December 2015 - dovecot

Sieve script replication glitches
by Apollon Oikonomopoulos 09 Dec '15

09 Dec '15

Hi, I'm running a replicated setup on Debian with Dovecot 2.2.18 and Pigeonhole 0.4.8 and run into some glitches with sieve script syncing. Replication is implemented via DSync/doveadm over SSL. While e-mail replication works as expected, sieve script replication is not always successful. In general I have observed the following behaviour re. sieve script replication: - Sometimes sieve scripts are replicated but have an mtime of 0 (Jan 1 1970) on the destination. - Subsequent updates to these "bad" scripts are not replicated at all. - Sometimes after creating *another* sieve script (which may or may not get the correct timestamp), the "bad" script's timestamp/contents are corrected on the receiving side. The following is a minimal real-world example of what happens (including transaction log entries): 1. On the master create a new sieve script: cat test.sieve | doveadm sieve put -u user(a)example.com test1 Sieve storage on the master: -rw------- 1 vmail vmail 2001 Dec 9 15:53 test1.sieve Transaction log on the master: record: offset=44756, type=attribute-update (ext), size=72, modseq=514 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test1 : timestamp=2015-12-09 15:53:28 value_len=0 record: offset=44828, type=modseq-update (ext), size=20 - uid=0 modseq=723 Sieve storage on the slave: -rw------- 1 vmail vmail 2001 Jan 1 1970 test1.sieve Transaction log on the slave: record: offset=55356, type=attribute-update (ext), size=72, modseq=597 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test1 : timestamp=2015-12-09 15:53:29 value_len=0 record: offset=55428, type=attribute-update (sync), size=72, modseq=598 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test1 : timestamp=2015-12-09 15:53:29 value_len=0 2. Create a new script on the master: cat test.sieve | doveadm sieve put -u user(a)example.com test2 Sieve storage on the master: -rw------- 1 vmail vmail 2001 Dec 9 15:53 test1.sieve -rw------- 1 vmail vmail 2001 Dec 9 15:57 test2.sieve Transaction log on the master: record: offset=44848, type=attribute-update (ext), size=72, modseq=515 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test2 : timestamp=2015-12-09 15:57:14 value_len=0 record: offset=44920, type=modseq-update (ext), size=20 - uid=0 modseq=725 record: offset=44940, type=modseq-update (ext), size=20 - uid=0 modseq=728 Sieve storage on the slave: -rw------- 1 vmail vmail 2001 Dec 9 15:53 test1.sieve -rw------- 1 vmail vmail 2001 Dec 9 15:57 test2.sieve Transaction log on the slave: record: offset=55500, type=attribute-update (ext), size=72, modseq=599 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test2 : timestamp=2015-12-09 15:57:14 value_len=0 record: offset=55572, type=attribute-update (sync), size=72, modseq=600 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test2 : timestamp=2015-12-09 15:57:14 value_len=0 record: offset=55644, type=attribute-update (ext), size=72, modseq=601 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test1 : timestamp=2015-12-09 15:57:14 value_len=0 record: offset=55716, type=attribute-update (ext), size=72, modseq=602 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test2 : timestamp=2015-12-09 15:57:14 value_len=0 record: offset=55788, type=attribute-update (sync), size=132, modseq=603 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test1 : timestamp=2015-12-09 15:53:28 value_len=0 - add: private/vendor/vendor.dovecot/pvt/server/sieve/files/test2 : timestamp=2015-12-09 15:57:14 value_len=0 Apparently something gets messed up on the receiving side and strange things appear in the transaction log (e.g. test1 with a timestamp of 15:57:14 which it never had). The fact that it happens sometimes (around 30-40% of the time) likely indicates a race condition somewhere. Also I'm not at all sure sieve transactions should appear twice on the receiving side (ext and then sync); I see other entries (e.g. ext-intro) as a single "(ext) (sync)" record. Any ideas? Thanks, Apollon doveconf -n output follows: # 2.2.18: /etc/dovecot/dovecot.conf # Pigeonhole version 0.4.8 (0c4ae064f307+) # OS: Linux 3.16.0-4-amd64 x86_64 Debian 8.2 auth_mechanisms = plain login default_vsz_limit = 320 M mail_location = mdbox:~/mdbox:ALT=/srv/mail/alt/%d/%n mail_plugins = " acl notify replication acl" managesieve_notify_capability = mailto managesieve_sieve_capability = fileinto reject envelope encoded-character vacation subaddress comparator-i;ascii-numeric relational regex imap4flags copy include variables body enotify environment mailbox date index ihave duplicate namespace { location = maildir:/home/vmail/example.com/staging/Maildir:INDEX=/home/vmail/%d/%n/sta… prefix = Staging. subscriptions = no type = public } namespace inbox { inbox = yes location = mailbox Drafts { special_use = \Drafts } mailbox Junk { special_use = \Junk } mailbox Sent { special_use = \Sent } mailbox "Sent Messages" { special_use = \Sent } mailbox Trash { special_use = \Trash } prefix = separator = . type = private } passdb { args = /etc/dovecot/dovecot-sql.conf driver = sql } plugin { acl = vfile sieve = ~/.dovecot.sieve sieve_before = /home/vmail/global.sieve sieve_default = /home/vmail/global.sieve sieve_dir = ~/sieve } protocols = " imap sieve" service aggregator { fifo_listener replication-notify-fifo { user = vmail } unix_listener replication-notify { user = vmail } } service auth { unix_listener /var/spool/postfix/private/auth { group = postfix mode = 0660 user = postfix } unix_listener auth-userdb { group = vmail mode = 0777 } } service doveadm { inet_listener { port = 12345 ssl = yes } vsz_limit = 1 G } service imap-login { process_limit = 500 process_min_avail = 5 } service imap-postlogin { executable = script-login /usr/local/sbin/ldap-postlogin user = $default_internal_user } service imap { executable = imap imap-postlogin process_limit = 1024 } service replicator { process_min_avail = 1 unix_listener replicator-doveadm { mode = 0600 user = vmail } } ssl_cert = </etc/ssl/certs/dovecot.crt ssl_client_ca_dir = /etc/ssl/certs ssl_key = </etc/ssl/private/dovecot.key userdb { args = /etc/dovecot/dovecot-sql.conf driver = sql } protocol lda { log_path = /var/log/dovecot/deliver.log mail_plugins = sieve replication notify } protocol imap { mail_max_userip_connections = 500 }

1 0

v2.2.20 released
by Timo Sirainen 09 Dec '15

09 Dec '15

http://dovecot.org/releases/2.2/dovecot-2.2.20.tar.gz http://dovecot.org/releases/2.2/dovecot-2.2.20.tar.gz.sig This could be (one of) the last v2.2.x release. We're starting v2.3 development soon. + Added mailbox { autoexpunge=<time> } setting. See http://wiki2.dovecot.org/MailboxSettings for details. + ssl_options: Added support for no_ticket + imap/pop3/managesieve-login: Added postlogin_socket=path passdb extra field. This allows replacing the default service imap/pop3/managesieve {} settings for specific users (e.g. running their imap process via valgrind or strace). + doveadm fetch: Added date.sent/received/saved.unixtime + fs-posix: Added mode=auto parameter to set the created files' and directories' mode based on the parent dir if it has setgid-bit. + director: Support backends having hostnames, which makes it possible to verify their SSL certificates. - director: Directors' state became desynchronized if doveadm director commands were used to modify the same backend in multiple directors at the same time with conflicting changes. This fix includes some extra checks, which makes sure that if such a conflict still happens it's automatically fixed. In some situations such an automatic fix may now be unnecessarily triggered and an error logged. - director: Backend tags weren't working correctly. - ldap: tls_* settings weren't used for ldaps URIs. - ldap, mysql: Fixed setting connect timeout. - auth: userdb lookups via auth-worker couldn't change username - dsync: Fixed handling deleted directories. Make sure we don't go to infinite mailbox renaming loop. - imap: Fixed crash in NOTIFY when there were watched namespaces that didn't support NOTIFY. - imap: After SETMETADATA was used, various commands (especially FETCH) could have started hanging when their output was large. - stats: Idle sessions weren't refreshed often enough, causing stats process to forget them and log errors about unknown sessions when they were updated later. - stats: Fixed "Duplicate session ID" errors when LMTP delivered to multiple recipients and fts_autoindex=yes. - zlib plugin: Fixed copying causing cache corruption when zlib_save wasn't set, but the source message was compressed. - fts-solr: Fixed escaping Solr query parameters. - lmtp: quota_full_tempfail=yes was ignored with lmtp_rcpt_check_quota=yes

9 12

core-dump in imap (Dovecot 2.2.19)
by Peter Eriksson 09 Dec '15

09 Dec '15

Just found a coredump from the imap process for one of our users on dovecot 2.2.19 on a Solaris 10/x86 system: > Dec 8 14:33:17 mail dovecot: [ID 583609 mail.crit] imap(leijo): Fatal: master: service(imap): child 14465 killed with signal 11 (core dumped) Please find attached dovecot -n output and some gdb backtrace. It seems that cmd->client was NULL when dereferencing it at line 178 in imap-commands.c (in the function command_exec): > 178 cmd->bytes_in += i_stream_get_absolute_offset(cmd->client->input) - > 179 cmd_start_bytes_in; Please let me know if you need more information. I don't know what the users was doing at that specific time. --- [Ll.U] Peter Eriksson KITVS-IFM & ITI-NET IT.LiU.SE +46-70 518 2786

2 1

Autoexpunge and special user namespace
by Joseph Tam 09 Dec '15

09 Dec '15

I'm interested in replacing my cron script that expunges old deleted mail with the new feature announced in 2.2.20 > + Added mailbox { autoexpunge=<time> } setting. See > http://wiki2.dovecot.org/MailboxSettings for details. To use it, I need to specify which mailbox needs them using the namespace configuration. For example namespace inbox { mailbox Trash { auto = no special_use = \Trash autoexpunge = 7d mailbox_list_index = yes } mailbox trash { auto = no special_use = \Trash autoexpunge = 7d mailbox_list_index = yes } mailbox "Deleted Messages" { auto = no special_use = \Trash autoexpunge = 7d mailbox_list_index = yes } mailbox "Deleted Items" { auto = no special_use = \Trash autoexpunge = 7d mailbox_list_index = yes } ... } Different readers use different mailboxes, so I include the common ones. Some questions: 1) Do I need the special_user configuration at all? Will some clients start using different mailboxes to store deleted mail if I keep the special_use lines? 2) I also expunge old mail in INBOX marked as deleted -- is there any way to handle this case? Create virtual mailbox? Joseph Tam <jtam.home(a)gmail.com>

2 1

v2.2.20 release candidate released
by Timo Sirainen 08 Dec '15

08 Dec '15

http://dovecot.org/releases/2.2/rc/dovecot-2.2.20.rc1.tar.gz http://dovecot.org/releases/2.2/rc/dovecot-2.2.20.rc1.tar.gz.sig v2.2.20 probably will be released tomorrow or maybe during weekend. + Added mailbox { autoexpunge=<time> } setting. See http://wiki2.dovecot.org/MailboxSettings for details. + ssl_options: Added support for no_ticket + imap/pop3/managesieve-login: Added postlogin_socket=path passdb extra field. This allows replacing the default service imap/pop3/managesieve {} settings for specific users (e.g. running their imap process via valgrind or strace). + doveadm fetch: Added date.sent/received/saved.unixtime + fs-posix: Added mode=auto parameter to set the created files' and directories' mode based on the parent dir if it has setgid-bit. + director: Support backends having hostnames, which makes it possible to verify their SSL certificates. - director: Directors' state became desynchronized if doveadm director commands were used to modify the same backend in multiple directors at the same time with conflicting changes. This fix includes some extra checks, which makes sure that if such a conflict still happens it's automatically fixed. In some situations such an automatic fix may now be unnecessarily triggered and an error logged. - director: Backend tags weren't working correctly. - ldap: tls_* settings weren't used for ldaps URIs. - ldap, mysql: Fixed setting connect timeout. - auth: userdb lookups via auth-worker couldn't change username - dsync: Fixed handling deleted directories. Make sure we don't go to infinite mailbox renaming loop. - imap: Fixed crash in NOTIFY when there were watched namespaces that didn't support NOTIFY. - imap: After SETMETADATA was used, various commands (especially FETCH) could have started hanging when their output was large. - stats: Idle sessions weren't refreshed often enough, causing stats process to forget them and log errors about unknown sessions when they were updated later. - stats: Fixed "Duplicate session ID" errors when LMTP delivered to multiple recipients and fts_autoindex=yes. - zlib plugin: Fixed copying causing cache corruption when zlib_save wasn't set, but the source message was compressed. - fts-solr: Fixed escaping Solr query parameters.

6 11

Which mail does a certain attachment belong to?
by Ralf Hildebrandt 08 Dec '15

08 Dec '15

I'm seeing this: ./17/3c/173c81da2aab91f225f8eb5e8fcc87119be4eff0-b38d990442316a54a90c00008f4649af: SecuriteInfo.com.TR.Kryptik.180736.5155.16748.22498.UNOFFICIAL FOUND ./17/3c/173c81da2aab91f225f8eb5e8fcc87119be4eff0-bc70851b03216a54483f00008f4649af: SecuriteInfo.com.TR.Kryptik.180736.5155.16748.22498.UNOFFICIAL FOUND When scanning the "attachments" directory of my mdbox. How can I find out WHICH EMAIL these two attachments belong to? mail_location = mdbox:~/mdbox mail_plugins = zlib fts fts_lucene mdbox_rotate_size = 128M mdbox_rotate_interval = 0 mail_attachment_dir = /some/path/attachments mail_attachment_min_size = 128k mail_attachment_fs = sis posix mail_attachment_hash = %{sha1} -- [*] 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

3 2

Released Pigeonhole v0.4.10.rc2 for Dovecot v2.2.20
by Stephan Bosch 08 Dec '15

08 Dec '15

Hello Dovecot users, Here's the Pigeonhole RC that goes with Dovecot v2.2.20. Due to the earlier described conflict, RC1 will not compile with the new Dovecot release. I am not really in hurry to make a definitive release of Pigeonhole, since v0.4.9 still compiles fine with Dovecot v2.2.20. Still, if no big bugs emerge in the next few days, I will make a final release. So, please test this new RC and especially the new features. Changelog v0.4.10: + Implemented the Sieve mime and foreverypart extensions (RFC 5703). These are fully implemented. The interaction with the editheader extension needs some work, but this should not influence most uses; i.e., changes by the editheader extension are not always visible using foreverypart/mime. + Sieve body extension: Properly implemented the `:text' body transform. It now extracts text for HTML message parts. + Sieve enotify extension: mailto method: Implemented the sieve_notify_mailto_envelope_from setting. This allows configuring the source of the notification sender address for e-mail notifications. This is similar to what already can be configured for redirect. + Added a sieve_enabled (defaults to 'yes') setting that allows explicitly disabling Sieve processing for particular users. This used to be possible by setting `sieve=', but ever since the sieve_before, sieve_after and sieve_default settings were added, this method was not reliable anymore. - variables extension: Fixed handling of empty string by the `:length' set modifier. An empty string yielded an empty string rather than "0". - Fixed memory leak in the Sieve script byte code dumping facility. Extension contexts were never actually freed. - Fixed handling of implicit keep when the last Sieve script is a global one. In that case the implicit keep action was executed in global context, which could mean that trivial (quota) errors ended up in the system log file, rather than the user log file. - doveadm sieve plugin: Fixed crashes caused by incorrect context allocation in the sieve command implementations. The release is available as follows: http://pigeonhole.dovecot.org/releases/2.2/rc/dovecot-2.2-pigeonhole-0.4.10… http://pigeonhole.dovecot.org/releases/2.2/rc/dovecot-2.2-pigeonhole-0.4.10… Refer to http://pigeonhole.dovecot.org and the Dovecot v2.x wiki for more information. Have fun testing this release candidate and don't hesitate to notify me when there are any problems. Regards, -- Stephan Bosch stephan(a)rename-it.nl

1 0

EVP_PKEY_get1_EC_KEY:expecting a ec key
by Oliver Riesen-Mallmann 07 Dec '15

07 Dec '15

Hi, since my last update from the Dovecot Prebuilt Binary for Debian I get a lot of messages like this in mail.log: dovecot: imap-login: Error: SSL: Stacked error: error:0608308E:digital envelope routines:EVP_PKEY_get1_EC_KEY:expecting a ec key Nevertheless Dovecot seems to work normally. Email client doesn't mention any error. This was my last update: Start-Date: 2015-12-04 14:00:31 Commandline: apt-get upgrade Upgrade: dovecot-core:amd64 (2.2.19-1~auto+98, 2.2.20~rc1-1~auto+3), dovecot-managesieved:amd64 (2.2.19-1~auto+98, 2.2.20~rc1-1~auto+3), dovecot-sieve:amd64 (2.2.19-1~auto+98, 2.2.20~rc1-1~auto+3), dovecot-imapd:amd64 (2.2.19-1~auto+98, 2.2.20~rc1-1~auto+3), dovecot-pop3d:amd64 (2.2.19-1~auto+98, 2.2.20~rc1-1~auto+3), openssl:amd64 (1.0.1e-2+deb7u17, 1.0.1e-2+deb7u18), libssl1.0.0:amd64 (1.0.1e-2+deb7u17, 1.0.1e-2+deb7u18) End-Date: 2015-12-04 14:00:53 My current Dovecot version: 2.2.20.rc1 (0b81127e53da) on Linux 3.2.0-4-amd64 #1 SMP Debian 3.2.68-1+deb7u1 x86_64 Is it a bug in Dovecot or in openssl/libssl? Regards Oliver

2 1

Dovecot cluster using GlusterFS
by Filip Pytloun 07 Dec '15

07 Dec '15

Hello, I have recently setup mailserver solution using 2-node master-master setup (mainly based on MySQL M-M replication and GlusterFS with 2 replica volume) on Ubuntu 14.04 (Dovecot 2.2.9). Unfortunately even with shared-storage-aware setting: mail_nfs_index = yes mail_nfs_storage = yes mail_fsync = always mmap_disable = yes ..I have hit strange issues pretty soon especially when user was manipulating same mailbox from multiple devices at the same time. Most issues was about corrupted indexes which was solved easily by just putting them on local storage of each node: mail_location = maildir:/srv/mail/%d/%u:INDEX=/var/lib/dovecot/index/%d/%u But I still hit issues like this one: dovecot: lmtp(6276, user(a)example.com) Error: Broken file /srv/mail/example.com/user(a)example.com/dovecot-uidlist line 8529: UIDs not ordered (8527 >= 8527) Which I am not sure how serious it is or if it's possible to solve or workaround? Anyway because of the above and high possibility of GlusterFS split-brains, I have decided to setup Dovecot Director according to the docs [1] but I have a couple of questions: - is custom monitoring still required? Poolmon [2] is 4 year old so I would suppose there's some progress since that? - it's not possible to have same backends and directors in Dovecot <2.2.17. I can backport newer Dovecot for Ubuntu Trusty, so this is not an issue, but.. - documentation states that it still doesn't work for LMTP [3]? Which is probably important for my setup, because both Postfix servers are using dovecot-lmtp for mail delivery so there can be still some issues (but probably less frequent?) when both servers will deliver new mails for one user at once. So do I really have to split directors from backends? Anyone has experience with clustered Dovecot setup? Why is Dovecot behaving so bad when it pretends to be shared storage friendly? Are these issues only specific for older Dovecot? Or is there something wrong in my architecture design? Thanks for any help, Filip --- [1] http://wiki2.dovecot.org/Director [2] https://github.com/brandond/poolmon/ [3] "LMTP however doesn't currently support mixing recipients to both being proxied and store locally." --- BTW if someone is interested in SaltStack, here are Salt formulas for Dovecot + Postfix + GlusterFS + Roundcube + Mailman setup which we are using: https://github.com/tcpcloud/salt-formula-dovecot https://github.com/tcpcloud/salt-formula-postfix https://github.com/tcpcloud/salt-formula-roundcube https://github.com/tcpcloud/salt-formula-glusterfs

6 7

mmap_disable=yes not honored always
by Peter Benko 06 Dec '15

06 Dec '15

<Sorry for the resend, it seems that the formatting of my original mail got garbled> Hi all, I'm experimenting with having the mail store on a 9p file system that lacks mmap() functionality. So I disabled it in dovecot: mmap_disable = yes However, I keep getting the following error messages in my log: Oct 15 16:55:00 computer-name dovecot: imap user(a)domain.com[192.168.1.3] Error: mmap() failed with file /mail/mailboxes/domain.com/user/indexpvt-accounts/other-user/mailboxes/INBOX/dovecot.index.pvt.log: Invalid argument I only get those messages for the shared folder's private indexes. This setup works perfectly on an ext2/3/4 fs. Any opinions, is this a bug? Is mmap_disable not mandatory for all operations? Best regards, Peter

2 1