rollback"; + } + + start_query_on_pgc(pgc); +} + +static void abort_transaction_query(struct multi_pgsql_query *q, + struct sql_result *r) +{ + char const *errmsg; + PGresult *pgr; + + pgr = ((struct multi_pgsql_transaction_query *)q)->pgr; + errmsg = pgr ? + kill_pg_errmsg_newline(PQresultErrorMessage(pgr)) : r->v.get_error(r); + transaction_query_xact_done(q, errmsg); +} + +static time_t transaction_query_eol(struct multi_pgsql_query *q) +{ + return ((struct multi_pgsql_transaction_query *)q)->cur_eol; +} + +static void transaction_query_dtor(struct multi_pgsql_query *q) +{ + struct multi_pgsql_transaction_query *qry; + PGresult *pgr; + pool_t pool; + + qry = (void *)q; + + pgr = qry->pgr; + if (pgr) PQclear(pgr); + + pool = qry->x_ctx->pool; + pool_unref(&pool); +} + +static char const *transaction_query_get_query(struct multi_pgsql_query *q) +{ + struct multi_pgsql_transaction_query *qry; + + qry = (void *)q; + return qry->sub->query; +} + +static void init_transaction_query(struct multi_pgsql_transaction_query *qry, + struct multi_pgsql_transaction_query_vtable *vtable, + struct multi_pgsql_transaction_context *x_ctx) +{ + init_multi_pgsql_query(&qry->super, &vtable->super); + + qry->x_ctx = x_ctx; + qry->cur_eol = time(NULL) + USER_QUERY_LIFETIME; + qry->pgr = NULL; +} + +/*** async transaction query */ +struct multi_pgsql_async_transaction_query { + struct multi_pgsql_transaction_query super; + + sql_commit_callback_t *cb; + void *ctx; +}; + +static void async_transaction_query_xact_done(struct multi_pgsql_transaction_query *q, + char const *errmsg) +{ + struct multi_pgsql_async_transaction_query *qry; + + qry = (void *)q; + qry->cb(errmsg, qry->ctx); +} + +/**** vtable/ init */ +static struct multi_pgsql_transaction_query_vtable async_transaction_query_vtable = { + { + start_transaction_query, + transaction_query_result, + abort_transaction_query, + transaction_query_dtor, + transaction_query_eol, + NULL, + transaction_query_get_query }, + + async_transaction_query_xact_done +}; + +static void +init_async_transaction_query(struct multi_pgsql_async_transaction_query *qry, + struct multi_pgsql_transaction_context *x_ctx, + sql_commit_callback_t *cb, void *cb_ctx) +{ + init_transaction_query(&qry->super, + &async_transaction_query_vtable, x_ctx); + + qry->cb = cb; + qry->ctx = cb_ctx; +} + +/*** sync transaction query */ +struct multi_pgsql_sync_transaction_query { + struct multi_pgsql_transaction_query super; + + struct ioloop *ioloop; + char const *errmsg; +}; + +static void sync_transaction_query_dtor(struct multi_pgsql_query *q) +{ + struct multi_pgsql_sync_transaction_query *qry; + + qry = (void *)q; + io_loop_stop(qry->ioloop); + transaction_query_dtor(q); +} + +static void set_sync_transaction_query_sync_ioloop(struct multi_pgsql_query *q, + struct ioloop *ioloop) +{ + struct multi_pgsql_sync_transaction_query *qry; + + qry = (void *)q; + qry->ioloop = ioloop; +} + +static void sync_transaction_query_xact_done(struct multi_pgsql_transaction_query *q, + char const *errmsg) +{ + struct multi_pgsql_sync_transaction_query *qry; + + qry = (void *)q; + qry->errmsg = i_strdup(errmsg); +} + +/**** vtable/ init */ +static struct multi_pgsql_transaction_query_vtable sync_transaction_query_vtable = { + { + start_transaction_query, + transaction_query_result, + abort_transaction_query, + sync_transaction_query_dtor, + transaction_query_eol, + set_sync_transaction_query_sync_ioloop, + transaction_query_get_query }, + + sync_transaction_query_xact_done +}; + +static void init_sync_transaction_query(struct multi_pgsql_sync_transaction_query *qry, + struct multi_pgsql_transaction_context *x_ctx) +{ + init_transaction_query(&qry->super, + &sync_transaction_query_vtable, x_ctx); +} + +/** pgc (connection) code */ +struct multi_pgsql_pgc { + struct multi_pgsql_pgc *p; + + char *connect_string; + PGconn *pgc; + PGresult *pgr; + struct multi_pgsql_db *pgdb; + struct multi_pgsql_query *qry; + struct io *io; + struct timeout *timeout; +}; + +/*** misc */ +static void start_pgc_connect(struct multi_pgsql_pgc *); + +static void init_pgc(struct multi_pgsql_pgc *pgc, + struct multi_pgsql_db *pgdb, char *connect_string) +{ + pgc->connect_string = connect_string; + pgc->pgdb = pgdb; + + dprintf(("%s: pgc %p, '%s'", __func__, pgc, connect_string)); + + start_pgc_connect(pgc); +} + +static void reset_pgc(struct multi_pgsql_pgc *pgc) +{ + dprintf(("%s: %p", __func__, pgc)); + + if (pgc->io) io_remove(&pgc->io); + if (pgc->timeout) timeout_remove(&pgc->timeout); + + if (pgc->pgr) { + PQclear(pgc->pgr); + pgc->pgr = NULL; + } + + if (pgc->pgc) { + PQfinish(pgc->pgc); + pgc->pgc = NULL; + } +} + +static void pgc_ioloop_switch(struct multi_pgsql_pgc *pgc) +{ + union { + struct io io; + struct timeout timeout; + } old; + + if (pgc->io) { + old.io = *pgc->io; + io_remove(&pgc->io); + + pgc->io = io_add(PQsocket(pgc->pgc), + old.io.condition, old.io.callback, old.io.context); + } + + if (pgc->timeout) { + old.timeout = *pgc->timeout; + timeout_remove(&pgc->timeout); + + pgc->timeout = timeout_add(old.timeout.msecs, + old.timeout.callback, old.timeout.context); + } +} + +static void log_pg_error(struct multi_pgsql_pgc *pgc, char const *fnc, char const *pg_call) +{ + i_error("%s: %p: %s: %s", + fnc, pgc, pg_call, kill_pg_errmsg_newline(PQerrorMessage(pgc->pgc))); +} + +/*** connection establishment */ +static void add_ready_pgc(struct multi_pgsql_db *, struct multi_pgsql_pgc *); + +static void restart_pgc_connect(struct multi_pgsql_pgc *pgc) +{ + timeout_remove(&pgc->timeout); + start_pgc_connect(pgc); +} + +static void continue_pgc_connect(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + char const *pg_call; + PostgresPollingStatusType status; + int io_dir, rc; + + io_remove(&pgc->io); + the_pgc = pgc->pgc; + + status = PQconnectPoll(the_pgc); + dprintf(("%s: %p: PQconnectPoll returned %d", + __func__, pgc, status)); + + switch (status) { + /* + According to the PostgreSQL source, this is an + unused legacy constant. Here to prevent gcc from + complaining about it. + */ + case PGRES_POLLING_ACTIVE: + case PGRES_POLLING_OK: + rc = PQsetnonblocking(the_pgc, 1); + if (rc == -1) { + pg_call = "PQsetnonblocking"; + goto error; + } + + i_info("%s: %p: connected", __func__, pgc); + add_ready_pgc(pgc->pgdb, pgc); + return; + + case PGRES_POLLING_FAILED: + pg_call = "PQconnectPoll"; + goto error; + + case PGRES_POLLING_READING: + io_dir = IO_READ; + break; + + case PGRES_POLLING_WRITING: + io_dir = IO_WRITE; + } + + pgc->io = io_add(PQsocket(the_pgc), io_dir, continue_pgc_connect, pgc); + return; + +error: + log_pg_error(pgc, __func__, pg_call); + + PQfinish(the_pgc); + pgc->pgc = NULL; + + pgc->timeout = timeout_add(RECONNECT_DELAY, + restart_pgc_connect, pgc); +} + +static void start_pgc_connect(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + + dprintf(("%s: %p", __func__, pgc)); + + pgc->pgc = the_pgc = PQconnectStart(pgc->connect_string); + if (!the_pgc) + i_fatal("%s: %p: out of memory", + __func__, pgc); + + if (PQstatus(the_pgc) == CONNECTION_BAD) { + log_pg_error(pgc, __func__, "PQconnectStart"); + pgc->timeout = timeout_add(RECONNECT_DELAY, + restart_pgc_connect, pgc); + return; + } + + pgc->io = io_add(PQsocket(the_pgc), IO_WRITE, + continue_pgc_connect, pgc); +} + +/*** query processing */ +/**** error handling */ +static void requeue_query_to_pgdb(struct multi_pgsql_query *, struct multi_pgsql_db *); + +static void pgc_query_io_timeout(struct multi_pgsql_pgc *pgc) +{ + i_info("%s: %p", __func__, pgc); + + requeue_query_to_pgdb(pgc->qry, pgc->pgdb); + pgc->qry = NULL; + + reset_pgc(pgc); + start_pgc_connect(pgc); +} + +static void pgc_query_processing_failure(struct multi_pgsql_pgc *pgc, + char const *fnc, + char const *pg_call) +{ + log_pg_error(pgc, fnc, pg_call); + + requeue_query_to_pgdb(pgc->qry, pgc->pgdb); + pgc->qry = NULL; + + reset_pgc(pgc); + start_pgc_connect(pgc); +} + +/**** result processing */ +static void done_with_query(struct multi_pgsql_pgc *pgc) +{ + struct multi_pgsql_query *qry; + + qry = pgc->qry; + pgc->qry = NULL; + + destroy_query(qry); +} + +static inline int got_result(PGconn *pgc) +{ + PGresult *pgr; + + pgr = PQgetResult(pgc); + if (!pgr) return 0; + + PQclear(pgr); + return 1; +} + +static void eat_results(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + PGresult *pgr; + int rc; + + dprintf(("%s: %p", __func__, pgc)); + + the_pgc = pgc->pgc; + do { + rc = PQconsumeInput(the_pgc); + if (rc == 0) { + pgc_query_processing_failure(pgc, __func__, + "PQconsumeInput"); + return; + } + + if (PQisBusy(the_pgc)) { + timeout_reset(pgc->timeout); + return; + } + } while (got_result(the_pgc)); + + io_remove(&pgc->io); + timeout_remove(&pgc->timeout); + + pgr = pgc->pgr; + pgc->pgr = NULL; + query_result(pgc->qry, pgc, &pgr); + if (pgr) PQclear(pgr); + + if (!pgc->qry) add_ready_pgc(pgc->pgdb, pgc); +} + +static void get_first_result(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + PGresult *pgr; + int rc; + + dprintf(("%s: %p", __func__, pgc)); + the_pgc = pgc->pgc; + + rc = PQconsumeInput(the_pgc); + if (rc == 0) { + pgc_query_processing_failure(pgc, __func__, + "PQconsumeInput"); + return; + } + + if (PQisBusy(the_pgc)) { + dprintf(("%s: %p: still busy", __func__, pgc)); + + timeout_reset(pgc->timeout); + return; + } + + io_remove(&pgc->io); + + pgr = pgc->pgr = PQgetResult(the_pgc); + if (pgr) { + pgc->io = io_add(PQsocket(the_pgc), IO_READ, eat_results, pgc); + timeout_reset(pgc->timeout); + + eat_results(pgc); + return; + } else + timeout_remove(&pgc->timeout); + + + query_result(pgc->qry, pgc, NULL); + if (!pgc->qry) add_ready_pgc(pgc->pgdb, pgc); +} + +/**** query transmission */ +static void flush_query(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + int rc; + + dprintf(("%s: %p", __func__, pgc)); + the_pgc = pgc->pgc; + + rc = PQflush(the_pgc); + switch (rc) { + case -1: + pgc_query_processing_failure(pgc, __func__, + "PQflush"); + return; + + case 0: + dprintf(("%s: %p: query sent", __func__, pgc)); + + io_remove(&pgc->io); + pgc->io = io_add(PQsocket(the_pgc), IO_READ, get_first_result, pgc); + } + + timeout_reset(pgc->timeout); +} + +static void start_query_on_pgc(struct multi_pgsql_pgc *pgc) +{ + PGconn *the_pgc; + char const *pg_call, *query; + void (*io_cb)(struct multi_pgsql_pgc *); + int rc, io_dir; + + query = get_query(pgc->qry); + dprintf(("%s: %p: %s", __func__, pgc, query)); + + the_pgc = pgc->pgc; + + rc = PQsendQuery(the_pgc, query); + if (rc == 0) { + pg_call = "PQsendQuery"; + goto error; + } + + rc = PQflush(the_pgc); + switch (rc) { + case -1: + pg_call = "PQflush"; + goto error; + + case 0: + io_cb = get_first_result; + io_dir = IO_READ; + break; + + default: + io_cb = flush_query; + io_dir = IO_WRITE; + } + + pgc->io = io_add(PQsocket(the_pgc), io_dir, io_cb, pgc); + pgc->timeout = timeout_add(PGC_IO_TIMEOUT * TIMEOUT_UNIT, + pgc_query_io_timeout, pgc); + return; + +error: + pgc_query_processing_failure(pgc, __func__, "PQflush"); +} + +static void start_new_query_on_pgc(struct multi_pgsql_query *qry, + struct multi_pgsql_pgc *pgc) +{ + dprintf(("%s: %p -> %p", __func__, qry, pgc)); + + pgc->qry = qry; + start_query_on_pgc(pgc); +} + +/** pgdb */ +struct multi_pgsql_db { + struct sql_db api; + + unsigned n_pgcs; + struct multi_pgsql_pgc *pgcs; + + struct { + struct multi_pgsql_pgc *first, **link_to; + } pgc_q; + + struct { + struct timeout *timeout; + struct multi_pgsql_query *first, **link_to; + } query_q; + + pool_t pool; +}; + +static struct multi_pgsql_query *dequeue_query_from_pgdb(struct multi_pgsql_db *); + +/*** pgc support code */ +static void do_init_pgcs(struct multi_pgsql_db *pgdb, char const *s, unsigned ofs) +{ + char const *r, *lws_start; + unsigned c, semi_count; + + while ((c = *s) && isspace(c)) ++s; + + if (!c) { + if (!ofs) i_fatal("%s: empty connect string", __func__); + + dprintf(("%s: end of connect_string, %u pgcs", + __func__, ofs)); + + pgdb->n_pgcs = ofs; + pgdb->pgcs = p_new(pgdb->pool, struct multi_pgsql_pgc, ofs); + return; + } + + semi_count = 0; + lws_start = NULL; + r = s; + do { + if (c == ';') { + if (++semi_count == 3) break; + continue; + } + + semi_count = 0; + + if (isspace(c)) { + if (!lws_start) lws_start = r; + continue; + } + + lws_start = NULL; + } while ((c = *++r)); + + do_init_pgcs(pgdb, c ? r + 1 : r, ofs + 1); + + if (lws_start) r = lws_start; + else if (c) r -= 2; + + init_pgc(pgdb->pgcs + ofs, pgdb, p_strndup(pgdb->pool, s, r - s)); +} + +static inline void init_pgcs(struct multi_pgsql_db *pgdb, char const *connect_string) +{ + do_init_pgcs(pgdb, connect_string, 0); +} + +static void add_ready_pgc(struct multi_pgsql_db *pgdb, struct multi_pgsql_pgc *pgc) +{ + struct multi_pgsql_query *qry; + + if (!pgdb->pgc_q.first) { + qry = dequeue_query_from_pgdb(pgdb); + + if (qry) { + start_query(qry, pgc); + return; + } + } + + *pgdb->pgc_q.link_to = pgc; + pgdb->pgc_q.link_to = &pgc->p; + + dprintf(("%s: %p -> %p", __func__, pgc, pgdb)); +} + +static struct multi_pgsql_pgc *get_ready_pgc(struct multi_pgsql_db *pgdb) +{ + struct multi_pgsql_pgc *pgc, *next_pgc; + + pgc = pgdb->pgc_q.first; + if (!pgc) return NULL; + + next_pgc = pgc->p; + if (next_pgc) pgc->p = NULL; + else pgdb->pgc_q.link_to = &pgdb->pgc_q.first; + pgdb->pgc_q.first = next_pgc; + + dprintf(("%s: %p -> %p", __func__, pgdb, pgc)); + return pgc; +} + +static void switch_pgdb_pgcs(struct multi_pgsql_db *pgdb) +{ + struct multi_pgsql_pgc *pgcs; + unsigned n; + + pgcs = pgdb->pgcs; + n = pgdb->n_pgcs; + do pgc_ioloop_switch(pgcs + --n); while (n); +} + +/*** query timeout */ +static void start_pgdb_query_timeout(struct multi_pgsql_db *, + struct multi_pgsql_query *); + +static void pgdb_query_timeout(struct multi_pgsql_db *pgdb) +{ + struct multi_pgsql_query *qry, *next_qry; + struct multi_pgsql_error_result timeout; + time_t now; + + dprintf(("%s: %p", __func__, pgdb)); + + timeout_remove(&pgdb->query_q.timeout); + qry = pgdb->query_q.first; + now = time(NULL); + init_error_result(&timeout, TIMEOUT_ERROR); + do { + next_qry = qry->p; + abort_query(qry, &timeout.api); + } while (next_qry && now > query_eol(next_qry)); + + pgdb->query_q.first = next_qry; + + if (next_qry) { + start_pgdb_query_timeout(pgdb, next_qry); + return; + } + + pgdb->query_q.link_to = &pgdb->query_q.first; +} + +static void start_pgdb_query_timeout(struct multi_pgsql_db *pgdb, + struct multi_pgsql_query *qry) +{ + time_t now, timeout, eol; + + now = time(NULL); + eol = query_eol(qry); + timeout = now < eol ? eol - now : 0; + if (timeout < MIN_QUERY_TIMEOUT) timeout = MIN_QUERY_TIMEOUT; + + pgdb->query_q.timeout = timeout_add(timeout * TIMEOUT_UNIT, + pgdb_query_timeout, pgdb); + dprintf(("%s: %p: timeout in %us", __func__, pgdb, (unsigned)timeout)); +} + +/*** query queueing */ +static void queue_query_to_pgdb(struct multi_pgsql_query *qry, struct multi_pgsql_db *pgdb) +{ + if (!pgdb->query_q.first) start_pgdb_query_timeout(pgdb, qry); + + *pgdb->query_q.link_to = qry; + pgdb->query_q.link_to = &qry->p; +} + +static struct multi_pgsql_query *dequeue_query_from_pgdb(struct multi_pgsql_db *pgdb) +{ + struct multi_pgsql_query *qry, *next_qry; + + qry = pgdb->query_q.first; + if (!qry) return NULL; + + timeout_remove(&pgdb->query_q.timeout); + + next_qry = qry->p; + if (next_qry) start_pgdb_query_timeout(pgdb, next_qry); + else pgdb->query_q.link_to = &pgdb->query_q.first; + pgdb->query_q.first = next_qry; + + qry->p = NULL; + return qry; +} + +static void requeue_query_to_pgdb(struct multi_pgsql_query *qry, + struct multi_pgsql_db *pgdb) +{ + struct multi_pgsql_error_result timeout; + struct multi_pgsql_pgc *pgc; + + if (time(NULL) >= query_eol(qry)) { + init_error_result(&timeout, TIMEOUT_ERROR); + abort_query(qry, &timeout.api); + return; + } + + pgc = get_ready_pgc(pgdb); + if (pgc) { + start_query(qry, pgc); + return; + } + + if (pgdb->query_q.timeout) { + timeout_remove(&pgdb->query_q.timeout); + qry->p = pgdb->query_q.first; + } else + pgdb->query_q.link_to = &qry->p; + + pgdb->query_q.first = qry; + start_pgdb_query_timeout(pgdb, qry); +} + + +/*** query submission */ +static void pgdb_async_query(struct multi_pgsql_db *pgdb, + struct multi_pgsql_query *qry) +{ + struct multi_pgsql_pgc *pgc; + + dprintf(("%s: %p: %p", __func__, pgdb, qry)); + + pgc = get_ready_pgc(pgdb); + if (pgc) { + start_query(qry, pgc); + return; + } + + dprintf(("%s: %p: no pgcs ready, queueing", __func__, pgdb)); + queue_query_to_pgdb(qry, pgdb); +} + + +static void pgdb_sync_query(struct multi_pgsql_db *pgdb, + struct multi_pgsql_query *qry) +{ + struct ioloop *old_ioloop, *sync_ioloop; + + dprintf(("%s: %p: %p", __func__, pgdb, qry)); + + old_ioloop = current_ioloop; + sync_ioloop = io_loop_create(); + switch_pgdb_pgcs(pgdb); + + set_query_sync_ioloop(qry, sync_ioloop); + pgdb_async_query(pgdb, qry); + + io_loop_run(sync_ioloop); + + io_loop_set_current(old_ioloop); + switch_pgdb_pgcs(pgdb); + io_loop_set_current(sync_ioloop); + io_loop_destroy(&sync_ioloop); +} + +/*** non-transaction SQL driver methods */ +extern struct sql_db driver_multi_pgsql_db; + +static struct sql_db *multi_pgsql_init_v(char const *connect_string) +{ + struct multi_pgsql_db *pgdb; + pool_t pool; + + dprintf(("%s: %s", __func__, connect_string)); + + pool = pool_alloconly_create("multi_pgsql_pool", MULTI_PGSQL_POOL); + + pgdb = p_new(pool, struct multi_pgsql_db, 1); + pgdb->api = driver_multi_pgsql_db; + pgdb->pool = pool; + + pgdb->query_q.link_to = &pgdb->query_q.first; + pgdb->pgc_q.link_to = &pgdb->pgc_q.first; + + init_pgcs(pgdb, connect_string); + return (void *)pgdb; +} + +static void multi_pgsql_deinit_v(struct sql_db *db) +{ + struct multi_pgsql_error_result deinit; + struct multi_pgsql_db *pgdb; + struct multi_pgsql_pgc *pgc; + struct multi_pgsql_query *qry, *next_qry; + pool_t pool; + unsigned n; + + dprintf(("%s: %p", __func__, db)); + + array_free(&db->module_contexts); + + pgdb = (void *)db; + init_error_result(&deinit, DEINIT_ERROR); + + n = pgdb->n_pgcs; + do { + pgc = pgdb->pgcs + --n; + + qry = pgc->qry; + if (qry) abort_query(qry, &deinit.api); + + reset_pgc(pgc); + + } while (n); + + qry = pgdb->query_q.first; + if (qry) { + timeout_remove(&pgdb->query_q.timeout); + + do { + next_qry = qry->p; + abort_query(qry, &deinit.api); + } while ((qry = next_qry)); + } + + pool = pgdb->pool; + pool_unref(&pool); +} + +static int multi_pgsql_connect(struct sql_db *db) +{ + dprintf(("%s: %p", __func__, db)); + return 1; +} + +static enum sql_db_flags +multi_pgsql_get_flags(struct sql_db *db) +{ + dprintf(("%s: %p", __func__, db)); + return 0; +} + +static const char * +multi_pgsql_escape_string(struct sql_db *db, const char *in) +{ + struct multi_pgsql_db *pgdb; + char *out; + size_t len; + + pgdb = (void *)db; + len = strlen(in); + out = t_buffer_get(len * 2 + 1); + len = PQescapeStringConn(pgdb->pgcs->pgc, out, in, len, + NULL); + t_buffer_alloc(len + 1); + + dprintf(("%s: %p: '%s' -> '%s'", + __func__, pgdb, in, out)); + return out; +} + +static void multi_pgsql_query(struct sql_db *db, const char *query, + sql_query_callback_t *cb, void *ctx) +{ + struct multi_pgsql_query *qry; + + dprintf(("%s: %p: %s", __func__, db, query)); + + qry = create_user_query(query, cb, ctx); + pgdb_async_query((void *)db, qry); +} + +static void multi_pgsql_exec(struct sql_db *db, const char *query) +{ + i_info("%s: %p: %s", __func__, db, query); + i_info("%s: not implemented", __func__); +} + +static struct sql_result * +multi_pgsql_query_s(struct sql_db *db, const char *query) +{ + struct multi_pgsql_db *pgdb; + struct multi_pgsql_sync_query sync_qry; + + pgdb = (void *)db; + dprintf(("%s: %p: %s", __func__, pgdb, query)); + + init_sync_query(&sync_qry, query); + pgdb_sync_query(pgdb, (void *)&sync_qry); + return result_from_sync_query(&sync_qry); +} + +/*** transaction support code */ +static void add_subquery_to(struct multi_pgsql_transaction_context *x_ctx, + char const *query, unsigned *affected_rows) + +{ + struct multi_pgsql_transaction_subquery *sub; + pool_t pool; + + pool = x_ctx->pool; + + sub = p_new(pool, struct multi_pgsql_transaction_subquery, 1); + sub->query = p_strdup(pool, query); + sub->affected_rows = affected_rows; + + *x_ctx->link_to = sub; + x_ctx->link_to = &sub->p; + + dprintf(("%s: %p: '%s'", __func__, x_ctx, query)); +} + +static struct sql_transaction_context * +multi_pgsql_transaction_begin(struct sql_db *db) +{ + struct multi_pgsql_transaction_context *x_ctx; + pool_t pool; + + pool = pool_alloconly_create("multi_pgsql_xact_pool", + MULTI_PGSQL_XACT_POOL); + x_ctx = p_new(pool, struct multi_pgsql_transaction_context, 1); + x_ctx->super.db = db; + x_ctx->link_to = &x_ctx->first; + x_ctx->pool = pool; + + add_subquery_to(x_ctx, "begin", NULL); + + dprintf(("%s: db %p: %p", __func__, db, x_ctx)); + return (void *)x_ctx; +} + +static void +multi_pgsql_transaction_commit(struct sql_transaction_context *ctx, + sql_commit_callback_t *cb, void *cb_ctx) +{ + struct multi_pgsql_transaction_context *x_ctx; + struct multi_pgsql_async_transaction_query *qry; + + x_ctx = (void *)ctx; + dprintf(("%s: %p", __func__, x_ctx)); + + add_subquery_to(x_ctx, "commit", NULL); + + qry = p_new(x_ctx->pool, struct multi_pgsql_async_transaction_query, 1); + init_async_transaction_query(qry, x_ctx, cb, cb_ctx); + pgdb_async_query((void *)ctx->db, (void *)qry); +} + +static int +multi_pgsql_transaction_commit_s(struct sql_transaction_context *ctx, + const char **error_r) +{ + struct multi_pgsql_transaction_context *x_ctx; + struct multi_pgsql_sync_transaction_query sync_qry; + char *errmsg; + + x_ctx = (void *)ctx; + dprintf(("%s: %p", __func__, x_ctx)); + + add_subquery_to(x_ctx, "commit", NULL); + + init_sync_transaction_query(&sync_qry, x_ctx); + pgdb_sync_query((void *)x_ctx->super.db, &sync_qry.super.super); + + errmsg = (char *)sync_qry.errmsg; + if (errmsg) { + *error_r = t_strdup(errmsg); + i_free(errmsg); + + return -1; + } + + return 0; +} + +static void +multi_pgsql_transaction_rollback(struct sql_transaction_context *ctx) +{ + struct multi_pgsql_transaction_context *x_ctx; + + x_ctx = (void *)ctx; + dprintf(("%s: %p", __func__, x_ctx)); + + pool_unref(&x_ctx->pool); +} + +static void +multi_pgsql_transaction_update(struct sql_transaction_context *ctx, const char *query, + unsigned int *affected_rows) +{ + struct multi_pgsql_transaction_context *x_ctx; + + x_ctx = (void *)ctx; + add_subquery_to(x_ctx, query, affected_rows); +} + +/*** db driver vtable */ +struct sql_db driver_multi_pgsql_db = { + "multi-pgsql", + + MEMBER(v) { + multi_pgsql_init_v, + multi_pgsql_deinit_v, + multi_pgsql_get_flags, + multi_pgsql_connect, + multi_pgsql_escape_string, + multi_pgsql_exec, + multi_pgsql_query, + multi_pgsql_query_s, + + multi_pgsql_transaction_begin, + multi_pgsql_transaction_commit, + multi_pgsql_transaction_commit_s, + multi_pgsql_transaction_rollback, + multi_pgsql_transaction_update + } +}; +#endif

3 4

[Dovecot] windows live mail + dovecot and nfs
by Anes Muhametov 17 Apr '10

17 Apr '10

Hi all. My clients have got problem accessing dovecot by "Windows live mail (Windows 7)" client. The problem is that users can't delete two or more folders one-by-one. There is no errors deleting first folder (new dir /storage/vol1/mail/domain/user/Maildir/..DOVECOT-TRASHED appeared and wasn't deleted). If I try to delete second one, error occurs (user, domain, IP address changed): dovecot: IMAP(user@domain): unlink_directory(/storage/vol1/mail/domain/user/Maildir/..DOVECOT-TRASHED) failed: Directory not empty I'm using dovecot 1.2.10 @FreeBSD 7.2-RELEASE-p1 amd64 and NFS storage backend. dovecot -n # 1.2.10: /usr/local/etc/dovecot.conf # OS: FreeBSD 7.2-RELEASE-p1 amd64 syslog_facility: local7 protocols: imap imaps pop3 pop3s ssl_cert_file: /etc/ssl/1.crt ssl_key_file: /etc/ssl/1.key.unsec disable_plaintext_auth: no verbose_ssl: yes shutdown_clients: no login_dir: /var/run/dovecot/login login_executable(default): /usr/local/libexec/dovecot/imap-login login_executable(imap): /usr/local/libexec/dovecot/imap-login login_executable(pop3): /usr/local/libexec/dovecot/pop3-login login_greeting: mail server ready. login_process_per_connection: no login_process_size: 512 login_processes_count: 60 login_max_processes_count: 256 login_max_connections: 100 max_mail_processes: 4096 mail_max_userip_connections(default): 50 mail_max_userip_connections(imap): 50 mail_max_userip_connections(pop3): 30 verbose_proctitle: yes first_valid_uid: 90 last_valid_uid: 90 first_valid_gid: 90 last_valid_gid: 90 mail_privileged_group: mail mail_location: maildir:%h/Maildir mmap_disable: yes mail_nfs_storage: yes mail_nfs_index: yes lock_method: dotlock maildir_copy_preserve_filename: yes mail_executable(default): /usr/local/libexec/dovecot/imap mail_executable(imap): /usr/local/libexec/dovecot/imap mail_executable(pop3): /usr/local/libexec/dovecot/pop3 mail_process_size: 128 mail_plugins(default): quota imap_quota mail_plugins(imap): quota imap_quota mail_plugins(pop3): quota mail_plugin_dir(default): /usr/local/lib/dovecot/imap mail_plugin_dir(imap): /usr/local/lib/dovecot/imap mail_plugin_dir(pop3): /usr/local/lib/dovecot/pop3 imap_client_workarounds(default): delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep imap_client_workarounds(imap): delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep imap_client_workarounds(pop3): pop3_client_workarounds(default): pop3_client_workarounds(imap): pop3_client_workarounds(pop3): outlook-no-nuls oe-ns-eoh lda: postmaster_address: postmaster(a)domain.ru mail_plugins: quota sendmail_path: /usr/sbin/sendmail log_path: /var/log/dovecot/deliver.log info_log_path: /var/log/dovecot/deliver.log auth default: mechanisms: plain login cram-md5 cache_size: 262144 cache_ttl: 180 cache_negative_ttl: 60 failure_delay: 1 worker_max_count: 2000 process_size: 512 passdb: driver: sql args: /usr/local/etc/dovecot-sql.conf userdb: driver: prefetch userdb: driver: sql args: /usr/local/etc/dovecot-sql.conf socket: type: listen master: path: /var/run/dovecot/auth-master mode: 384 user: user group: domain plugin: quota: maildir:Quota Here is listing of ..DOVECOT-TRASHED: ls -la /storage/vol1/mail/domain/user/Maildir/..DOVECOT-TRASHED: total 16 drwx------ 2 user user 4096 Apr 9 11:32 . drwx------ 27 user user 8192 Apr 9 11:41 .. -rw------- 1 user user 156 Apr 9 11:32 .nfs.014cd02e.ffffd11b4.4 -rw------- 1 user user 17 Apr 9 11:32 .nfs.014cd031.ffffd11b4.4 fstat output: [root@ /storage/vol1/mail/domain/user/Maildir/..DOVECOT-TRASHED]# fstat .* USER CMD PID FD MOUNT INUM MODE SZ|DV R/W NAME root fstat 82081 wd /storage/vol1 32454159 drwx------ 4096 r . user imap 53042 7 /storage/vol1 40573622 -rw------- 156 rw .nfs.014cd02e.ffffd11b4.4 user imap 53042 9 /storage/vol1 40573631 -rw------- 17 rw .nfs.014cd031.ffffd11b4.4 root bash 22223 wd /storage/vol1 32454159 drwx------ 4096 r . # ps uxwa|grep user user 53042 0.0 0.1 9812 2400 ?? S 11:31AM 0:00.02 imap [user@domain 1.1.1.1] user 53650 0.0 0.1 10836 2908 ?? S 11:32AM 0:00.05 imap [user@domain 1.1.1.1] It seems like first imap process locked ..DOVECOT-TRASHED. Best regards Anes Muhametoc.

2 2

[Dovecot] get the messages ID after APPEND
by Christian Fischer 17 Apr '10

17 Apr '10

Hi there! I am new to imap "Command Line" and trying to write a script doing the following on a dovecot imap server: - APPEND a message (e.g. to Drafts) - get the ID to reopen the message for reading In short: How do i get the ID back from the APPEND command or which command has to be invoked to get the uid of the newly added message? Thank you! Christian ----------------- dovecot --version 1.0.15 dovecot -n # 1.0.15: /etc/dovecot/dovecot.conf log_timestamp: %Y-%m-%d %H:%M:%S ssl_cert_file: /etc/postfix/tls/mail.cert ssl_key_file: /etc/postfix/tls/mail.key disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable: /usr/lib/dovecot/imap-login mail_privileged_group: mail mail_location: maildir:%h/.EMAIL auth default: mechanisms: plain login passdb: driver: pam userdb: driver: passwd socket: type: listen client: path: /var/spool/postfix/private/auth mode: 432 user: postfix group: postfix

3 3

[Dovecot] mysql log entry
by Terry Barnum 17 Apr '10

17 Apr '10

I'm sure this is a really newbie question, but I notice in /var/log/mail.log that there are between 3 and 15 of the following log entries every day for the last five days, which is far back as my logs go--what do they mean? Apr xx hh:mm:ss mail dovecot[109]: auth-worker(default): mysql: Connected to localhost (postfix) Is it dovecot connecting to mysql for password or user info? If so, wouldn't it be more frequent? I've got 20 users with some connecting from outside many times per day. Thanks, -Terry

2 2

[Dovecot] Problems with masteruser
by Неворотин Вадим 16 Apr '10

16 Apr '10

I have very strange problem with masteruser. See two logs below: # telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR LOGIN-REFERRALS ID ENABLE STARTTLS AUTH=PLAIN AUTH=LOGIN] Welcome to our post server! x login nevorotin password x OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR LOGIN-REFERRALS ID ENABLE SORT SORT=DISPLAY THREAD=REFERENCES THREAD=REFS MULTIAPPEND UNSELECT IDLE CHILDREN NAMESPACE UIDPLUS LIST-EXTENDED I18NLEVEL=1 CONDSTORE QRESYNC ESEARCH ESORT SEARCHRES WITHIN CONTEXT=SEARCH LIST-STATUS ACL RIGHTS=texk] Logged in x list "" "*" * LIST (\HasNoChildren) "/" "INBOX" x OK List completed. x getacl INBOX * ACL "INBOX" "nevorotin" lrwstipekxacd x OK Getacl completed. All work perfectly. And then I log in throw masteruser: # telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR LOGIN-REFERRALS ID ENABLE STARTTLS AUTH=PLAIN AUTH=LOGIN] Welcome to our post server! x login nevorotin*master masterpassword x OK [CAPABILITY IMAP4rev1 LITERAL+ SASL-IR LOGIN-REFERRALS ID ENABLE SORT SORT=DISPLAY THREAD=REFERENCES THREAD=REFS MULTIAPPEND UNSELECT IDLE CHILDREN NAMESPACE UIDPLUS LIST-EXTENDED I18NLEVEL=1 CONDSTORE QRESYNC ESEARCH ESORT SEARCHRES WITHIN CONTEXT=SEARCH LIST-STATUS ACL RIGHTS=texk] Logged in x list "" "*" * LIST (\Unmarked) "/" "INBOX" x OK List completed. x getacl INBOX x NO [NONEXISTENT] Mailbox doesn't exist: INBOX I've turned on debug logging, but there aren't any errors. I only see that masteruser succesfully logged in as nevorotin. How can I make a masteruser login to user account exactly the same as simple user login? Now it don't work at all [?] I use 1.2.10 version of dovecot.

3 10

Re: [Dovecot] zlib Plugin Dovecot 2.0
by Alex Baule 16 Apr '10

16 Apr '10

Hi Paschal You do not understand what I want. I developed a plugin based on the zlib plugin, but want to port it to dovecot 2.0. There are some differences in the 2.0 plugin, which I did not quite understand how it works for me to readjust it in my plugin. As in the source code there are not many comments, I do not quite know what each function does, if they are all necessary in my adaptation. In developing the plugin in version 1.4, Timo helped me a lot. What I need to know is how the hooks on write emails do this, and what the differences in the read function, since the code is different from version 1.4 to 2.0.

2 6

[Dovecot] 2.0beta4 and latest Mercurial: Folder not SELECT-able until LIST-ed
by B. Johannessen 16 Apr '10

16 Apr '10

Have a look at the attached dialog. It shows the folder "shared/a(a)db.org/Archive" not being SELECT-able until after it's been mentioned in a LIST response. This can't possibly be right, can it? Bob

2 4

[Dovecot] flags, keywords in a shared mailbox setup
by Thomas Hummel 16 Apr '10

16 Apr '10

Hello Timo, I'm trying to get a clear understanding of dovecot's (1.2.x) flags and keywords sharing possibilities or impossibilities in a shared-mailboxes setup. By definition, a shared mailbox is a mailbox from a personal namespace which has been made available to someone else who sees it in a shared namespace. There can be some combinatorics according to the fact that INDEX and CONTROL files location can either be user-dependent or user-independent and this in either the private or the shared namespace. Besides, it's not clear to me if the 'dovecot-shared' file is still relevant in 1.2x and shared-namespaces to store flags only in indexes. So what combination is to be avoided because it would break things ? More specifically : . is, in 1.2x, the method for having private flags of a shared mailbox still to have the 'dovecot-shared' file (in order to store flags in indexes) ? . how can shared keywords in a shared mailbox even work ? If they're shared (how ? by pointing CONTROL to the same user-independent location ?), keywords known by a client (user1) won't necessary be known or have the same meaning for another client (user2), will they ? [Note : isn't it the same event with no shared mailboxes but with one user using several clients ?] But, if they're private, they have to be stored outside the filename : how can we do that ? Thanks -- Thomas Hummel | Institut Pasteur <hummel(a)pasteur.fr> | Pôle informatique - systèmes et réseau

2 3

[Dovecot] dovecot-auth question
by rmcgraw＠math.purdue.edu 16 Apr '10

16 Apr '10

/export/rpoolbackup# dovecot -n # 1.2.10: /local/etc/dovecot.conf Warning: fd limit 256 is lower than what Dovecot can use under full load (more than 768). Either grow the limit or change login_max_processes_count and max_mail_processes settings # OS: SunOS 5.10 i86pc protocols: imaps pop3s ssl_cert_file: /local/ssl/certs/dovecot-cert.pem ssl_key_file: /local/ssl/private/dovecot-key.pem verbose_ssl: yes login_dir: /local/var/run/dovecot/login login_executable(default): /local/libexec/dovecot/imap-login login_executable(imap): /local/libexec/dovecot/imap-login login_executable(pop3): /local/libexec/dovecot/pop3-login first_valid_uid: 101 mail_location: mbox:~/mail:INBOX=/var/mail/%u mail_executable(default): /local/libexec/dovecot/imap mail_executable(imap): /local/libexec/dovecot/imap mail_executable(pop3): /local/libexec/dovecot/pop3 mail_plugin_dir(default): /local/lib/dovecot/imap mail_plugin_dir(imap): /local/lib/dovecot/imap mail_plugin_dir(pop3): /local/lib/dovecot/pop3 auth default: verbose: yes debug: yes passdb: driver: pam userdb: driver: passwd ----- We recently moved to dovecot and after few little configuration problem our dovecot seems to be running without problems for the most part, but I am seeing the following error in /var/adm/messages about ever 3 minutes. Apr 5 09:13:23 gauss.math.purdue.edu dovecot-auth[14359]: [ID 293258 user.error] libsldap: Status: 49 Mesg: openConnection: simple bind failed - Invalid credentials I would like to try and find out what user connection is causing this message. I have turned on most of the verbose logging but I cannot see anything in any logs that might give me the information. Can someone give me guidance on how I can track this down. Thanks Robert _____________________________________________________________________ Robert P. McGraw, Jr. Manager, Computer System EMAIL: rmcgraw(a)purdue.edu Purdue University ROOM: MATH-807 Department of Mathematics PHONE: (765) 494-6055 150 N. University Street West Lafayette, IN 47907-2067

2 1

[Dovecot] dovecot 1.2.11 + virtual plugin + deliver
by Stef Simoens 16 Apr '10

16 Apr '10

Hello List and Timo, I configured the virtual mailboxes which are really nice. To create the virtual-folder, I made a postlogin-script as described in the wiki. However, as my mail is delivered by dovecot's deliver ... I'm getting errors for people not yet logged in... Namespace 'virtual.': virtual : Root mail directory doesn't exist: /var/vmail/domain.org/username/virtual I tried removing the virtual-plugin from the protocol lda-section; and putting the namespace only in the protocol imap-section; but then the error becomes Namespace 'virtual.': Unknown mail storage driver virtual After an IMAP login (thanks to the postlogin-script) the virtual folder is created. However, I would prefer that mails to existing users and new users works even before they logged-in (and before the postlogin-script creates the virtual folder). Any help? Greetings, -- Stef Simoens

3 3

[Dovecot] Per-user flags for shared folders
by Неворотин Вадим 16 Apr '10

16 Apr '10

As I see in 2.0 there willn't be ability to control which flags should be shared, and which should be per-user. But may be there is some plans to implement settings to control that? It'll be great if you can write in dovecot-shared file which flags should be per-user. It's very usefull for group mailboxes, when each user should work with such mailbox like with their own. Now I can't add permission for delete on group mailbox for anybody, because if somebody delete a message, it dissapear for other users of group mailbox. So my group mailboxes look like a big scrapyard[?] I can compile from sources and add also \Deleted flag to private flags, but it's not an elegant solution[?] And my idea is easy to implement :) Only add here: 305 shared_path = t_strconcat(box->path, "/dovecot-shared", NULL); 306 if (stat(shared_path, &st) == 0) 307 box->private_flags_mask = MAIL_SEEN; in maildir-storage.c some code for reading a mask from dovecot-shared file. Unfortunatelly I'm not a good programmer, so I can't make a patch myself(((

2 1

[Dovecot] 2.0beta4 doesn't respond to DONE after untagged FETCH during IDLE
by Daniel I. Applebaum 16 Apr '10

16 Apr '10

Using K-9 Mail to IDLE on a test folder "MyIncTmp". When there is a flag change on a message in the folder, which generates an untagged FETCH, K-9 Mail responds with a DONE, but dovecot does not exit IDLE state (LOG 1). However, if a new message is delivered, generating an untagged EXISTS, K-9 Mail responds the same way, with a DONE, and dovecot does exit IDLE state. (LOG 2) You can also see in LOG 1 that the "OK Still here" continues to be sent, corroborating that dovecot is still in IDLE state. I hope this information is helpful. Sincerely, Dan. LOG 1: failure to exit IDLE after untagged FETCH 04-16 08:25:48.911: VERBOSE/k9(10160): Sending IMAP command IDLE on connection conn1127686952 04-16 08:25:48.921: VERBOSE/k9(10160): conn1127686952>>> 13 IDLE 04-16 08:25:48.921: VERBOSE/k9(10160): Sent IMAP command IDLE with tag 13 for conn1127686952 04-16 08:25:49.191: VERBOSE/k9(10160): <<< #+# [idling] 04-16 08:25:49.201: VERBOSE/k9(10160): conn1127686952<<<#+# [idling] 04-16 08:25:49.201: VERBOSE/k9(10160): Got async response: #+# [idling] 04-16 08:25:49.201: DEBUG/k9(10160): Idling dantest:MyIncTmp/Thread-24/conn1127686952 04-16 08:25:49.201: DEBUG/k9(10160): Considering releasing WakeLock for Pushing 04-16 08:25:49.201: DEBUG/k9(10160): Releasing WakeLock for Pushing for thread Thread-24 04-16 08:25:54.491: VERBOSE/k9(10160): <<< #null# [16, FETCH, [FLAGS, [\Seen, \Recent]]] 04-16 08:25:54.491: VERBOSE/k9(10160): conn1127686952<<<#null# [16, FETCH, [FLAGS, [\Seen, \Recent]]] 04-16 08:25:54.491: VERBOSE/k9(10160): Got async response: #null# [16, FETCH, [FLAGS, [\Seen, \Recent]]] 04-16 08:25:54.501: DEBUG/k9(10160): Acquired WakeLock for Pushing for thread Thread-24 04-16 08:25:54.501: DEBUG/k9(10160): Got useful async untagged response: #null# [16, FETCH, [FLAGS, [\Seen, \Recent]]] for dantest:MyIncTmp/Thread-24/conn1127686952 04-16 08:25:54.501: VERBOSE/k9(10160): conn1127686952>>> DONE 04-16 08:26:54.571: DEBUG/k9(10160): Acquired WakeLock for Pushing for thread Thread-24 04-16 08:26:54.581: ERROR/k9(10160): Error 'Read timed out' 04-16 08:26:54.581: ERROR/k9(10160): java.net.SocketTimeoutException: Read timed out 04-16 08:27:00.141: VERBOSE/k9(10160): Sending IMAP command NOOP on connection conn1127686952 04-16 08:27:00.141: VERBOSE/k9(10160): conn1127686952>>> 14 NOOP 04-16 08:27:00.141: VERBOSE/k9(10160): Sent IMAP command NOOP with tag 14 for conn1127686952 04-16 08:27:26.651: VERBOSE/k9(10160): conn1127686952<<<#null# [OK, Still, here] LOG 2: successful exit from IDLE after untagged EXISTS 04-16 08:28:35.271: INFO/k9(10160): About to IDLE for dantest:MyIncTmp/Thread-24/conn1126759992 04-16 08:28:35.271: VERBOSE/k9(10160): Sending IMAP command IDLE on connection conn1126759992 04-16 08:28:35.281: VERBOSE/k9(10160): conn1126759992>>> 10 IDLE 04-16 08:28:35.281: VERBOSE/k9(10160): Sent IMAP command IDLE with tag 10 for conn1126759992 04-16 08:28:35.531: VERBOSE/k9(10160): <<< #+# [idling] 04-16 08:28:35.531: VERBOSE/k9(10160): conn1126759992<<<#+# [idling] 04-16 08:28:35.531: VERBOSE/k9(10160): Got async response: #+# [idling] 04-16 08:28:35.531: DEBUG/k9(10160): Idling dantest:MyIncTmp/Thread-24/conn1126759992 04-16 08:28:35.531: DEBUG/k9(10160): Considering releasing WakeLock for Pushing 04-16 08:28:35.531: DEBUG/k9(10160): Releasing WakeLock for Pushing for thread Thread-24 04-16 08:28:51.621: VERBOSE/k9(10160): <<< #null# [18, EXISTS] 04-16 08:28:51.621: VERBOSE/k9(10160): conn1126759992<<<#null# [18, EXISTS] 04-16 08:28:51.621: VERBOSE/k9(10160): Got async response: #null# [18, EXISTS] 04-16 08:28:51.621: DEBUG/k9(10160): Acquired WakeLock for Pushing for thread Thread-24 04-16 08:28:51.631: DEBUG/k9(10160): Got useful async untagged response: #null# [18, EXISTS] for dantest:MyIncTmp/Thread-24/conn1126759992 04-16 08:28:51.631: VERBOSE/k9(10160): conn1126759992>>> DONE 04-16 08:28:51.631: VERBOSE/k9(10160): <<< #null# [1, RECENT] 04-16 08:28:51.631: VERBOSE/k9(10160): conn1126759992<<<#null# [1, RECENT] 04-16 08:28:51.631: VERBOSE/k9(10160): Got async response: #null# [1, RECENT] 04-16 08:28:51.901: VERBOSE/k9(10160): <<< #10# [OK, Idle, completed.] 04-16 08:28:51.901: VERBOSE/k9(10160): conn1126759992<<<#10# [OK, Idle, completed.] 04-16 08:28:51.901: VERBOSE/k9(10160): Got async response: #10# [OK, Idle, completed.]

2 2

[Dovecot] Dovecot 2.0 beta 4 errors reported
by Ken A 16 Apr '10

16 Apr '10

Timo, Thank you for an amazing imap/pop server. I'm using Dovecot 2.0 beta 4 and mbox storage, and have encountered a couple of errors worth reporting: ------ 1. When I delete a parent folder that contains sub folders, I get this error message pop up in Thunderbird 3.04: > "The current command did not succeed. The mail server responded: > [CANNOT] Mailbox GUIDs are not permanent without index files." But, the folders are deleted properly, and no error is logged on the server. I have the mail_log plugin currently enabled like so: > mail_plugins = quota mail_log notify > mail_log_events = delete undelete expunge copy mailbox_delete mailbox_rename > mail_log_fields = uid box msgid size ------ 2. segfault reported in the system log (unrelated to the error above): > kernel: pop3[6794]: segfault at 0000000000000028 rip 0000000000403489 > rsp 00007fff20a24a20 error 4 ------ Thanks, Ken -- Ken Anderson Pacific Internet - http://www.pacific.net

2 1

[Dovecot] 2.0.beta4 "OK Still here" not flushed?
by Daniel I. Applebaum 16 Apr '10

16 Apr '10

Hello, everybody! I've noticed that my installation of K-9 Mail on Android has been consuming a lot of battery power recently. I've been running it connecting to 2.0.beta3 for the past few weeks and just upgraded to 2.0.beta4 while debugging this issue. There is a conspicuous lack of "OK Still here" messages in my IDLE connections. The lack of these "OK Still here" messages means that the idling connections get dropped very frequently, which causes K-9 Mail to expend much more effort than just receiving the periodic "OK Still here" and going back to sleep. The "OK Still here" messages sometimes arrive, however, just before an untagged FETCH, however. I do not know anything about dovecot internals, but I suspected that the "OK Still here" messages were not being flushed, so I modified line 270 of src/imap/imap-client.c to be: if (1 || o_stream_get_buffer_used_size(client->output) >= which I am sure is the wrong solution, since it forces a flush for every output line and would be quite inefficient. However, with that modification, the "OK Still here" messages appear reliably. Hopefully, there is a way to specifically flush just the "OK Still here" messages appropriately. I'd be happy to test any real solutions to this issue. Sincerely, Dan.

2 2

[Dovecot] Dovecot 2.0b4 configuration issue, listen config option?
by Brandon Lamb 16 Apr '10

16 Apr '10

Hi, In my 1.2 setup I have pop3 running on ip x.x.x.2 and imap on x.x.x.7 In 2.0 how do I say listen x.x.x.2:110 x.x.x.7:143 so that pop3 is not listening on the .7 ip? I tried putting a listen = ip in the listener section but it complained. I cant figure it out and cant seem to find any 2.0 specific documentation, trying to search in the wiki directory of /usr/local/share/doc/dovecot/wiki but so far no help. I realize I can firewall that off but I dont want to rely on that. Is this no longer possible?

3 9

[Dovecot] Thunderbird subscription bug ?
by Thomas Hummel 16 Apr '10

16 Apr '10

Hello Timo, I'm having a hard time trying to understand why Thunderbird 3.0.3 doesn't allow me to subscribe to a shared mailbox which I can subscribe to via Kmail for instance. I'm running dovecot-1.2.11/Maildir. The user 'doveimap' shares his mailbox folder/subfolder to the 'dovepop' user which should see it in the '#shared' shared namespace at the following "location" : #shared/doveimap/folder/subfolder ACL (and unix permissions) are ok : # cat /courriel/boites/doveimap/.folder.subfolder/dovecot-acl user=dovepop akxeilprwts rawlogs give : a) kmail case : in : 4 NAMESPACE 5 LIST "" "" 6 LSUB "" "%" 7 LIST "" "%" 8 LSUB "" "#shared/%" 9 LIST "" "#shared/%" 10 LIST "" "INBOX" 11 SELECT "INBOX" 12 NOOP 13 UID FETCH 1:* (UID FLAGS) 14 UID FETCH 1:2 (UID RFC822.SIZE FLAGS ENVELOPE BODY.PEEK[HEADER.FIELDS (REFERENCES)]) 15 LIST "" "*" 16 LIST "" "#shared/*" 17 LSUB "" "*" 18 LSUB "" "#shared/*" 19 LIST "" "#shared/doveimap/folder/subfolder" 20 SUBSCRIBE "#shared/doveimap/folder/subfolder" out : * NAMESPACE (("" "/")) (("#shared/" "/")) NIL 4 OK Namespace completed. * LIST (\Noselect) "/" "" 5 OK List completed. 6 OK Lsub completed. * LIST (\HasNoChildren) "/" "Trash" * LIST (\HasNoChildren) "/" "INBOX" * LIST (\Noselect \HasChildren) "/" "#shared" 7 OK List completed. 8 OK Lsub completed. * LIST (\Noselect \HasChildren) "/" "#shared/doveimap" 9 OK List completed. * LIST (\HasNoChildren) "/" "INBOX" 10 OK List completed. * FLAGS (\Answered \Flagged \Deleted \Seen \Draft) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft \*)] Flags permitted. * 2 EXISTS * 0 RECENT * OK [UIDVALIDITY 1270130617] UIDs valid * OK [UIDNEXT 3] Predicted next UID * OK [HIGHESTMODSEQ 1] Highest 11 OK [READ-WRITE] Select completed. 12 OK NOOP completed. * 1 FETCH (UID 1 FLAGS (\Seen)) * 2 FETCH (UID 2 FLAGS (\Seen)) 13 OK Fetch completed. * 1 FETCH (UID 1 RFC822.SIZE 1140 FLAGS (\Seen) ENVELOPE [...] * 2 FETCH (UID 2 RFC822.SIZE 1138 FLAGS (\Seen) ENVELOPE [...] 14 OK Fetch completed. * LIST (\HasNoChildren) "/" "Trash" * LIST (\HasNoChildren) "/" "INBOX" * LIST (\Noselect \HasChildren) "/" "#shared/doveimap" * LIST (\HasNoChildren) "/" "#shared/doveimap/folder/subfolder" 15 OK List completed. * LIST (\Noselect \HasChildren) "/" "#shared/doveimap" * LIST (\HasNoChildren) "/" "#shared/doveimap/folder/subfolder" 16 OK List completed. 17 OK Lsub completed. 18 OK Lsub completed. * LIST (\HasNoChildren) "/" "#shared/doveimap/folder/subfolder" 19 OK List completed. 20 OK Subscribe completed. b) Thunderbird case : in : 4 namespace 5 ENABLE CONDSTORE 6 lsub "" "*" 7 lsub "" "#shared/*" 8 list "" "INBOX" 9 list "" "Trash" 10 create "Trash" 11 select "INBOX" (CONDSTORE) 12 myrights "INBOX" 13 getacl "INBOX" 14 UID fetch 1:* (FLAGS) 15 UID fetch 1:2 (UID RFC822.SIZE FLAGS BODY.PEEK[HEADER.FIELDS (From To Cc Bcc Subject Date Message-ID Priority X-Priority References Newsgroups In-Reply-To Content-Type)]) 16 UID fetch 1,2 (UID RFC822.SIZE BODY.PEEK[]) 17 IDLE DONE 18 lsub "" "*" 19 list "" "%" 20 list "" "%/%" 21 lsub "" "#shared/*" 22 list "" "#shared/%" 23 list "" "#shared/%/%" 24 IDLE DONE 25 list "" "#shared/#shared/%" 26 list "" "#shared/#shared/%/%" 27 IDLE DONE 28 noop 29 IDLE DONE out : * NAMESPACE (("" "/")) (("#shared/" "/")) NIL 4 OK Namespace completed. * ENABLED CONDSTORE 5 OK Enabled. 6 OK Lsub completed. 7 OK Lsub completed. * LIST (\HasNoChildren) "/" "INBOX" 8 OK List completed. * LIST (\HasNoChildren) "/" "Trash" 9 OK List completed. 10 NO [ALREADYEXISTS] Mailbox exists. * FLAGS (\Answered \Flagged \Deleted \Seen \Draft) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft \*)] Flags permitted. * 2 EXISTS * 0 RECENT * OK [UIDVALIDITY 1270130617] UIDs valid * OK [UIDNEXT 3] Predicted next UID * OK [HIGHESTMODSEQ 1] Highest 11 OK [READ-WRITE] Select completed. * MYRIGHTS "INBOX" lrwstipekxacd 12 OK Myrights completed. * ACL "INBOX" "dovepop" lrwstipekxacd 13 OK Getacl completed. * 1 FETCH (UID 1 FLAGS (\Seen)) * 2 FETCH (UID 2 FLAGS (\Seen)) 14 OK Fetch completed. * 1 FETCH (UID 1 RFC822.SIZE 1140 FLAGS (\Seen) [...] * 2 FETCH (UID 2 RFC822.SIZE 1138 FLAGS (\Seen) [...] 15 OK Fetch completed. * 1 FETCH (UID 1 RFC822.SIZE 1140 BODY[] [...] * 2 FETCH (UID 2 RFC822.SIZE 1138 BODY[] [...] 16 OK Fetch completed. + idling 17 OK Idle completed. 18 OK Lsub completed. * LIST (\HasNoChildren) "/" "Trash" * LIST (\HasNoChildren) "/" "INBOX" * LIST (\Noselect \HasChildren) "/" "#shared" 19 OK List completed. * LIST (\Noselect \HasChildren) "/" "#shared/doveimap" 20 OK List completed. 21 OK Lsub completed. * LIST (\Noselect \HasChildren) "/" "#shared/doveimap" 22 OK List completed. * LIST (\Noselect \HasChildren) "/" "#shared/doveimap/folder" 23 OK List completed. + idling * OK Still here 24 OK Idle completed. 25 OK List completed. 26 OK List completed. + idling 27 OK Idle completed. 28 OK NOOP completed. + idling 29 OK Idle completed. So TB doesn't seem to ask for what's behind #shared/doveimap/folder although the server tells him it has children... As a matter of fact, in the subscribe dialog box in th UI, I see #shared doveimap folder but with no check box alongside My server is configured like this : namespace private { separator = / location = maildir:/courriel/boites/%u:CONTROL=/courriel/meta/%u:INDEX=/var/dovecot-test/indexes/%1u/%u inbox = yes hidden = no list = yes subscriptions = yes } namespace shared { separator = / prefix = "#shared/%%u/" location = maildir:/courriel/boites/%%u:INDEX=/var/dovecot-test/indexes/%1u/%u/shared/%%u subscriptions = no list = children } Is it a bug in TB ? Thanks -- Thomas Hummel | Institut Pasteur <hummel(a)pasteur.fr> | Pôle informatique - systèmes et réseau

3 2

[Dovecot] Shared mailboxes unix permissions
by Thomas Hummel 16 Apr '10

16 Apr '10

Hello Timo, I'm running dovecot-1.2.11/Maildir and plan to migrate to single UID mailboxes some day, but for now, I've got "system" users and I'm testing permissions handling in order to set up shared mailboxes. The private namespace mailboxes location is location = maildir:/courriel/boites/%u:CONTROL=/courriel/meta/%u:INDEX=/var/dovecot-test/indexes/%1u/%u I've . created a 'doveshared' unix group . added 'mail_access_groups = doveshared' . chmod/chgrp the maildir and control dirs like this : drwxrws--- 7 doveimap doveshared 4096 Mar 31 18:47 /courriel/boites/doveimap drwxrws--- 5 doveimap doveshared 4096 Mar 31 18:47 /courriel/meta/doveimap [which were empty] . added 'acl_shared_dict = file:/var/dovecot-test/dict/shared-mailboxes' with drwxrwxs-- 2 root doveshared 512 Mar 31 18:44 /var/dovecot-test/dict [which were empty] Everything gets created with the permission I was expecting, except : -rw------- 1 doveimap doveshared 8 Mar 31 18:47 /courriel/meta/doveimap/dovecot-uidvalidity -rw------- 1 doveimap doveshared 0 Mar 31 18:44 /courriel/meta/doveimap/dovecot-uidvalidity.4bb37be4 and -rw------- 1 doveimap doveshared 0 Mar 31 18:44 /var/dovecot-test/dict/shared-mailboxes I can't see the explanation in http://wiki.dovecot.org/SharedMailboxes/Permissions and the 'dovecot-shared' file doesn't help (besides, my understanding is that in 1.2x, it's for backward compatibility reason). Any idea ? Thanks -- Thomas Hummel | Institut Pasteur <hummel(a)pasteur.fr> | Pôle informatique - systèmes et réseau

2 3

[Dovecot] "try MD5-CRYPT scheme instead"?
by Ralf Hildebrandt 16 Apr '10

16 Apr '10

I'm getting a lot of entries like this in my log: Apr 16 10:10:18 postamt dovecot: auth(default): shadow(doxxxris,141.42.206.38): CRYPT(wrongpassword) != '$1$qSe71xxx$09MoMqNyll.wPLCdSaFuA0', try MD5-CRYPT scheme instead It seems that changing the password via SquirrelMail (poppasswd) breaks the authentication scheme somewhat. Users cannot log in anymore. -- Ralf Hildebrandt Geschäftsbereich IT | Abteilung Netzwerk Charité - Universitätsmedizin Berlin Campus Benjamin Franklin Hindenburgdamm 30 | D-12203 Berlin Tel. +49 30 450 570 155 | Fax: +49 30 450 570 962 ralf.hildebrandt(a)charite.de | http://www.charite.de

2 2

[Dovecot] Error in SimpleVirtualInstall
by Simone Neugierig 16 Apr '10

16 Apr '10

Hi, i still got a little Problem (Thunderbird cant copy my mail to "Gesendet") but besides that i can send and receive mail. ive used: http://wiki.dovecot.org/HowTo/SimpleVirtualInstall there it says that /etc/dovecot/passwd should look like: test:{PLAIN}pass bill:{PLAIN}secret timo@example.com:{PLAIN}hello123 dave@example.com:{PLAIN}world234 joe@elsewhere.org:{PLAIN}whee jane@elsewhere.org:{PLAIN}mypass but the syntax used for test and bill didnt work here, i had to add the virtual domain to get it working. or else deliver wasnt able to recognize the user (unknown user). Can anybody please confirm this? Also i would like to suggest to explicitely say that protocol lda { } is needed even if it contains nothing. or better change the example so that it contains the postmaster address. but the one in the howto just contains .. dovecot probably wont even start with that example. this is dovecot 1.2.11 -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

2 1

[Dovecot] dovecot2: authcache statistics
by Andreas Schulze 16 Apr '10

16 Apr '10

Hello, setting auth_cache_size to nonzero enables caching of authentication data. The cache hitrate may be logged by sending SIGUSR2 to dovecot/auth process. Is it possible that dovecot logs statistics on regular intervals ? ( like http://www.postfix.org/postconf.5.html#anvil_status_update_time ) Andreas -- Andreas Schulze Internetdienste | P532 DATEV eG 90329 Nürnberg | Telefon +49 911 319-0 | Telefax +49 911 319-3196 E-Mail info @datev.de | Internet www.datev.de Sitz: 90429 Nürnberg, Paumgartnerstr. 6-14 | Registergericht Nürnberg, GenReg Nr.70 Vorstand Prof. Dieter Kempf (Vorsitzender) Dipl.-Kfm. Wolfgang Stegmann (stellvertretender Vorsitzender) Dipl.-Kfm. Michael Leistenschneider Jörg Rabe v. Pappenheim Dipl.-Vw. Eckhard Schwarzer Vorsitzender des Aufsichtsrates: Reinhard Verholen

2 2

[Dovecot] manipulating mails
by Sven 16 Apr '10

16 Apr '10

Hi all, sorry for my inexperience. I am still learning. I searched for a while but found nothing useful. I am looking for the best way to manipulate headers, bodies, attachments, locations, folders and/or flags while using deliver and e.g. perl. In other words manipulating the whole mail in respect of the requirements of the LDA, MUA and User. I see two diffent ways. As i understand I can work via IMAP and pre-auth sessions, which keeps track of the index files from dovecot and handles complications with users already logged in. Can I use e.g. some perl imap module -> download the mail -> change it -> upload it again with all other parameters unchanged and no confusions on the user's side? This way I think its complex to rewrite the mail header/body/attachments as desired and keep everything else, eg times, flags, headers. the goal should be close to invisibility for a user. This maybe leads to open pre-authed sockets/ports and a lot of work. The other way would be to directly modify the file system. This keeps everything (i think), but changes mtime and needs some index file and MUA cache update. For performance reasons I would like to not use the index-automatism of dovecot but do it manualy at runtime of some script. How does the index update gets initialized from the outside? Same for the client cache. Is there some command available or is a IMAP login my workaround? This maybe leads to password problems or open pre-authed sockets/ports. Last question: Is it possible to add some received-tag to a mail-header when appending mails via imap to the mailbox? Can anyone turn my head in the desired direction? Thanks in advance Sven

3 6

[Dovecot] Small dovecot.org downtime
by Timo Sirainen 16 Apr '10

16 Apr '10

dovecot.org gets more memory in a few hours, so it'll be down for a while (maybe max 30 mins?)

1 0

[Dovecot] catch-all not working with postfix dovecot lda
by fakessh 16 Apr '10

16 Apr '10

its tha archive to the cross post to postfix-users help me http://www.mail-archive.com/postfix-users@postfix.org/msg22963.html On Fri, 16 Apr 2010 00:26:25 +0200, fakessh <fakessh(a)fakessh.eu> wrote: On Thu, 15 Apr 2010 08:33:43 -0500, Noel Jones <njones(a)megan.vbhcs.org> wrote: On 4/14/2010 3:42 PM, fakessh wrote: On Wed, 14 Apr 2010 13:50:34 -0500, Noel Jones<njones(a)megan.vbhcs.org> wrote: On 4/14/2010 1:45 PM, fakessh wrote: On Wed, 14 Apr 2010 14:12:25 -0400, Charles Marcus <CMarcus(a)Media-Brokers.com> wrote: I changed the entries @fakessh to root@localhost in /etc/postfix/virtual postmap then a file to the postfix restart. all without success, or rather the same mistake Then post your new "postconf -n", log entries showing the problem, and file contents. my postcon -n [root@r13151 ~]# postconf -n alias_database = hash:/etc/aliases alias_maps = hash:/etc/aliases body_checks = regexp:/etc/postfix/body_checks.cf bounce_notice_recipient = postmaster broken_sasl_auth_clients = yes command_directory = /usr/sbin config_directory = /etc/postfix content_filter = dksign:[127.0.0.1]:10028 daemon_directory = /usr/libexec/postfix debug_peer_level = 2 default_privs = nobody double_bounce_sender = no header_checks = regexp:/etc/postfix/header_checks.cf home_mailbox = Maildir/ html_directory = no in_flow_delay = 10 inet_interfaces = all mail_spool_directory = /var/spool/mail mailbox_command = /usr/libexec/dovecot/deliver mailq_path = /usr/bin/mailq.postfix manpage_directory = /usr/share/man maps_rbl_domains = bl.spamcop.net mime_header_checks = regexp:/etc/postfix/mime_header_checks.cf mydestination = $myhostname, localhost.$mydomain mydomain = r13151.ovh.net mynetworks = 127.0.0.0/8 ,87.98.186.232 myorigin = $mydomain newaliases_path = /usr/bin/newaliases.postfix queue_run_delay = 2000s readme_directory = /usr/share/doc/postfix-2.3.3/README_FILES recipient_delimiter = + relay_domains = sample_directory = /usr/share/doc/postfix-2.3.3/samples sendmail_path = /usr/sbin/sendmail.postfix setgid_group = postdrop smtp_sasl_security_options = noanonymous smtp_sasl_tls_security_options = noanonymous smtp_sender_dependent_authentication = yes smtp_tls_loglevel = 3 smtp_tls_session_cache_database = btree:/var/lib/postfix/smtp_tls_session_cache smtpd_banner = $myhostname ESMTP $mail_name ($mail_version) smtpd_client_restrictions = permit_mynetworks,reject_unknown_reverse_client_hostname,reject_unauth_pipelining, reject_non_fqdn_recipient , permit smtpd_milters = inet:[127.0.0.1]:10040 smtpd_recipient_restrictions = permit_mynetworks permit_inet_interfaces permit_sasl_authenticated reject_unverified_recipient reject_non_fqdn_sender reject_non_fqdn_recipient reject_unknown_sender_domain reject_unknown_recipient_domain reject_unknown_reverse_client_hostname reject_unauth_destination reject_unauth_pipelining reject_rbl_client zen.spamhaus.org reject_sender_login_mismatch check_policy_service unix:postgrey/socket check_sender_access hash:/etc/postfix/check_backscatterer check_policy_service unix:private/spfpolicy reject_rbl_client bl.spamcop.net reject_rhsbl_sender dbl.spamhaus.org reject_rbl_client cbl.abuseat.org reject_rbl_client b.barracudacentral.org smtpd_reject_unlisted_sender = no smtpd_sasl_auth_enable = yes smtpd_sasl_authenticated_header = yes smtpd_sasl_local_domain = $myhostname smtpd_sasl_path = private/auth smtpd_sasl_type = dovecot smtpd_tls_CAfile = /etc/pki/tls/sub.class4.server.ca.pem smtpd_tls_auth_only = yes smtpd_tls_cert_file = /etc/pki/tls/newcerts/01.pem smtpd_tls_key_file = /etc/pki/tls/private/r13151.ovh.net.key smtpd_tls_received_header = yes smtpd_tls_session_cache_database = btree:/var/lib/postfix/smtpd_tls_session_cache smtpd_use_tls = yes soft_bounce = no tls_random_source = dev:/dev/urandom unknown_local_recipient_reject_code = 550 virtual_alias_domains = fakessh.eu renelacroute.fr nicolaspichot.fr virtual_alias_maps = hash:/etc/postfix/virtual But you already have all the information you need to fix this yourself. Key points are 1) use fully qualified names in virtual_alias_maps. ie. user(a)example1.com user(a)example2.com *not* user(a)example1.com user my jed /etc/postfix/virtual # # AUTHOR(S) # Wietse Venema # IBM T.J. Watson Research # P.O. Box 704 # Yorktown Heights, NY 10598, USA # # VIRTUAL(5$ postmaster(a)fakessh.eu root(a)localhost.r13151.ovh.net fakessh(a)fakessh.eu fakessh(a)localhost.r13151.ovh.net webmail(a)fakessh.eu webmail(a)localhost.r13151.ovh.net serge(a)fakessh.eu serge(a)localhost.r13151.ovh.net @fakessh root(a)localhost.r13151.ovh.net renelacroute(a)renelacroute.fr renelacroute(a)localhost.r13151.ovh.net @renelacroute.fr root(a)localhost.r13151.ovh.net postmaster(a)renelacroute.fr root(a)localhost.r13151.ovh.net nicolaspichot(a)nicolaspichot.fr nicolaspichot(a)localhost.r13151.ovh.net @nicolaspichot.fr root(a)localhost.r13151.ovh.net 2) if you want local delivery of the mail, the new domain must be listed in mydestination. i use local delivery agent [root@r13151 ~]# rpm -qa | grep dovecot dovecot-sieve-devel-0.1.15-4.el5 dovecot-sieve-cmu-1.1.8-11.el5 dovecot-1.2.11-3_108.el5 dovecot-sieve-0.1.15-4.el5 dovecot-managesieve-0.11.11-0_4.el5 [root@r13151 ~]# to atrpms repo Your fix may be as simple as adding "localhost.$mydomain" to mydestination. my hostname is r13151.ovh.net -- Noel Jones my log to correct session Apr 16 00:36:16 r13151 postfix/postfix-script: stopping the Postfix mail system Apr 16 00:36:16 r13151 postfix/master[14544]: terminating on signal 15 Apr 16 00:36:17 r13151 postfix/postfix-script: starting the Postfix mail system Apr 16 00:36:17 r13151 postfix/master[14848]: daemon started -- version 2.3.3, configuration /etc/postfix Apr 16 00:36:31 r13151 postfix/tlsmgr[14861]: open smtp TLS cache btree:/var/lib/postfix/smtp_tls_session_cache Apr 16 00:36:31 r13151 postfix/tlsmgr[14861]: tlsmgr_cache_run_event: start TLS smtp session cache cleanup Apr 16 00:36:31 r13151 postfix/smtpd[14859]: connect from smtp20.orange.fr[80.12.242.26] Apr 16 00:36:31 r13151 postfix/cleanup[14866]: 6617457F6F: message-id=<20100415223631.6617457F6F(a)r13151.ovh.net> Apr 16 00:36:31 r13151 postfix/qmgr[14851]: 6617457F6F: from=<postmaster(a)r13151.ovh.net>, size=260, nrcpt=1 (queue active) Apr 16 00:36:31 r13151 postfix/local[14867]: 6617457F6F: to=<fakessh(a)localhost.r13151.ovh.net>, orig_to=<fakessh(a)fakessh.eu>, relay=local, delay=0.16, delays=0.15/0.01/0/0, dsn=2.0.0, status=deliverable (delivers to command: /usr/libexec/dovecot/deliver) Apr 16 00:36:31 r13151 postfix/qmgr[14851]: 6617457F6F: removed Apr 16 00:36:34 r13151 postgrey[838]: action=pass, reason=client whitelist, client_name=smtp20.orange.fr, client_address=80.12.242.26, sender=john.swilting(a)wanadoo.fr, recipient=fakessh(a)fakessh.eu Apr 16 00:36:34 r13151 postgrey[838]: cleaning up old logs... Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: : SPF none (No applicable sender policy available): Envelope-from: john.swilting(a)wanadoo.fr Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: handler sender_policy_framework: is decisive. Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: : Policy action=PREPEND Received-SPF: none (wanadoo.fr: No applicable sender policy available) receiver=r13151.ovh.net; identity=mailfrom; envelope-from="john.swilting(a)wanadoo.fr"; helo=smtp20.orange.fr; client-ip=80.12.242.26 Apr 16 00:36:35 r13151 postfix/smtpd[14859]: AD73957F6F: client=smtp20.orange.fr[80.12.242.26] Apr 16 00:36:35 r13151 postfix/cleanup[14866]: AD73957F6F: message-id=<1271370966.7979.13.camel(a)localhost.localdomain> Apr 16 00:36:37 r13151 postfix/qmgr[14851]: AD73957F6F: from=<john.swilting(a)wanadoo.fr>, size=1460, nrcpt=1 (queue active) Apr 16 00:36:37 r13151 postfix/smtpd[14859]: disconnect from smtp20.orange.fr[80.12.242.26] Apr 16 00:36:37 r13151 dkimproxy.out[817]: connect from 127.0.0.1 Apr 16 00:36:37 r13151 postfix/smtpd[14874]: connect from localhost.localdomain[127.0.0.1] Apr 16 00:36:37 r13151 postfix/smtp[14873]: discarding EHLO keywords: 8BITMIME STARTTLS Apr 16 00:36:38 r13151 postfix/smtpd[14874]: 30AB357F71: client=smtp20.orange.fr[80.12.242.26] Apr 16 00:36:38 r13151 dkimproxy.out[817]: DKIM signing - skipped; message-id=<1271370966.7979.13.camel(a)localhost.localdomain>, from=<john.swilting(a)wanadoo.fr> Apr 16 00:36:38 r13151 postfix/cleanup[14866]: 30AB357F71: message-id=<1271370966.7979.13.camel(a)localhost.localdomain> Apr 16 00:36:39 r13151 postfix/qmgr[14851]: 30AB357F71: from=<john.swilting(a)wanadoo.fr>, size=1866, nrcpt=1 (queue active) Apr 16 00:36:39 r13151 postfix/smtp[14873]: AD73957F6F: to=<fakessh(a)fakessh.eu>, relay=127.0.0.1[127.0.0.1]:10028, delay=7.9, delays=6.5/0.01/0.3/1.1, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as 30AB357F71) Apr 16 00:36:39 r13151 postfix/smtpd[14874]: disconnect from localhost.localdomain[127.0.0.1] Apr 16 00:36:39 r13151 clamd[7779]: SelfCheck: Database status OK. Apr 16 00:36:39 r13151 postfix/qmgr[14851]: AD73957F6F: removed Apr 16 00:36:40 r13151 postfix/smtpd[14880]: connect from localhost.localdomain[127.0.0.1] Apr 16 00:36:40 r13151 postfix/smtpd[14880]: 6C5AB57F6F: client=localhost.localdomain[127.0.0.1] Apr 16 00:36:40 r13151 postfix/cleanup[14866]: 6C5AB57F6F: message-id=<1271370966.7979.13.camel(a)localhost.localdomain> Apr 16 00:36:40 r13151 postfix/qmgr[14851]: 6C5AB57F6F: from=<john.swilting(a)wanadoo.fr>, size=2728, nrcpt=1 (queue active) Apr 16 00:36:40 r13151 postfix/smtpd[14880]: disconnect from localhost.localdomain[127.0.0.1] Apr 16 00:36:40 r13151 postfix/smtpd[14880]: connect from localhost.localdomain[127.0.0.1] Apr 16 00:36:40 r13151 postfix/smtpd[14880]: A3F4A57F74: client=localhost.localdomain[127.0.0.1] Apr 16 00:36:40 r13151 postfix/cleanup[14866]: A3F4A57F74: message-id=<1271370966.7979.13.camel(a)localhost.localdomain> Apr 16 00:36:40 r13151 postfix/qmgr[14851]: A3F4A57F74: from=<>, size=2240, nrcpt=1 (queue active) Apr 16 00:36:40 r13151 postfix/smtpd[14880]: disconnect from localhost.localdomain[127.0.0.1] Apr 16 00:36:41 r13151 amavis[9551]: (09551-05) Passed CLEAN, [127.0.0.1] [90.55.123.193] <john.swilting(a)wanadoo.fr> -> <fakessh(a)localhost.r13151.ovh.net>, Message-ID: <1271370966.7979.13.camel(a)localhost.localdomain>, mail_id: 2HnPlHlQZ5AT, Hits: 4.791, size: 2046, queued_as: 6C5AB57F6F, 1723 ms Apr 16 00:36:41 r13151 postfix/lmtp[14876]: 30AB357F71: to=<fakessh(a)localhost.r13151.ovh.net>, orig_to=<fakessh(a)fakessh.eu>, relay=127.0.0.1[127.0.0.1]:10024, delay=2.8, delays=1.1/0.01/0.01/1.7, dsn=2.0.0, status=sent (250 2.0.0 Ok, id=09551-05, from MTA([127.0.0.1]:10025): 250 2.0.0 Ok: queued as 6C5AB57F6F) Apr 16 00:36:41 r13151 postfix/qmgr[14851]: 30AB357F71: removed Apr 16 00:36:41 r13151 dovecot: deliver(fakessh): msgid=<1271370966.7979.13.camel(a)localhost.localdomain>: saved mail to INBOX Apr 16 00:36:41 r13151 postfix/local[14867]: 6C5AB57F6F: to=<fakessh(a)localhost.r13151.ovh.net>, relay=local, delay=0.93, delays=0.22/0/0/0.71, dsn=2.0.0, status=sent (delivered to command: /usr/libexec/dovecot/deliver) Apr 16 00:36:41 r13151 postfix/qmgr[14851]: 6C5AB57F6F: removed Apr 16 00:36:41 r13151 dovecot: deliver(root): msgid=<1271370966.7979.13.camel(a)localhost.localdomain>: saved mail to INBOX Apr 16 00:36:41 r13151 postfix/local[14886]: A3F4A57F74: to=<root(a)r13151.ovh.net>, orig_to=<postmaster(a)r13151.ovh.net>, relay=local, delay=0.99, delays=0.29/0.01/0/0.68, dsn=2.0.0, status=sent (delivered to command: /usr/libexec/dovecot/deliver) Apr 16 00:36:41 r13151 postfix/qmgr[14851]: A3F4A57F74: removed my error log to tentative a catch-all Apr 16 00:38:28 r13151 dovecot: IMAP(fakessh): Disconnected: Logged out bytes=329/7731 Apr 16 00:39:37 r13151 dovecot: dovecot: Killed with signal 15 (by pid=15053 uid=0 code=kill) Apr 16 00:39:38 r13151 dovecot: Dovecot v1.2.11 starting up (core dumps disabled) Apr 16 00:39:40 r13151 postfix/postfix-script: stopping the Postfix mail system Apr 16 00:39:40 r13151 postfix/master[14848]: terminating on signal 15 Apr 16 00:39:42 r13151 postfix/postfix-script: starting the Postfix mail system Apr 16 00:39:42 r13151 postfix/master[15149]: daemon started -- version 2.3.3, configuration /etc/postfix Apr 16 00:39:52 r13151 postfix/tlsmgr[15162]: open smtp TLS cache btree:/var/lib/postfix/smtp_tls_session_cache Apr 16 00:39:52 r13151 postfix/tlsmgr[15162]: tlsmgr_cache_run_event: start TLS smtp session cache cleanup Apr 16 00:39:52 r13151 postfix/smtpd[15160]: connect from smtp28.orange.fr[80.12.242.101] Apr 16 00:39:52 r13151 postfix/cleanup[15167]: B8F9157F6F: message-id=<20100415223952.B8F9157F6F(a)r13151.ovh.net> Apr 16 00:39:52 r13151 postfix/qmgr[15152]: B8F9157F6F: from=<postmaster(a)r13151.ovh.net>, size=260, nrcpt=1 (queue active) Apr 16 00:39:52 r13151 postfix/error[15168]: B8F9157F6F: to=<catch(a)fakessh.eu>, relay=none, delay=0.17, delays=0.15/0.01/0/0, dsn=5.0.0, status=undeliverable (User unknown in virtual alias table) Apr 16 00:39:52 r13151 postfix/qmgr[15152]: B8F9157F6F: removed Apr 16 00:39:55 r13151 postfix/smtpd[15160]: NOQUEUE: reject: RCPT from smtp28.orange.fr[80.12.242.101]: 450 4.1.1 <catch(a)fakessh.eu>: Recipient address rejected: undeliverable address: User unknown in virtual alias table; from=<john.swilting(a)wanadoo.fr> to=<catch(a)fakessh.eu> proto=ESMTP helo=<smtp28.orange.fr> Apr 16 00:39:55 r13151 postfix/smtpd[15160]: disconnect from smtp28.orange.fr[80.12.242.101] Apr 16 00:40:15 r13151 postfix/smtpd[15160]: connect from smtp28.orange.fr[80.12.242.101] Apr 16 00:40:15 r13151 postfix/smtpd[15160]: NOQUEUE: reject: RCPT from smtp28.orange.fr[80.12.242.101]: 450 4.1.1 <catch(a)fakessh.eu>: Recipient address rejected: undeliverable address: User unknown in virtual alias table; from=<john.swilting(a)wanadoo.fr> to=<catch(a)fakessh.eu> proto=ESMTP helo=<smtp28.orange.fr> Apr 16 00:40:15 r13151 postfix/smtpd[15160]: disconnect from smtp28.orange.fr[80.12.242.101]

3 5

[Dovecot] Fwd: Re: catch-all not working with postfix dovecot lda (fwd)
by fakessh 16 Apr '10

16 Apr '10

it may be a problem in dealing with amavisd perl milter Subject: Re: [Dovecot] catch-all not working with postfix dovecot lda On Fri, 16 Apr 2010 09:07:55 +1000, Noel Butler <noel.butler(a)ausics.net> wrote: > Postfix must first "know the user(s)" > therefore this isa postfix issue and not dovecot > dovecot deliver assumes the MTA has verified the user to accept mail > from and does not do further authentication > > how to build a catch-all with dovecot lda the question then. is not a postfix issue > On Fri, 2010-04-16 at 01:00 +0200, fakessh wrote: > >> its tha archive to the cross post to postfix-users >> help me >> >> http://www.mail-archive.com/postfix-users@postfix.org/msg22963.html >> >> >> >> On Fri, 16 Apr 2010 00:26:25 +0200, fakessh <fakessh(a)fakessh.eu> wrote: >> >> On Thu, 15 Apr 2010 08:33:43 -0500, Noel Jones >> <njones(a)megan.vbhcs.org> wrote: >> >> On 4/14/2010 3:42 PM, fakessh wrote: >> >> On Wed, 14 Apr 2010 13:50:34 -0500, Noel >> Jones<njones(a)megan.vbhcs.org> wrote: >> >> On 4/14/2010 1:45 PM, fakessh wrote: >> >> On Wed, 14 Apr 2010 14:12:25 -0400, Charles Marcus >> <CMarcus(a)Media-Brokers.com> wrote: >> >> I changed the entries @fakessh to root@localhost in >> >> /etc/postfix/virtual >> >> postmap then a file to the postfix restart. all without >> success, or rather the same mistake >> >> Then post your new "postconf -n", log entries showing the >> problem, >> and file contents. >> >> my postcon -n [root@r13151 ~]# postconf -n alias_database = >> hash:/etc/aliases alias_maps = hash:/etc/aliases body_checks = >> regexp:/etc/postfix/body_checks.cf bounce_notice_recipient = postmaster >> broken_sasl_auth_clients = yes command_directory = /usr/sbin >> config_directory = /etc/postfix content_filter = dksign:[127.0.0.1]:10028 >> daemon_directory = /usr/libexec/postfix debug_peer_level = 2 >> default_privs >> = nobody double_bounce_sender = no header_checks = >> regexp:/etc/postfix/header_checks.cf home_mailbox = Maildir/ >> html_directory >> = no in_flow_delay = 10 inet_interfaces = all mail_spool_directory = >> /var/spool/mail mailbox_command = /usr/libexec/dovecot/deliver >> mailq_path = >> /usr/bin/mailq.postfix manpage_directory = /usr/share/man >> maps_rbl_domains >> = bl.spamcop.net mime_header_checks = >> regexp:/etc/postfix/mime_header_checks.cf mydestination = $myhostname, >> localhost.$mydomain mydomain = r13151.ovh.net mynetworks = 127.0.0.0/8 >> ,87.98.186.232 myorigin = $mydomain newaliases_path = >> /usr/bin/newaliases.postfix queue_run_delay = 2000s readme_directory = >> /usr/share/doc/postfix-2.3.3/README_FILES recipient_delimiter = + >> relay_domains = sample_directory = /usr/share/doc/postfix-2.3.3/samples >> sendmail_path = /usr/sbin/sendmail.postfix setgid_group = postdrop >> smtp_sasl_security_options = noanonymous smtp_sasl_tls_security_options = >> noanonymous smtp_sender_dependent_authentication = yes smtp_tls_loglevel >> = >> 3 smtp_tls_session_cache_database = >> btree:/var/lib/postfix/smtp_tls_session_cache smtpd_banner = $myhostname >> ESMTP $mail_name ($mail_version) smtpd_client_restrictions = >> >> permit_mynetworks,reject_unknown_reverse_client_hostname,reject_unauth_pipelining, >> >> reject_non_fqdn_recipient , permit smtpd_milters = >> inet:[127.0.0.1]:10040 smtpd_recipient_restrictions = permit_mynetworks >> permit_inet_interfaces permit_sasl_authenticated >> reject_unverified_recipient reject_non_fqdn_sender >> reject_non_fqdn_recipient reject_unknown_sender_domain >> reject_unknown_recipient_domain reject_unknown_reverse_client_hostname >> reject_unauth_destination reject_unauth_pipelining reject_rbl_client >> zen.spamhaus.org reject_sender_login_mismatch check_policy_service >> unix:postgrey/socket check_sender_access >> hash:/etc/postfix/check_backscatterer check_policy_service >> unix:private/spfpolicy reject_rbl_client bl.spamcop.net >> reject_rhsbl_sender >> dbl.spamhaus.org reject_rbl_client cbl.abuseat.org reject_rbl_client >> b.barracudacentral.org smtpd_reject_unlisted_sender = no >> smtpd_sasl_auth_enable = yes smtpd_sasl_authenticated_header = yes >> smtpd_sasl_local_domain = $myhostname smtpd_sasl_path = private/auth >> smtpd_sasl_type = dovecot smtpd_tls_CAfile = >> /etc/pki/tls/sub.class4.server.ca.pem smtpd_tls_auth_only = yes >> smtpd_tls_cert_file = /etc/pki/tls/newcerts/01.pem smtpd_tls_key_file = >> /etc/pki/tls/private/r13151.ovh.net.key smtpd_tls_received_header = yes >> smtpd_tls_session_cache_database = >> btree:/var/lib/postfix/smtpd_tls_session_cache smtpd_use_tls = yes >> soft_bounce = no tls_random_source = dev:/dev/urandom >> unknown_local_recipient_reject_code = 550 virtual_alias_domains = >> fakessh.eu renelacroute.fr nicolaspichot.fr virtual_alias_maps = >> hash:/etc/postfix/virtual >> >> But you already have all the information you need to fix this >> yourself. Key points are 1) use fully qualified names in >> virtual_alias_maps. ie. user(a)example1.com user(a)example2.com *not* >> user(a)example1.com user >> >> my jed /etc/postfix/virtual # # AUTHOR(S) # Wietse Venema # IBM T.J. >> Watson Research # P.O. Box 704 # Yorktown Heights, NY 10598, USA # # >> VIRTUAL(5$ postmaster(a)fakessh.eu root(a)localhost.r13151.ovh.net >> fakessh(a)fakessh.eu fakessh(a)localhost.r13151.ovh.net webmail(a)fakessh.eu >> webmail(a)localhost.r13151.ovh.net serge(a)fakessh.eu >> serge(a)localhost.r13151.ovh.net @fakessh root(a)localhost.r13151.ovh.net >> renelacroute(a)renelacroute.fr renelacroute(a)localhost.r13151.ovh.net >> @renelacroute.fr root(a)localhost.r13151.ovh.net postmaster(a)renelacroute.fr >> root(a)localhost.r13151.ovh.net nicolaspichot(a)nicolaspichot.fr >> nicolaspichot(a)localhost.r13151.ovh.net @nicolaspichot.fr >> root(a)localhost.r13151.ovh.net >> >> 2) if you want local delivery of the mail, the new domain must be >> listed in mydestination. >> >> i use local delivery agent [root@r13151 ~]# rpm -qa | grep dovecot >> dovecot-sieve-devel-0.1.15-4.el5 dovecot-sieve-cmu-1.1.8-11.el5 >> dovecot-1.2.11-3_108.el5 dovecot-sieve-0.1.15-4.el5 >> dovecot-managesieve-0.11.11-0_4.el5 [root@r13151 ~]# to atrpms repo >> >> Your fix may be as simple as adding "localhost.$mydomain" to >> mydestination. >> >> my hostname is r13151.ovh.net >> >> -- Noel Jones >> >> >> >> my log to correct session >> >> Apr 16 00:36:16 r13151 postfix/postfix-script: stopping the Postfix mail >> system >> Apr 16 00:36:16 r13151 postfix/master[14544]: terminating on signal 15 >> Apr 16 00:36:17 r13151 postfix/postfix-script: starting the Postfix mail >> system >> Apr 16 00:36:17 r13151 postfix/master[14848]: daemon started -- version >> 2.3.3, configuration /etc/postfix >> Apr 16 00:36:31 r13151 postfix/tlsmgr[14861]: open smtp TLS cache >> btree:/var/lib/postfix/smtp_tls_session_cache >> Apr 16 00:36:31 r13151 postfix/tlsmgr[14861]: tlsmgr_cache_run_event: >> start TLS smtp session cache cleanup >> Apr 16 00:36:31 r13151 postfix/smtpd[14859]: connect from >> smtp20.orange.fr[80.12.242.26] >> Apr 16 00:36:31 r13151 postfix/cleanup[14866]: 6617457F6F: >> message-id=<20100415223631.6617457F6F(a)r13151.ovh.net> >> Apr 16 00:36:31 r13151 postfix/qmgr[14851]: 6617457F6F: >> from=<postmaster(a)r13151.ovh.net>, size=260, nrcpt=1 (queue active) >> Apr 16 00:36:31 r13151 postfix/local[14867]: 6617457F6F: >> to=<fakessh(a)localhost.r13151.ovh.net>, orig_to=<fakessh(a)fakessh.eu>, >> relay=local, delay=0.16, delays=0.15/0.01/0/0, dsn=2.0.0, >> status=deliverable (delivers to command: /usr/libexec/dovecot/deliver) >> Apr 16 00:36:31 r13151 postfix/qmgr[14851]: 6617457F6F: removed >> Apr 16 00:36:34 r13151 postgrey[838]: action=pass, reason=client >> whitelist, client_name=smtp20.orange.fr, client_address=80.12.242.26, >> sender=john.swilting(a)wanadoo.fr, recipient=fakessh(a)fakessh.eu >> Apr 16 00:36:34 r13151 postgrey[838]: cleaning up old logs... >> Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: : SPF none (No >> applicable sender policy available): Envelope-from: >> john.swilting(a)wanadoo.fr >> Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: handler >> sender_policy_framework: is decisive. >> Apr 16 00:36:35 r13151 postfix/policy-spf[14869]: : Policy action=PREPEND >> Received-SPF: none (wanadoo.fr: No applicable sender policy available) >> receiver=r13151.ovh.net; identity=mailfrom; >> envelope-from="john.swilting(a)wanadoo.fr"; helo=smtp20.orange.fr; >> client-ip=80.12.242.26 >> Apr 16 00:36:35 r13151 postfix/smtpd[14859]: AD73957F6F: >> client=smtp20.orange.fr[80.12.242.26] >> Apr 16 00:36:35 r13151 postfix/cleanup[14866]: AD73957F6F: >> message-id=<1271370966.7979.13.camel(a)localhost.localdomain> >> Apr 16 00:36:37 r13151 postfix/qmgr[14851]: AD73957F6F: >> from=<john.swilting(a)wanadoo.fr>, size=1460, nrcpt=1 (queue active) >> Apr 16 00:36:37 r13151 postfix/smtpd[14859]: disconnect from >> smtp20.orange.fr[80.12.242.26] >> Apr 16 00:36:37 r13151 dkimproxy.out[817]: connect from 127.0.0.1 >> Apr 16 00:36:37 r13151 postfix/smtpd[14874]: connect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:37 r13151 postfix/smtp[14873]: discarding EHLO keywords: >> 8BITMIME STARTTLS >> Apr 16 00:36:38 r13151 postfix/smtpd[14874]: 30AB357F71: >> client=smtp20.orange.fr[80.12.242.26] >> Apr 16 00:36:38 r13151 dkimproxy.out[817]: DKIM signing - skipped; >> message-id=<1271370966.7979.13.camel(a)localhost.localdomain>, >> from=<john.swilting(a)wanadoo.fr> >> Apr 16 00:36:38 r13151 postfix/cleanup[14866]: 30AB357F71: >> message-id=<1271370966.7979.13.camel(a)localhost.localdomain> >> Apr 16 00:36:39 r13151 postfix/qmgr[14851]: 30AB357F71: >> from=<john.swilting(a)wanadoo.fr>, size=1866, nrcpt=1 (queue active) >> Apr 16 00:36:39 r13151 postfix/smtp[14873]: AD73957F6F: >> to=<fakessh(a)fakessh.eu>, relay=127.0.0.1[127.0.0.1]:10028, delay=7.9, >> delays=6.5/0.01/0.3/1.1, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as >> 30AB357F71) >> Apr 16 00:36:39 r13151 postfix/smtpd[14874]: disconnect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:39 r13151 clamd[7779]: SelfCheck: Database status OK. >> Apr 16 00:36:39 r13151 postfix/qmgr[14851]: AD73957F6F: removed >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: connect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: 6C5AB57F6F: >> client=localhost.localdomain[127.0.0.1] >> Apr 16 00:36:40 r13151 postfix/cleanup[14866]: 6C5AB57F6F: >> message-id=<1271370966.7979.13.camel(a)localhost.localdomain> >> Apr 16 00:36:40 r13151 postfix/qmgr[14851]: 6C5AB57F6F: >> from=<john.swilting(a)wanadoo.fr>, size=2728, nrcpt=1 (queue active) >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: disconnect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: connect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: A3F4A57F74: >> client=localhost.localdomain[127.0.0.1] >> Apr 16 00:36:40 r13151 postfix/cleanup[14866]: A3F4A57F74: >> message-id=<1271370966.7979.13.camel(a)localhost.localdomain> >> Apr 16 00:36:40 r13151 postfix/qmgr[14851]: A3F4A57F74: from=<>, >> size=2240, nrcpt=1 (queue active) >> Apr 16 00:36:40 r13151 postfix/smtpd[14880]: disconnect from >> localhost.localdomain[127.0.0.1] >> Apr 16 00:36:41 r13151 amavis[9551]: (09551-05) Passed CLEAN, [127.0.0.1] >> [90.55.123.193] <john.swilting(a)wanadoo.fr> -> >> <fakessh(a)localhost.r13151.ovh.net>, Message-ID: >> <1271370966.7979.13.camel(a)localhost.localdomain>, mail_id: 2HnPlHlQZ5AT, >> Hits: 4.791, size: 2046, queued_as: 6C5AB57F6F, 1723 ms >> Apr 16 00:36:41 r13151 postfix/lmtp[14876]: 30AB357F71: >> to=<fakessh(a)localhost.r13151.ovh.net>, orig_to=<fakessh(a)fakessh.eu>, >> relay=127.0.0.1[127.0.0.1]:10024, delay=2.8, delays=1.1/0.01/0.01/1.7, >> dsn=2.0.0, status=sent (250 2.0.0 Ok, id=09551-05, from >> MTA([127.0.0.1]:10025): 250 2.0.0 Ok: queued as 6C5AB57F6F) >> Apr 16 00:36:41 r13151 postfix/qmgr[14851]: 30AB357F71: removed >> Apr 16 00:36:41 r13151 dovecot: deliver(fakessh): >> msgid=<1271370966.7979.13.camel(a)localhost.localdomain>: saved mail to >> INBOX >> Apr 16 00:36:41 r13151 postfix/local[14867]: 6C5AB57F6F: >> to=<fakessh(a)localhost.r13151.ovh.net>, relay=local, delay=0.93, >> delays=0.22/0/0/0.71, dsn=2.0.0, status=sent (delivered to command: >> /usr/libexec/dovecot/deliver) >> Apr 16 00:36:41 r13151 postfix/qmgr[14851]: 6C5AB57F6F: removed >> Apr 16 00:36:41 r13151 dovecot: deliver(root): >> msgid=<1271370966.7979.13.camel(a)localhost.localdomain>: saved mail to >> INBOX >> Apr 16 00:36:41 r13151 postfix/local[14886]: A3F4A57F74: >> to=<root(a)r13151.ovh.net>, orig_to=<postmaster(a)r13151.ovh.net>, >> relay=local, >> delay=0.99, delays=0.29/0.01/0/0.68, dsn=2.0.0, status=sent (delivered to >> command: /usr/libexec/dovecot/deliver) >> Apr 16 00:36:41 r13151 postfix/qmgr[14851]: A3F4A57F74: removed >> >> >> my error log to tentative a catch-all >> >> Apr 16 00:38:28 r13151 dovecot: IMAP(fakessh): Disconnected: Logged out >> bytes=329/7731 >> Apr 16 00:39:37 r13151 dovecot: dovecot: Killed with signal 15 (by >> pid=15053 uid=0 code=kill) >> Apr 16 00:39:38 r13151 dovecot: Dovecot v1.2.11 starting up (core dumps >> disabled) >> Apr 16 00:39:40 r13151 postfix/postfix-script: stopping the Postfix mail >> system >> Apr 16 00:39:40 r13151 postfix/master[14848]: terminating on signal 15 >> Apr 16 00:39:42 r13151 postfix/postfix-script: starting the Postfix mail >> system >> Apr 16 00:39:42 r13151 postfix/master[15149]: daemon started -- version >> 2.3.3, configuration /etc/postfix >> Apr 16 00:39:52 r13151 postfix/tlsmgr[15162]: open smtp TLS cache >> btree:/var/lib/postfix/smtp_tls_session_cache >> Apr 16 00:39:52 r13151 postfix/tlsmgr[15162]: tlsmgr_cache_run_event: >> start TLS smtp session cache cleanup >> Apr 16 00:39:52 r13151 postfix/smtpd[15160]: connect from >> smtp28.orange.fr[80.12.242.101] >> Apr 16 00:39:52 r13151 postfix/cleanup[15167]: B8F9157F6F: >> message-id=<20100415223952.B8F9157F6F(a)r13151.ovh.net> >> Apr 16 00:39:52 r13151 postfix/qmgr[15152]: B8F9157F6F: >> from=<postmaster(a)r13151.ovh.net>, size=260, nrcpt=1 (queue active) >> Apr 16 00:39:52 r13151 postfix/error[15168]: B8F9157F6F: >> to=<catch(a)fakessh.eu>, relay=none, delay=0.17, delays=0.15/0.01/0/0, >> dsn=5.0.0, status=undeliverable (User unknown in virtual alias table) >> Apr 16 00:39:52 r13151 postfix/qmgr[15152]: B8F9157F6F: removed >> Apr 16 00:39:55 r13151 postfix/smtpd[15160]: NOQUEUE: reject: RCPT from >> smtp28.orange.fr[80.12.242.101]: 450 4.1.1 <catch(a)fakessh.eu>: Recipient >> address rejected: undeliverable address: User unknown in virtual alias >> table; from=<john.swilting(a)wanadoo.fr> to=<catch(a)fakessh.eu> proto=ESMTP >> helo=<smtp28.orange.fr> >> Apr 16 00:39:55 r13151 postfix/smtpd[15160]: disconnect from >> smtp28.orange.fr[80.12.242.101] >> Apr 16 00:40:15 r13151 postfix/smtpd[15160]: connect from >> smtp28.orange.fr[80.12.242.101] >> Apr 16 00:40:15 r13151 postfix/smtpd[15160]: NOQUEUE: reject: RCPT from >> smtp28.orange.fr[80.12.242.101]: 450 4.1.1 <catch(a)fakessh.eu>: Recipient >> address rejected: undeliverable address: User unknown in virtual alias >> table; from=<john.swilting(a)wanadoo.fr> to=<catch(a)fakessh.eu> proto=ESMTP >> helo=<smtp28.orange.fr> >> Apr 16 00:40:15 r13151 postfix/smtpd[15160]: disconnect from >> smtp28.orange.fr[80.12.242.101] >> >> >> >>

1 0

[Dovecot] version 1.2.11 creating symlink loop for config file
by Adrian Pepper 15 Apr '10

15 Apr '10

What is the following at the end of "main_init" in dovecot 1.2.11 (src/master/main.c) intended to accomplish? base_config_path = t_strconcat(settings_root->defaults->base_dir, "/"PACKAGE".conf", NULL); (void)unlink(base_config_path); if (symlink(configfile, base_config_path) < 0) { i_error("symlink(%s, %s) failed: %m", configfile, base_config_path); } } With our compilation parameters and runtime setup it ends up removing the working base_config_path file and making it instead a symlink to itself. (I.e. a non-working loop). Strangely, the one instance of dovecot continues and seems to mostly work. But (for now) I need to put back the proper dovecot.conf before I can restart it. I think that would be the expected result if strcmp(configfile, base_config_path) == 0. Is that in fact not expected to be the case? (See my initial question). I did not expect self-modifying of configuration like this. (See my initial question). Pattern matches (e.g. "config_path") find no apparent mention of this in recent postings. Adrian Pepper Computer Science Computing Facility David R. Cheriton School of Computer Science University of Waterloo arpepper(a)uwaterloo.ca ========================================================================= After the bug has happened... services110.cs# servers/dovecot -n # 1.2.11: /software/imap-dovecot-1.2/data/etc/dovecot/dovecot.conf Error: Can't open configuration file /software/imap-dovecot-1.2/data/etc/dovecot/dovecot.conf: Number of symbolic links encountered during path name traversal exceeds MAXSYMLINKS Fatal: Invalid configuration in /software/imap-dovecot-1.2/data/etc/dovecot/dovecot.conf services110.cs# services110.cs# servers/dovecot -n # 1.2.11: /software/imap-dovecot-1.2/data/etc/dovecot/dovecot.conf Warning: fd limit 256 is lower than what Dovecot can use under full load (more than 768). Either grow the limit or change login_max_processes_count and max_mail_processes settings # OS: SunOS 5.8 sun4u base_dir: /software/imap-dovecot-1.2/data/etc/dovecot/ protocols: imap imaps pop3 pop3s ssl_cert_file: /software/sslCerts/config/certs/dovecot.pem ssl_key_file: /software/sslCerts/config/certs/private/dovecot.key disable_plaintext_auth: no login_dir: /software/imap-dovecot-1.2/data/etc/dovecot/home/ login_executable(default): /software/imap-dovecot-1.2/libexec/dovecot/imap-login login_executable(imap): /software/imap-dovecot-1.2/libexec/dovecot/imap-login login_executable(pop3): /software/imap-dovecot-1.2/libexec/dovecot/pop3-login first_valid_uid: 100 last_valid_uid: 65535 first_valid_gid: 100 last_valid_gid: 65535 mail_location: mbox:~:INDEX=~/.imap:INBOX=/var/mail/%u mmap_disable: yes dotlock_use_excl: no mail_nfs_storage: yes mail_nfs_index: yes mail_executable(default): /software/imap-dovecot-1.2/libexec/dovecot/imap mail_executable(imap): /software/imap-dovecot-1.2/libexec/dovecot/imap mail_executable(pop3): /software/imap-dovecot-1.2/libexec/dovecot/pop3 mail_plugin_dir(default): /software/imap-dovecot-1.2/lib/dovecot/imap mail_plugin_dir(imap): /software/imap-dovecot-1.2/lib/dovecot/imap mail_plugin_dir(pop3): /software/imap-dovecot-1.2/lib/dovecot/pop3 namespace: type: private separator: / prefix: maildir_mail/ location: maildir:~/.maildir list: yes subscriptions: yes namespace: type: private separator: / location: mbox:~/.:INBOX=/var/mail/%u:INDEX=~/.imap1 inbox: yes list: yes subscriptions: yes namespace: type: private separator: / prefix: mail/ location: mbox:~/mail:INDEX=~/.imap2 list: yes subscriptions: yes lda: postmaster_address: postmaster(a)cs.uwaterloo.ca hostname: cs.uwaterloo.ca auth default: verbose: yes debug: yes passdb: driver: passwd-file args: /etc/shadow userdb: driver: passwd args: blocking=yes services110.cs# This is not a production setup; some/most configuration is experimental or speculative. "dovecot -n" does not appear to remove the config file.

4 3

[Dovecot] imap login fails with "no auth attempts"
by Niels Richter 15 Apr '10

15 Apr '10

Hi, I'm currently facing a strange behaviour of dovecot 1.2.10 (and before): after some days the login into our mail servers stops working for imap-connections. pop3-users can go on with using dovecot. Looking into the logfile I can only find: [...] imap-login: Disconnected (no auth attempts) imap-login: Disconnected: Shutting down (no auth attempts) [...] as a hint what is maybe happening. After restarting the system is working quite well again for some time. There is no extra load and not other special at the server running. The authentication is handled over mysql. This is our dovecot -n output (relevant parts): # 1.2.10: /usr/local/etc/dovecot.conf # OS: FreeBSD 7.2-STABLE i386 ufs protocols: imaps pop3s pop3 imap ssl_ca_file: /usr/local/etc/ssl/... ssl_cert_file: /usr/local/etc/ssl/... ssl_key_file: /usr/local/etc/ssl/... disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable(default): /usr/local/libexec/dovecot/imap-login login_executable(imap): /usr/local/libexec/dovecot/imap-login login_executable(pop3): /usr/local/libexec/dovecot/pop3-login mail_max_userip_connections(default): 20 mail_max_userip_connections(imap): 20 mail_max_userip_connections(pop3): 10 verbose_proctitle: yes first_valid_gid: 0 mail_access_groups: mail mail_privileged_group: mail mail_location: maildir:/usr/... mail_executable(default): /usr/local/libexec/dovecot/imap mail_executable(imap): /usr/local/libexec/dovecot/imap mail_executable(pop3): /usr/local/libexec/dovecot/pop3 mail_process_size: 512 mail_plugin_dir(default): /usr/local/lib/dovecot/imap mail_plugin_dir(imap): /usr/local/lib/dovecot/imap mail_plugin_dir(pop3): /usr/local/lib/dovecot/pop3 imap_client_workarounds(default): delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep imap_client_workarounds(imap): delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep imap_client_workarounds(pop3): namespace: type: private prefix: INBOX. inbox: yes list: yes subscriptions: yes Do you see anything which is unclear or unnecessary and could lead to a crash of the imap-login process after some time? Thanks for your help! -- Niels.

2 1

[Dovecot] Any secret dovecot 2.0 lda configuration documentation anywhere? =)
by Brandon Lamb 15 Apr '10

15 Apr '10

I successfully got imap/pop3 working using 2.0b4, now im trying to also get lda working with exim but im totally confused. I dont know what should run as what user, getting confused on auth sockets, what to call them or set or whatever. My exim runs as user/group mail/mail, dovecot has user dovecot but user mail for mail_user/group settings. At one point running deliver -d user -f user gave me an error userdb lookup not possible with only userdb prefetch. Here is my config at this particular point in time, probably will be hacked up again in a few seconds =P # 2.0.beta4: /etc/dovecot/dovecot.conf # OS: Linux 2.6.32-3-amd64 x86_64 Debian squeeze/sid auth_cache_negative_ttl = 3600 s auth_cache_size = 4096 auth_debug = yes auth_debug_passwords = yes auth_default_realm = olypen.com auth_mechanisms = plain login auth_socket_path = /var/run/dovecot/auth-userdb auth_username_format = %Lu auth_worker_max_count = 60 base_dir = /var/run/dovecot debug_log_path = /var/log/dovecot/debug.log default_vsz_limit = 536870912 deliver_log_format = %s from %f %$ disable_plaintext_auth = no dotlock_use_excl = yes first_valid_gid = 8 first_valid_uid = 8 fsync_disable = yes info_log_path = /var/log/dovecot/info.log last_valid_gid = 8 last_valid_uid = 8 lda_mailbox_autocreate = yes lda_mailbox_autosubscribe = yes log_path = /var/log/dovecot/dovecot.log log_timestamp = %Y-%m-%d %H:%M:%S login_log_format = %s login_log_format_elements = (%u) (%r) mail_access_groups = mail mail_debug = yes mail_gid = 8 mail_location = maildir:~/:INDEX=~/dovecot-index mail_log_prefix = %Ls <%u>: mail_plugins = quota mail_privileged_group = mail mail_uid = 8 mailbox_idle_check_interval = 60 s namespace { hidden = yes inbox = yes list = yes location = maildir:%h/:INDEX=~/dovecot-index prefix = separator = / type = private } namespace { hidden = yes list = no location = virtual:/etc/dovecot/virtual:INDEX=~/dovecot-index prefix = virtual/ separator = / type = private } passdb { args = /etc/dovecot/dovecot-sql.conf.ext driver = sql } plugin { autocreate = Trash autocreate2 = Spam autocreate3 = Sent autocreate4 = Drafts quota = maildir quota_rule2 = Trash:ignore quota_rule3 = Deleted Items:ignore quota_rule4 = Spam:ignore quota_warning = storage=95%% /etc/dovecot/quota-warning.sh 95 quota_warning2 = storage=80%% /etc/dovecot/quota-warning.sh 80 } postmaster_address = postmaster(a)olypen.com protocols = imap pop3 quota_full_tempfail = yes rejection_reason = Your message to <%t> with subject of <%s> was automatically rejected:%n%r rejection_subject = Mailbox Full: %s sendmail_path = /usr/sbin/exim service auth { unix_listener /var/spool/exim/auth-userdb { group = mail mode = 0666 user = } } service imap-login { inet_listener imap { port = 143 } process_min_avail = 5 service_count = 0 vsz_limit = 134217728 } service imap { process_limit = 1024 vsz_limit = 1073741824 } service pop3-login { inet_listener pop3 { port = 110 } process_min_avail = 5 service_count = 0 vsz_limit = 134217728 } service pop3 { process_limit = 1024 vsz_limit = 1073741824 } ssl = no userdb { driver = prefetch } valid_chroot_dirs = /mail verbose_proctitle = yes protocol pop3 { mail_max_userip_connections = 1 mail_plugins = $mail_plugins virtual pop3_client_workarounds = outlook-no-nuls oe-ns-eoh pop3_enable_last = no pop3_lock_session = no pop3_logout_format = retr=%r/%b, del=%d/%m, size=%s pop3_no_flag_updates = yes pop3_reuse_xuidl = no pop3_save_uidl = yes pop3_uidl_format = %08Xu%08Xv } protocol lda { auth_socket_path = auth-master info_log_path = /var/log/dovecot/deliver.log log_path = /var/log/dovecot/deliver.log mail_plugins = quota } protocol imap { imap_client_workarounds = delay-newmail netscape-eoh tb-extra-mailbox-sep imap_idle_notify_interval = 120 s imap_logout_format = bytes=%i/%o imap_max_line_length = 65536 mail_max_userip_connections = 15 mail_plugins = $mail_plugins imap_quota autocreate virtual }

1 0