- dovecot - dovecot.org

Re: [Dovecot] Shared namespace
by Daniel L. Miller 08 May '12

08 May '12

On 5/7/2012 6:46 AM, Daniel L. Miller wrote: > On 5/6/2012 8:34 PM, Daniel L. Miller wrote: >> I'm not sure what's triggered it - I THINK it has something to do >> with a force-resync, but not sure. >> >> I'm no longer viewing my full list of shared mailboxes. I used to >> have a number of users shown - they had all had their ACL's set >> individually and were shown in the list. "doveadm acl debug" shows >> they SHOULD be shown - at least I think so, as the output for the >> mailboxes that DO appear is identical for those that don't. >> >> Where should I look for breakage? > > After upgrading to 2.1.6 - mailboxes are back. Not sure I want to TRY > to break it again... > Ok - I broke it again. Tried changing mail_shared_explicit_inbox to no - problem mailboxes disappeared again. Changed back to yes - they came back. Problem with both Thunderbird and Roundcube as clients. -- Daniel

2 9

[Dovecot] courier to dovecot
by Grafibit Rendszergazda 08 May '12

08 May '12

Dear All, I know there were many threads on this topic. I was searching the archive and the net also, but did not find a solution on my problem. I hope you can help me. I am migrating from an old server using courier to a new server with dovecot and i can not manage to have the same uids after the migration. After copying the message files from the old server to the new, and executing the script from the wiki page the messages are still re-downloaded by the mail clients. After telnet and logging in to the old server port 110, I got the following on the command *uidl*: 1 UID377-1278276092 2 UID387-1278276092 3 UID394-1278276092 4 UID498-1278276092 5 UID499-1278276092 6 UID564-1278276092 On the new server i set up the *pop3_uidl_format = UID%u-%v* because it seems for me that the old server is using the same uidl format. Either if i execute the /./courier-dovecot-migrate.pl --to-dovecot --recursive --convert /command or i don't the uidls on the new server won't be the same. Both the %u and %v numbers are differ. Please give me any hints how to continue the investigation to have a successful migration. Thank you in advance, Tamás * dovecot -n* # 1.2.15: /etc/dovecot/dovecot.conf # OS: Linux 3.2.0-0.bpo.2-amd64 x86_64 Debian 6.0.4 protocols: imap imaps pop3 pop3s managesieve ssl_cert_file: /var/lib/dtc/etc/ssl/dovecot/new.cert.cert ssl_key_file: /var/lib/dtc/etc/ssl/dovecot/new.cert.key disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable(default): /usr/lib/dovecot/imap-login login_executable(imap): /usr/lib/dovecot/imap-login login_executable(pop3): /usr/lib/dovecot/pop3-login login_executable(managesieve): /usr/lib/dovecot/managesieve-login mail_location: maildir:%h/Maildir mbox_write_locks: fcntl dotlock mail_executable(default): /usr/lib/dovecot/imap mail_executable(imap): /usr/lib/dovecot/imap mail_executable(pop3): /usr/lib/dovecot/pop3 mail_executable(managesieve): /usr/lib/dovecot/managesieve mail_plugins(default): quota imap_quota mail_plugins(imap): quota imap_quota mail_plugins(pop3): quota mail_plugins(managesieve): mail_plugin_dir(default): /usr/lib/dovecot/modules/imap mail_plugin_dir(imap): /usr/lib/dovecot/modules/imap mail_plugin_dir(pop3): /usr/lib/dovecot/modules/pop3 mail_plugin_dir(managesieve): /usr/lib/dovecot/modules/managesieve pop3_uidl_format: UID%u-%v managesieve_logout_format(default): bytes=%i/%o managesieve_logout_format(imap): bytes=%i/%o managesieve_logout_format(pop3): bytes=%i/%o managesieve_logout_format(managesieve): bytes ( in=%i : out=%o ) namespace: type: private separator: . prefix: INBOX. inbox: yes list: yes subscriptions: yes lda: postmaster_address: mail(a)exemple.com mail_plugin_dir: /usr/lib/dovecot/modules/lda auth_socket_path: /var/run/dovecot-auth-master log_path: /var/log/sieve.log info_log_path: /var/log/sieve.info mail_plugins: sieve quota auth default: passdb: driver: sql args: /var/lib/dtc/etc/dovecot-mysql.conf userdb: driver: prefetch userdb: driver: sql args: /var/lib/dtc/etc/dovecot-mysql.conf socket: type: listen client: path: /var/run/dovecot/dovecot-auth-client mode: 438 master: path: /var/run/dovecot-auth-master mode: 384 user: dtc plugin: sieve: ~/.dovecot.sieve sieve_dir: ~/sieve quota: maildir

3 3

Re: [Dovecot] Upgrading known problems (2.0 to 2.1) ?
by Timo Sirainen 08 May '12

08 May '12

Yes, that's possible. The IMAP/POP3 servers behind Dovecot director wouldn't even need to be Dovecots at all. Although you really want to use v2.1 director since it has a lot of fixes compared to v2.0. On 8.5.2012, at 0.02, Jean Michel wrote: > Thanks, an other thing is that I didnt saw anything about a cenario like this, its possible to have a director working on 2.0 and a backend with 2.1 simultaneously and vice-versa? > > 2012/5/7 Timo Sirainen <tss(a)iki.fi> > On 7.5.2012, at 23.21, Jean Michel wrote: > > > Is there any known/possible problem while upgrading from dovecot 2.0 to 2.1 > > ? > > Did anybody had any trouble with this ? > > http://wiki2.dovecot.org/Upgrading/2.1 lists all known problems. > > > > > -- > ------------------------------ > Jean Michel Feltrin >

1 0

[Dovecot] Upgrading known problems (2.0 to 2.1) ?
by Jean Michel 07 May '12

07 May '12

Is there any known/possible problem while upgrading from dovecot 2.0 to 2.1 ? Did anybody had any trouble with this ? ------------------------------ Jean Michel Feltrin

2 1

[Dovecot] dsync with virtual users
by Ken Stevenson 07 May '12

07 May '12

I'm using dovecot v2.0.16

3 3

[Dovecot] Panic in index-sync.c: line 25 on dovecot-1.2.17
by Leon Meßner 07 May '12

07 May '12

Hi, with dovecot-1.2.17 (yes i know it's old) i'm getting the folowing error after doing a system upgrade. May 7 22:04:26 mail dovecot: deliver(elon): Panic: file index-sync.c: line 25 (index_mailbox_set_recent_uid): assertion failed: (seq_range_exists(&ibox->recent_flags, uid)) May 7 22:04:27 mail kernel: pid 79989 (deliver), uid 4873: exited on signal 6 (core dumped) this only happens to one user and only with mail from one mailinglist (this one). Mails do get filtered by sieve. After some delivery attempts the mail finaly gets stored alright in the maildir. May 7 22:23:27 mail dovecot: deliver(elon): sieve: msgid=<assp.047487a390.4FA82AA5.60907 09(a)amfes.com>: stored mail into mailbox 'dovecot' May 7 22:23:27 mail postfix/local[82289]: 7DED511402: to=<elon(a)physik.tu-berlin.de>, orig_to=<l.messner(a)physik.tu-berlin.de>, relay=local, delay=1141, delays=1141/0.01/0/0.26, dsn=2.0.0, status=sent (delivered to command: /usr/local/libexec/dovecot/deliver) May 7 22:23:27 mail postfix/qmgr[1303]: 7DED511402: removed After that a new mail does the same dance (perhaps until the next one to this ML arrives). The maildirs are on NFS and in this case only beeing used with my instance of mutt and two other postfix+dovecot servers which are quite idle. Whats the way to debug this ? Thanks, Leon

2 2

[Dovecot] Proxying for some users and having the rest local?
by Dr Josef Karthauser 07 May '12

07 May '12

I'm migrating a load of mailboxes from cyrus to a new dovecot server. What I'd like to do is to switch the DNS to dovecot, and set things up so that all old imap mailboxes are accessed through the dovecot proxy. That way I can migrate them one at a time independent of DNS issues. I've got a load of users specified in mysql, and that configuration works well for local mailboxes. How do I configure things so that I can switch proxy access on and off? What's I've got so far is: password_query = SELECT user, domain, password, "masteruser" as master, "masterpass" as pass, proxy, "remoteimap.server" as host FROM mailboxes \ WHERE user = '%n' AND isMailbox AND active AND domain = '%d' I imagined that I could switch the proxy per user by setting proxy='n' or proxy=null for any users that want to be local, and proxy='y' for any users that I want to forward on to the original server. It doesn't appear to work that way though. Irrespective of the proxy value dovecot appears to try and proxy all users. What am I doing wrong? I can't work it out from the docs, and the source will take quite a bit of studying to get to the bottom of it all. Thanks :), Joe

2 3

Re: [Dovecot] mdbox packing
by Daniel L. Miller 07 May '12

07 May '12

On 5/6/2012 8:12 PM, Daniel L. Miller wrote: > On 5/6/2012 8:09 PM, Daniel L. Miller wrote: >> On 4/30/2012 5:23 PM, Timo Sirainen wrote: >>> On Mon, 2012-04-30 at 12:43 -0700, Daniel L. Miller wrote: >>>> Is there a way to manually force an mdbox storage to be rebuilt >>>> into new >>>> files? Particularly files of the maximum size? >>> You could do it with dsync. >>> >>> >> dsync seems to do a marvelous job - new question.? >> >> Given a mailbox with a larger number of older mails - assuming any >> new mails will have later dates - will the new mail storage files be >> identical? I'm probably not saying the right - let me try this: >> >> 1. "dsync backup mdbox" (with appropriate args) is run for a given >> user. >> 2. "dsync mirror mdbox" is done just to catch up. >> 3. Old mailstore moved off. >> 4. New mailstore moved to active location. >> >> I'm sure there's a safer way to do the above - but I've got a >> low-traffic site and I can just shut down mail service altogether for >> a few minutes if I want during this. >> >> Now - repeat the above four steps. Will mail files m.1 through >> m.(n-1) be identical to the last run? Is this a valid packing >> strategy prior to performing an rsync type backup - assuming no >> changes are being made to the archived mails between pack runs? > > Given the above, and SIS - if the backup is performed in the same > spool, i.e. original is /var/mail/domain/user, backup is > /var/mail/domain/user-new - will this result in any problems with > SIS? If after the pack and rename operation, the old mail tree is > simply deleted - will that leave SIS files unreferenced if the user > later deletes the messages from their active store? > And yet another one - dsync does not APPEAR to be copying ACL's in this process. -- Daniel

2 1

[Dovecot] This binary should probably be called with process group set to (vmail) instead of (userid)
by jeep＠rahul.net 07 May '12

07 May '12

Im sorry to ask another basic question, but Ive tried to find this answer for several hours now and it eludes me. Im getting the following when dovecot tries to deliver an email: May 6 19:29:21 mydomain dovecot: lda: Debug: auth input: jeff home=/opt/imapdata/j/jeff/INBOX uid=1001 gid=999 May 6 19:29:21 mydomain dovecot: lda(jeff): Fatal: setgid(999(vmail) from userdb lookup) failed with euid=999(vmail), gid=500(jeff), egid=500(jeff): Operation not permitted (This binary should probably be called with process group set to 999(vmail) instead of 500(jeff)) I cannot seem to find where its trying to call dovecot-lda with uid(jeff) at? AFAIK Ive set everything up to use/deliver as 'vmail'. 'jeff' is the only person on this box with a unix account which is uid 500, but how do I make it use vmail instead? Thanks for the help! Jeff postfix master.cf: virtual_transport = vmail mailbox_command=/opt/dovecot/libexec/dovecot/dovecot-lda -f "$SENDER" -a "$RECIPIENT" dovecot -n: # 2.1.5: /opt/dovecot/etc/dovecot/dovecot.conf # OS: Linux 2.6.35.14-106.fc14.x86_64 x86_64 Fedora release 14 (Laughlin) ext4 auth_debug = yes auth_debug_passwords = yes auth_mechanisms = plain login auth_socket_path = /opt/dovecot215/var/run/dovecot/auth-userdb auth_verbose = yes auth_verbose_passwords = plain default_client_limit = 225 default_internal_user = vmail disable_plaintext_auth = no first_valid_gid = 999 first_valid_uid = 999 listen = * lock_method = flock mail_debug = yes mail_gid = vmail mail_home = /opt/imapdata/vmailhome mail_location = mbox:/opt/imapdata/%1n/%n:INDEX=/opt/imapdata/%1n/%n mail_privileged_group = vmail mail_uid = vmail mbox_lock_timeout = 1 mins mbox_write_locks = fcntl namespace { inbox = yes location = prefix = separator = / type = private } passdb { args = /opt/dovecot/etc/dovecot/conf.d/dovecot-sql.conf.ext driver = sql } plugin { mail_log_events = delete undelete expunge copy mailbox_delete mailbox_rename mail_log_fields = uid box msgid size } protocols = imap service auth { inet_listener { port = 12345 } unix_listener auth-userdb { group = vmail mode = 0666 user = vmail } user = $default_internal_user } service imap-login { inet_listener imap { port = 143 } service_count = 1 } ssl_cert = </etc/pki/dovecot/certs/dovecot.pem ssl_key = </etc/pki/dovecot/private/dovecot.pem userdb { args = /opt/dovecot/etc/dovecot/conf.d/dovecot-sql.conf.ext driver = sql } userdb { args = /opt/dovecot/etc/dovecot/conf.d/dovecot-sql.conf.ext driver = sql } userdb { args = /opt/dovecot/etc/dovecot/conf.d/dovecot-sql.conf.ext driver = sql } protocol imap { imap_idle_notify_interval = 1 mins imap_max_line_length = 64 k mail_max_userip_connections = 5 } /mf/home/jeep/shell/.signature

2 2

Re: [Dovecot] Shared namespace
by Daniel L. Miller 07 May '12

07 May '12

On 5/6/2012 8:34 PM, Daniel L. Miller wrote: > I'm not sure what's triggered it - I THINK it has something to do with > a force-resync, but not sure. > > I'm no longer viewing my full list of shared mailboxes. I used to > have a number of users shown - they had all had their ACL's set > individually and were shown in the list. "doveadm acl debug" shows > they SHOULD be shown - at least I think so, as the output for the > mailboxes that DO appear is identical for those that don't. > > Where should I look for breakage? After upgrading to 2.1.6 - mailboxes are back. Not sure I want to TRY to break it again... -- Daniel

1 0