August 2022 - dovecot

Dovecot mail-crypt webmail can't read encrypted messages
by Serveria Support 18 Oct '22

18 Oct '22

I'm trying to setup Dovecot with mail-crypt plugin with per-user encryption. I have configured mail-crypt plugin as per official guide here: https://doc.dovecot.org/configuration_manual/mail_crypt_plugin/ After that I created a user and an encrypted key by running this command: doveadm -o \plugin/mail_crypt_private_password=12345 mailbox cryptokey generate -u mail(a)example.org -URf (replacing dummy data ofc) I can log in to webmail (and Dovecot) just fine, emails are getting sent and delivered. I have also checked the storage and the messages seem to be stored encrypted. However, I can't read the emails in webmail (just headers can be seen) and in Dovecot logs I can see the following error: failed: Private key not available: Cannot decrypt key ### Cannot decrypt key ### <8632: Password not available (FETCH RFC822.HEADER) There seems to be an issue with mySQL query. The query I'm using (Select username as "user", password,"%w" as userdb_mail_crypt_private_password from mailbox;) seems to work just fine, when run from mysql prompt it outputs the usernames and passwords, but the error is still there (Cannot decrypt key ### Password not available). Any ideas? What am I missing?

16 70

Multidomain ssl config ?
by Maurizio Caloro 16 Oct '22

16 Oct '22

on postfix now this seems to run, and with dovecot i need also handle this two domains, but appairing this error messages. like: Jun 29 20:49:28 Dovecot/imap-login: Info: Disconnected (no auth attempts in 0 secs): user=<>, rip=a.b.c.d, lip=37.120.190.188, TLS handshaking: SSL_accept() failed: error:14094416:SSL routines: ssl3_read_bytes:sslv3 alert certificate unknown: SSL alert number 46, session=<FdklDjkdfrkfi> Running with Debian Buster # dovecot --version 2.3.4.1 (f79e8e7e4) # nmail.caloro.ch local_name nmail.caloro.ch { ssl_cert = </etc/letsencrypt/live/nmail.caloro.ch/privkey.pem ssl_key = </etc/letsencrypt/live/nmail.caloro.ch/fullchain.pem } # nmail.calm-ness.ch local_name nmail.calm-ness.ch { ssl_cert = </etc/letsencrypt/live/nmail.calm-ness.ch/privkey.pem ssl_key = </etc/letsencrypt/live/nmail.calm-ness.ch/fullchain.pem } thanks for possible help

6 5

Duplicate UID warning in dsync backups
by Doug 23 Sep '22

23 Sep '22

Hi, Running dovecot version 2.3.18 with mdbox storage. I run daily backups using dsync on the same server into maildir format. Several times per week I will get an error like the one below that prevents backups from completing. The error is always on my user, never any of my other couple dozen users. dsync(doug): Warning: Deleting mailbox 'INBOX': UID=133060 already exists locally for a different mail: GUIDs don't match (1657786027.M158587P19048.maildev.domain.com vs 581a08178e8ecf62f04700002bad4ea1) dsync(doug): Error: Couldn't delete mailbox INBOX: INBOX can't be deleted. Subsequent backups typically have a slightly different error. dsync(doug): Warning: Deleting mailbox 'INBOX': UID=132994 GUID=407c871b0ec1c562b51000002bad4ea1 is missing locally dsync(doug): Error: Couldn't delete mailbox INBOX: INBOX can't be deleted. Notice the UID and GUID are different from the first error. At this point I will identify 407c871b0ec1c562b51000002bad4ea1 in the source mailbox, and move it to another folder. More often than not that clears things up. The backup command I am using is dsync -u doug all backup maildir:/home/doug/.mailbkup/mailboxes My questions are: Is the "locally" referring to the source or the target of the backup? What might be causing this problem? What the heck is this guid: 1657786027.M158587P19048.maildev.domain.com" Note: The server is actually named mail.domain.com, not maildev.domain.com All file names in the backup maildir have mail.domain.com in them No files exist in the target maildir directory that begin with 1657786027* My original migration from mbox format to mdbox was tested and run on a server named maildev.domain.com. Is that why I am seeing a reference to maildev here? Is that name cached somewhere? Can it be corrected? Does it matter? Are there any options I should add to the dsync command to fix this problem? -- Doug

1 1

Panic: file mail-index-transaction-finish.c: line 185
by tobisworld＠gmail.com 19 Sep '22

19 Sep '22

Hi list I have weird issue with my Dovecot 2.3.17.1 (476cd46418) When deleting a certain amount of messages from my INBOX via my MUA (Evolution) all over sudden dovecot starts to panic Panic: file mail-index-transaction-finish.c: line 185 (mail_index_transaction_get_uid): assertion failed: (seq <= t->view- >map->hdr.messages_count) imap(REDACTED)<24075><P0/vLkXUNt4KQmMB>: Error: Raw backtrace: /usr/lib64/dovecot/libdovecot.so.0(backtrace_append+0x42) [0x7f09274d4142] -> /usr/lib64/dovecot/libdovecot.so.0(backtrace_get+0x1e) [0x7f09274d424e] -> /usr/lib64/dovecot/libdovecot.so.0(+0xf72fe) [0x7f09274e22fe] -> /usr/lib64/dovecot/libdovecot.so.0(+0xf73a1) [0x7f09274e23a1] -> /usr/lib64/dovecot/libdovecot.so.0(i_fatal+0) [0x7f0927430e38] -> /usr/lib64/dovecot/libdovecot-storage.so.0(+0xf438d) [0x7f092789d38d] - > /usr/lib64/dovecot/libdovecot-storage.so.0(+0xf43da) [0x7f092789d3da] -> /usr/lib64/dovecot/libdovecot- storage.so.0(mail_index_transaction_finish+0x177) [0x7f092789da87] -> /usr/lib64/dovecot/libdovecot-storage.so.0(+0xf9ad6) [0x7f09278a2ad6] - > /usr/lib64/dovecot/libdovecot- storage.so.0(mail_index_transaction_commit_full+0x9d) [0x7f09278a2fad] -> /usr/lib64/dovecot/libdovecot- storage.so.0(mail_index_transaction_commit+0x1c) [0x7f09278a306c] -> /usr/lib64/dovecot/libdovecot- storage.so.0(mail_cache_set_seq_corrupted_reason+0x61) [0x7f092788c921] -> /usr/lib64/dovecot/libdovecot- storage.so.0(mail_set_mail_cache_corrupted+0xed) [0x7f09277fbbbd] -> /usr/lib64/dovecot/libdovecot-storage.so.0(+0x7a3e9) [0x7f09278233e9] - > /usr/lib64/dovecot/libdovecot-storage.so.0(+0x7a6b0) [0x7f09278236b0] -> /usr/lib64/dovecot/libdovecot- storage.so.0(mail_get_physical_size+0x38) [0x7f09277e8ba8] -> /usr/libexec/dovecot//lib20_mail_log_plugin.so(+0x2465) [0x7f09263dd465] -> /usr/libexec/dovecot//lib15_notify_plugin.so(notify_contexts_mail_copy+ 0x5c) [0x7f09265e13ec] -> /usr/libexec/dovecot//lib15_notify_plugin.so(+0x25d4) [0x7f09265e25d4] -> /usr/libexec/dovecot//lib10_quota_plugin.so(+0x102ae) [0x7f09269f92ae] -> /usr/libexec/dovecot//lib01_acl_plugin.so(+0xef3e) [0x7f0926c10f3e] -> /usr/lib64/dovecot/libdovecot- storage.so.0(+0x50c4c) [0x7f09277f9c4c] -> /usr/lib64/dovecot/libdovecot-storage.so.0(mailbox_move+0x58) [0x7f09277f9ec8] -> dovecot/imap(+0x1175b) [0x55a2c0ea575b] -> dovecot/imap(command_exec+0x64) [0x55a2c0eb3f04] -> dovecot/imap(+0x1dde2) [0x55a2c0eb1de2] -> dovecot/imap(+0x1de71) [0x55a2c0eb1e71] -> dovecot/imap(+0x1e039) [0x55a2c0eb2039] -> dovecot/imap(client_handle_input+0x205) [0x55a2c0eb22f5] -> dovecot/imap(client_input+0x79) [0x55a2c0eb28d9] imap(REDACTED)<24075><P0/vLkXUNt4KQmMB>: Fatal: master: service(imap): child 24075 killed with signal 6 (core dumps disabled - https://dovecot.org/bugreport.html#coredumps) >From that point on ANY access to INBOX from my MUA causes the error. The only way that I found to get around is to unsubscribe the INBOX folder and subscribe again. Then the error is gone until I delete a bunch of messages again. The error only occurs if I delete the messages within short time or select multiple (about 20msg at once). If I delete them one-by-one always waiting for expunge message in dovecot logs to appear then I can delete as many as I want. I have not verified with other MUA but in Evolution I can reproduce the issue. Cheers and have a good one tobi

4 14

Unclear how to get both user and domain quotas to work
by Hans van Zijst 10 Sep '22

10 Sep '22

Hi, First of all: my apologies for this long message. I'm trying to post all the information anybody could need to help me on my way. What I'm trying to make is an installation where both users and domains can have quotas. Ideally I would configure a quota for a domain and set a quota of 0 for all users within that domain, so that mail will be accepted up to the point where the domain quota is reached. Setting up user quota was fairly easy, there's lots of documentation around and examples to take inspiration from. Not so much for domain quota however, or I'm really bad at searching the Net. I'm running Debian Bullseye here: MariaDB 10.5.12 Dovecot 2.3.13 Postfixadmin 3.3.10 I want to use dictionaries in MySQL, because eventually this will be a platform with several Dovecot and Postfix servers. At this time I'm only running one Dovecot installation and a few Postfix servers (two for communication with the outside world and one to do the virus and spam scanning). My current setup for user quota is how Postfixadmin wants it: MariaDB [postfixadmin]> desc quota2; +----------+--------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +----------+--------------+------+-----+---------+-------+ | username | varchar(100) | NO | PRI | NULL | | | bytes | bigint(20) | NO | | 0 | | | messages | int(11) | NO | | 0 | | +----------+--------------+------+-----+---------+-------+ 3 rows in set (0.001 sec) In /etc/dovecot/conf.d/90-quota.conf I have this: plugin { quota = dict:User quota::proxy::sqlquota quota_rule = *:bytes=1024M quota_rule2 = Trash:bytes=+100M quota_vsizes = yes quota_status_success = DUNNO quota_status_nouser = DUNNO quota_status_overquota = "552 5.2.2 Mailbox is full" } dict { sqlquota = mysql:/etc/dovecot/dovecot-dict-sql.conf.ext } The SQL-configuration /etc/dovecot/dovecot-dict-sql.conf.ext is this: connect = host=... dbname=... user=... password=... map { pattern = priv/quota/storage table = quota2 username_field = username value_field = bytes } map { pattern = priv/quota/messages table = quota2 username_field = username value_field = messages } map { pattern = shared/expire/$user/$mailbox table = expires value_field = expire_stamp fields { username = $user mailbox = $mailbox } } I use LMTP to accept mail from Postfix; the quota-plugin is activated for LMTP, so every incoming mail should be counted. This works like a charm: if I don't configure a quota for a user, he gets the default 1GB, if I override it in Postfixadmin, then that's what he gets. I'm running quota-status on a TCP-port, so the gateway SMTP-servers can connect to Dovecot for a quotacheck to decide whether to accept or reject an incoming message. Again, this works flawlessly. I can't find any serious documentation about this mapping thing. In particular, what's with that "pattern"? All I can find in the documentation is: "The dict key must match exactly priv/quota/storage. The dict keys are hardcoded in the Dovecot code, so depending on what functionality you’re configuring you need to know the available dict keys used it." -- https://doc.dovecot.org/configuration_manual/dict/#dict-sql If anyone can point me at better documentation than the actual source code: please! I browsed through the quota-plugin source code, but I'm not a programmer and couldn't find what I thought I was looking for. That said, I'm not sure where I got the third "map" from, with the shared/expire pattern. I seem to have forgotten to document that bit of my journey. Now, domain quotas... I've tried all kinds of stuff, but at the moment I'm stuck with something that seems to almost work. I added a second quota root: plugin { quota2 = dict:Domain quota:%d:proxy::sqldomainquota quota2_rule = *:bytes=1024M quota2_status_success = DUNNO quota2_status_nouser = DUNNO quota2_status_overquota = "552 5.2.2 Domain quota exceeded" quota2_vsizes = yes } dict { sqldomainquota = mysql:/etc/dovecot/dovecot-dict-sql.conf.ext } This dict refers to the same mapping as the one for user quotas, and I think that's the way it should be. Please correct me if I'm wrong. This creates a new user in the quota2 table, by the name of just the domain. Unfortunately, it simply fills it with the numbers from the first user it can find in that domain, not all users in that domain combined. What I think I should do, is change something in the dict so that it sums up all users in that domain, something like this: SELECT SUM(bytes) FROM quota2 WHERE username LIKE '%@domain'; Problem is that I can't find anything useful online with regards to this, I'm not even sure if I'm thinking in the right direction here. Any hints, suggestions or corrections? Or pointers to articles or documentation I might have missed? Kind regards, Hans

2 1

Permission denied UNIX perms appear ok (ACL/MAC wrong?))
by Austin Witmer 07 Sep '22

07 Sep '22

Hello all! Recently I upgraded my mail server to Ubuntu 22.04 LTS and ever since then I am periodically getting some dovecot errors like the below in my mail log. As far as I can tell, my unix perms are just fine. What is ACL/MAC? Aug 20 14:41:58 mail dovecot: imap(user(a)domain.com)<56316><1NieGKPmuOdKwxVI>: Error: Mailbox INBOX: stat(/mnt/volume1/mailserver/plain/maildir/domain.com/user/dovecot.index.lo… failed: Permission denied (euid=1000(austin) egid=1000(austin) UNIX perms appear ok (ACL/MAC wrong?)) And here is the listing showing the permissions for that file. austin@mail:~$ ls -la /mnt/volume1/mailserver/plain/maildir/domain.com/user/dovecot.index.log -rwxrwxr-- 1 austin austin 15796 Aug 20 14:41 /mnt/volume1/mailserver/plain/maildir/domain.com/user/dovecot.index.log What in the world is causing these errors, and what can I do about them? Thanks in advance! Austin Witmer

7 20

Call procmail after sieve
by George Asenov 01 Sep '22

01 Sep '22

Hello, I'm trying to achieve the following scenario with dovecot + sieve + procmail. Dovecot is set as LDA in postfix (if this makes any difference) I need to set sieve that after it has executed all it's scripts (including before, ~/.sieve etc) and if keep is in still in effect to call procmail to also execute the custom user filters and then return (if it is not forwarded elsewhere or some) to dovecot for normal delivery. If you ask this I need this for backward compatibility. I now use procmail for local delivery,spam/virus filtering, and users have set a lot of custom filters/redirects. mailbox_command = /usr/bin/procmail-wrapper -o -a $DOMAIN -d $LOGNAME I want to migrate to dovecot as lda and sieve for filtering but to keep procmail working for a while. I have tried to follow (partially) this guide https://www.enricozini.org/blog/2022/debian/migrating-from-procmail-to-siev… but procmail fails to execute ########################## Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail: Establishing connection Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail: Forked child process Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Connected to program Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Finished streaming payload to program Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Finished input to program Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Disconnected Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Waiting for program to finish after 0 msecs (timeout = 10000 msecs) Aug 30 15:22:26 uk7 dovecot[112153]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Fatal: execvp(/var/lib/dovecot/sieve-pipe/procmail) failed: Permission denied Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Child process ended Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Terminated with non-zero exit code 75 Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Debug: program exec:/var/lib/dovecot/sieve-pipe/procmail (112153): Destroy Aug 30 15:22:26 uk7 dovecot[112151]: lda(test-uk7.teststring567.tk)<112151><BlNhECIdDmMXtgEAsVjU+A>: Error: sieve: failed to execute to program `procmail': refer to server log for more information. [2022-08-30 15:22:26] ############################### i couldn't find the logs referred here some relative (on my opinion) configs: ################################ #/etc/dovecot/conf.d/90-sieve.conf plugin { sieve = file:~/sieve;active=~/.dovecot.sieve sieve_plugins = sieve_extprograms sieve_extensions = +vnd.dovecot.pipe sieve_pipe_bin_dir = /var/lib/dovecot/sieve-pipe sieve_after = /var/lib/dovecot/sieve_after sieve_trace_dir = ~/.sieve-trace sieve_trace_level = matching sieve_trace_debug = yes } #################################### #/var/lib/dovecot/sieve-pipe/procmail #!/bin/sh exec /usr/bin/procmail ##################################### #/var/lib/dovecot/sieve_after/procmail.sieve require "vnd.dovecot.pipe"; pipe "procmail"; ##################################### #cat /etc/procmailrc SHELL="/bin/bash" SENDMAIL="/usr/sbin/sendmail" SENDMAILFLAGS="-oi -t"" LOGFILE="/var/log/procmail.log" # fully define here so user .procmailrc can use it too DELIVER="/usr/libexec/dovecot/deliver -d $LOGNAME" # fallback: DEFAULT="$HOME/Maildir/" MAILDIR="$HOME/Maildir/" # deliver spam to spam folder :0 w * ^X-Spam-Status: Yes | $DELIVER -m spam # uncomment next line to support user .procmailrc files INCLUDERC=$HOME/.procmailrc # deliver to INBOX and stop :0 w | $DELIVER ######################################### Thank you in advance

5 8

where to setup virtual_mailbox_maps
by frakass 31 Aug '22

31 Aug '22

Hello, I have a domain in virtual_mailbox_domains: aaa.com I have also the virtual_alias_domains which include: bbb.com I know how to forward x(a)bbb.com to y(a)aaa.com by setting up the file "virtual_alias_maps": x(a)bbb.com y(a)aaa.com (and run postmap after the changes.) But, how can I setup virtual_mailbox_maps (if I name this correctly)? for example, 111(a)aaa.com forwards to 222(a)aaa.com. Thank you.

2 1

multiple / backup sql servers for sql server access
by Jack Snodgrass 31 Aug '22

31 Aug '22

I am using this file: dovecot-sql.conf.ext and in there I have a connect = "host=<dbserver> dbname=<dbname> user=<dbuser> password=<dbpass>" My <dbserver> was down and I lost mail access for a few days before I realized that there was an issue. How can I specify a backup server so that if my primary sql server goes down, a backup sql server will be queried? -- jack

3 4

Redhat 9 Repository for Dovecot
by Istiak Ferdous 29 Aug '22

29 Aug '22

Hello, Redhat 9 is publicly released for some time. Is there any plan to provide repository for it? http://repo.dovecot.org/ce-2.3-latest/rhel/ https://developers.redhat.com/articles/2022/05/18/whats-new-red-hat-enterpr… Thanks, Istiak Ferdous

2 4