January 2016 - dovecot

2.2.13 fts-squat corrupt uidlist - wrong indexid
by dilieto＠lineone.net 09 Jan '16

09 Jan '16

I have the same problem described in this old thread: http://dovecot.org/list/dovecot/2012-April/135162.html I have done some debugging, by adding log messages to squat_uidlist_map_header and squat_trie_refresh. I have found that the "wrong indexid" warning is always triggered under the following conditions - uidlist->hdr.indexid != 0 - uidlist->trie->hdr.indexid == 0 - squat_trie_is_file_stale(trie) == 0 In other words it looks like the squat_uidlist_map_header function is called with an uninitialized trie, an initialized header, and a current (not stale) file. Because of the latter, squat_trie_refresh never calls squat_trie_open, the trie stays uninitialized and later triggers the warning. In addition the search results will only include mails added to the mailbox after the warning. By replacing the call to squat_trie_refresh with squat_trie_open in squat_uidlist_map_header, the warning no longer happens. Patch is attached. I am not sure this is the correct way of fixing the problem, but it works reliably.

1 0

Released Pigeonhole v0.4.11 for Dovecot v2.2.21
by Stephan Bosch 08 Jan '16

08 Jan '16

Hello Dovecot users, Here is the final v0.4.11 release of Pigeonhole for Dovecot v2.2.21. Nothing of consequence changed since the RC. Changelog v0.4.11: - Sieve mime extension: Fixed the header :mime :anychild test to work properly outside a foreverypart loop. - Several fixes in message body part handling: - Fixed assert failure occurring when text extraction is attempted on an empty or broken text part. - Fixed assert failure in handling of body parts that are converted to text. - Fixed header unfolding for (mime) headers parsed from any mime part. - Fixed trimming for (mime) headers parsed from any mime part. - Fixed erroneous changes to the message part tree structure performed when re-parsing the message. - LDA Sieve plugin: Fixed logging of actions; sometimes the configured log format was not followed. - LDA Sieve plugin: Fixed bug in error handling of script storage initialization. - Sieve Extprograms plugin: Ignored ENOTCONN error in shutdown(fd, SHUT_WR) call. - Fixed duplication of discard actions in the script result. Each discard was counted as a separate action, which means that action limit would be crossed too early. - Made sure that quota errors never get logged as errors in syslog. - Fixed handling of implicit keep for a partially executed transaction that yielded a temporary failure. - Fixed handling of global errors. If master and user error handler were identical, in some cases the log message could be lost. - Fixed AIX compile issue in message body parser. The release is available as follows: http://pigeonhole.dovecot.org/releases/2.2/dovecot-2.2-pigeonhole-0.4.11.ta… http://pigeonhole.dovecot.org/releases/2.2/dovecot-2.2-pigeonhole-0.4.11.ta… Refer to http://pigeonhole.dovecot.org and the Dovecot v2.x wiki for more information. Have fun testing this release and don't hesitate to notify me when there are any problems. Regards, -- Stephan Bosch stephan(a)rename-it.nl

1 0

doveadm search -A tries to create mailboxes
by Bill Shirley 08 Jan '16

08 Jan '16

[0:root@elmo ~]$ rpm -q dovecot dovecot-2.2.18-2.fc22.x86_64 I got a surprise today when I tried the doveadm search function from: http://wiki2.dovecot.org/Plugins/Expire [0:root@elmo ~]$ doveadm search -A mailbox Trash savedbefore 30d doveadm(clamscan): Error: User initialization failed: Namespace '': mkdir(//Maildir) failed: Permission denied (euid=982(clamscan) egid=289(clamscan) missing +w perm: /, dir owned by 0:0 mode=0555) doveadm(clamscan): Error: User init failed doveadm(nm-openconnect): Error: User initialization failed: Namespace '': mkdir(//Maildir) failed: Permission denied (euid=986(nm-openconnect) egid=293(nm-openconnect) missing +w perm: /, dir owned by 0:0 mode=0555) doveadm(nm-openconnect): Error: User init failed doveadm(openvpn): Error: User initialization failed: Namespace '': mkdir(/etc/openvpn/Maildir) failed: Permission denied (euid=987(openvpn) egid=296(openvpn) missing +w perm: /etc/openvpn, dir owned by 0:0 mode=0755) doveadm(openvpn): Error: User init failed .. plus several more I wouldn't think the search function (or any operation) would try to create mailboxes. These are system user ids that don't have or need mailboxes. Is this correct? Bill

3 5

doveadm import - strange warning
by Ralf Hildebrandt 08 Jan '16

08 Jan '16

I'm using the doveadm import command to extract mails from my backup: % doveadm import -u restore(a)backup.invalid mdbox:/home/copymail/mdbox "" \ mailbox INBOX header Delivered-To backup+$SEARCHPATTERN(a)backup.invalid This works like a charm, I'm getting all the mails, and I can access the newly generated mailbox. But when executing the command I'm getting an annoying warning: doveadm(restore(a)backup.invalid): Error: chdir(/root/) failed: Permission denied (euid=1001(copymail) egid=1001(copymail) missing +x perm: /root, dir owned by 0:0 mode=0700) doveadm(restore(a)backup.invalid): Error: chdir(/root) failed: Permission denied I'm using a file based authentication: # cat passwd restore@backup.invalid:{SHA}+/...the.hash...:1001:1001:(gecos):/home/restore:: So why is "restore(a)backup.invalid" being associated with /root? I'm using 2:2.2.18-1~auto+91, but the warning has been around for ages :) -- [*] sys4 AG http://sys4.de, +49 (89) 30 90 46 64 Franziskanerstraße 15, 81669 München Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263 Vorstand: Patrick Ben Koetter, Marc Schiffbauer Aufsichtsratsvorsitzender: Florian Kirstein

5 7

strange interaction with MacOSX calendar server
by Jim Reid 08 Jan '16

08 Jan '16

I’ve installed and configured Apple’s calendar server. It seems to be working just fine. However it keeps trying to speak to my IMAP server — don’t know why — and the authentication attempts fail every 30 seconds or so. Here’s what’s in the calendar server logs: 2016-01-07 22:42:38+0000 [-] [caldav-1] [IMAP4DownloadProtocol (TLSMemoryBIOProtocol),client] [txdav.caldav.datastore.scheduling.imip.inbound.IMAP4DownloadProtocol#error] IMAP login failed for com.apple.calendarserver 2016-01-07 22:43:08+0000 [-] [caldav-1] [IMAP4DownloadProtocol (TLSMemoryBIOProtocol),client] [txdav.caldav.datastore.scheduling.imip.inbound.IMAP4DownloadProtocol#error] IMAP login failed for com.apple.calendarserver 2016-01-07 22:43:39+0000 [-] [caldav-1] [IMAP4DownloadProtocol (TLSMemoryBIOProtocol),client] [txdav.caldav.datastore.scheduling.imip.inbound.IMAP4DownloadProtocol#error] IMAP login failed for com.apple.calendarserver And in dovecot’s logs I see: Jan 7 22:47:43 hutch dovecot[63067]: imap-login: Disconnected (auth failed, 2 attempts in 4 secs): user=<com.apple.calendarserver>, method=PLAIN, rip=195.54.233.70, lip=195.54.233.70, TLS, session=<wkMAQsYoVADDNulG> Jan 7 22:48:10 hutch dovecot[63067]: auth-worker(65378): Error: pam(com.apple.calendarserver,195.54.233.70): pam_acct_mgmt() failed: permission denied Jan 7 22:48:14 --- last message repeated 1 time --- Jan 7 22:48:14 hutch dovecot[63067]: imap-login: Disconnected (auth failed, 2 attempts in 4 secs): user=<com.apple.calendarserver>, method=PLAIN, rip=195.54.233.70, lip=195.54.233.70, TLS, session=<dSvTQ8YoXQDDNulG> Jan 7 22:48:40 hutch dovecot[63067]: auth-worker(65378): Error: pam(com.apple.calendarserver,195.54.233.70): pam_acct_mgmt() failed: permission denied Jan 7 22:48:44 --- last message repeated 1 time — I tweaked dovecot to use AUTH PLAIN as well as CRAM-MD5. [The MD5 stuff has always worked fine for SMTP and IMAP over TLS.] I think /etc/pam.d/dovecot is OK too: % cat /etc/pam.d/dovecot # # as documented on Dovecot wiki auth required pam_opendirectory.so try_first_pass account required pam_nologin.so account required pam_opendirectory.so password required pam_opendirectory.so Here’s my dovecot.conf: # 2.2.5: dovecot.conf # OS: Darwin 12.5.0 x86_64 auth_mechanisms = plain login cram-md5 base_dir = /var/run/dovecot/ listen = *, [::] mail_debug = yes mail_privileged_group = mail mbox_write_locks = fcntl namespace inbox { hidden = yes inbox = yes list = no location = mbox:~/mail:INBOX=/var/mail/%u prefix = "#mbox/" separator = / type = private } namespace { inbox = no list = yes location = maildir:/var/imap/%u:LAYOUT=fs prefix = separator = / type = private } passdb { driver = pam args = failure_show_msg=yes } passdb { args = /usr/local/etc/dovecot-md5 driver = passwd-file } protocols = imap service auth { executable = /usr/local/libexec/dovecot/auth unix_listener /var/spool/postfix/private/auth { group = postfix mode = 0660 user = postfix } user = root } service imap-login { client_limit = 32 executable = /usr/local/libexec/dovecot/imap-login inet_listener imap { port = 0 } process_limit = 128 process_min_avail = 3 service_count = 1 vsz_limit = 64 M } service imap { executable = /usr/local/libexec/dovecot/imap process_limit = 32 } service pop3-login { client_limit = 32 process_limit = 128 process_min_avail = 3 service_count = 1 vsz_limit = 64 M } service pop3 { process_limit = 32 } ssl_cert = </usr/local/etc/ssl/certs/dovecot.pem ssl_cipher_list = ALL:!LOW:!SSLv2 ssl_key = </usr/local/etc/ssl/private/dovecot.pem syslog_facility = local4 userdb { driver = passwd } protocol imap { mail_max_userip_connections = 10 mail_plugin_dir = /usr/local/lib/dovecot/imap } protocol lda { postmaster_address = postmaster(a)example.com } Can someone please explain what’s going on and how to fix/troubleshoot it? Thanks. The endless error messages in the log don’t seem to be causing any problems to either the mail setup or the calendaer server. Famous last words… However they are a bit annoying and probably point to an underlying problem or misconfiguration that really should be fixed. By the way, the password used by the calendar server is for my MacOSX login (which should be getting picked up via PAM, right?) while a different one is used for the SMTP/IMAP MD5 challenge-response. I hope someone can clear up my confusion.

1 0

Mapping of INBOX.INBOX breaks imapc_list_prefix=INBOX?
by Peer Heinlein 07 Jan '16

07 Jan '16

We're running a huge imapc-Migration from an old Courier-IMAP to Dovecot. Courier always uses "INBOX."-prefix for his folders, so we set namespace inbox { [...] separator = . prefix = INBOX. [...] } on our new Dovecot 2.2.20 system. And also we added: imapc_list_prefix=INBOX to avoid the dublication of the INBOX prefix. At the end, doveadm/dsync/imapc see "INBOX.Virus", adds his Prefix, has "INBOX.INBOX.Virus", deletes his imapc_list_prefix and the end we have "INBOX.Virus" again which is identical to the source folder from Courier. Works perfect BUT: It looks like there is a "hidden feature" in Dovecot that always maps "INBOX.INBOX" back to "INBOX". For me that looks like a workaround for broken mailclients to fix some of the common problems after changing the namespace prefix. I never heard about that "hidden feature", but even if I don't (!) have INBOX.INBOX: root@imap01:~/MIGRATION # doveadm mailbox list -u test INBOX INBOX.Virus INBOX.Spam INBOX.Trash I'm always able to access INBOX.INBOX which is identical to INBOX: root@imap01:~/MIGRATION # doveadm fetch -u test uid mailbox INBOX.INBOX uid: 1 Looks like there is such a magic-mapping and looks like this "hard coded mapping" (?) breaks the imapc_list_prefix-magic. If the user has on the SOURCE-server (Courier): Source-Server: INBOX (5 messages) INBOX.Inbox (3 messages) and I do migrate that user using doveadm/dsync/imapc, this user has on the DESTINATION-server (Dovecot) Destination-server: INBOX (5 messages) INBOX.Inbox (5 messages from "real INBOX") and his 3 messages from INBOX.Inbox are lost. I can NOT find any errors in the debug/error-output of doveadm/dsync/imapc. From the view of doveadm everything worked perfect as you can see in the attached logfile. Or am I completly wrong?! Peer -- Heinlein Support GmbH Schwedter Str. 8/9b, 10119 Berlin http://www.heinlein-support.de Tel: 030 / 405051-42 Fax: 030 / 405051-19 Zwangsangaben lt. §35a GmbHG: HRB 93818 B / Amtsgericht Berlin-Charlottenburg, Geschäftsführer: Peer Heinlein -- Sitz: Berlin

2 4

doveadm proxy kick shows always "0 connections kicked"?!
by Peer Heinlein 06 Jan '16

06 Jan '16

On Dovecot 2.2.20: There *IS* a user connected: root@mailproxy02:~ # doveadm proxy list | grep testuser username proto src ip dest ip port testuser imap xx.xx.xx.xx yy.yy.yy.yy 143 But if I kick him there are always "0 connections kicked": root@mailproxy02:~ # doveadm proxy kick testuser 0 connections kicked ...even if the user's gone after I kicked him. Peer -- Heinlein Support GmbH Schwedter Str. 8/9b, 10119 Berlin http://www.heinlein-support.de Tel: 030 / 405051-42 Fax: 030 / 405051-19 Zwangsangaben lt. §35a GmbHG: HRB 93818 B / Amtsgericht Berlin-Charlottenburg, Geschäftsführer: Peer Heinlein -- Sitz: Berlin

1 0

allow_nets=local in passdb gets "auth: Panic"
by SATOH Fumiyasu 05 Jan '16

05 Jan '16

Hi, I have the following configuration in my dovecot.conf for Dovecot 2.2.21: passdb { driver = ldap args = /etc/dovecot/dovecot-ldap.conf.ext default_fields = allow_nets=local,127.0.0.1,10.255.1.0/24 } This triggers "auth: Panic" on POP3/IMAP logins as the below: Dec 22 14:57:39 localhost dovecot: auth: ldap(u0000,::1,<oiF8SHYngqsAAAAAAAAAAAAAAAAAAAAB>): allow_nets: Invalid network 'local' Dec 22 14:57:39 localhost dovecot: auth: Panic: file net.c: line 1137 (net_is_in_network): assertion failed: (IPADDR_IS_V6(ip) == IPADDR_IS_V6(net_ip)) `doveadm auth test <username> <password>` is no problem as expected. What's wrong? Regards, -- -- Name: SATOH Fumiyasu @ OSS Technology Corp. (fumiyas @ osstech co jp) -- Business Home: http://www.OSSTech.co.jp/ -- GitHub Home: https://GitHub.com/fumiyas/ -- PGP Fingerprint: BBE1 A1C9 525A 292E 6729 CDEC ADC2 9DCA 5E1C CBCA

2 3

Replication - user permissions
by Petter Gunnerud 05 Jan '16

05 Jan '16

When setting up dsync for replication, what should the user permissions be for sync over ssh? I'm running virtual users only. Postfix and Dovecot services run as mail:mail. All the maildir folders are owned by mail:mail and permissions are 700. User mail is not allowed login. So whats the best practice in respect to security to allow for dsync over ssh? So of the options I consider: 1) change postfix/dovecot settings so that maildirs are created with 770 permissions, then create a user dsync:mail that is allowed for ssh login. 2) permit user mail to login using ssh 3)go with tcp sync rather than ssh Other suggestions? PG

1 0

Replication - hostname issue
by Petter Gunnerud 05 Jan '16

05 Jan '16

I have postfix/dovecot setup on a virtual gentoo server. It's been in service for almost two years without any issues. Now I want to set up a spare server to replicate mails from the one running. I copied the vm files to the second host server, changed the ip, hostname and hosts file for the copy and followed the dovecot replication doc. The last point in the doc tells to rundovecot --hostdomainto make sure the hostnames differ... To my surprise the command returned "localhost" on both servers. How do I set the hostname for dovecot service?(The hostname command returns the servers hostnames.)Will a change in dovecot hostname make all clients redownload all their mail? A second question. When configuring dovecot replication. Should the settings be done on the master server or the stand by server, or both? (Doc doesn't say anything regarding this) PG

2 2