- dovecot - dovecot.org

2.4 ballpark release date?
by Shawn Heisey 05 Apr '24

05 Apr '24

I'm wondering whether there is a ballpark release date for Dovecot 2.4. I know an exact release date is probably not available, but can it at least be narrowed to a specific quarter? Is it likely in 2023? I have seen on previous messages that an Ubuntu repository for 22.04 (jammy) will appear with the 2.4 release. My mail server in AWS is running Ubuntu 20.04 (focal), which I have avoided upgrading for various reasons, including websites that don't work with PHP 8.1. Now the last remaining hurdle is the fact that there is no dovecot repository for jammy. I would prefer the official Dovecot repo over dovecot packages included with Ubuntu, because it's likely to be a newer release, and built from vanilla project code. At the moment jammy includes dovecot 2.3.16, which is even older than the 2.3.20 that I am running on focal from the official dovecot repo. Again, I know that a precise release date won't be possible. I'm just hoping for something like "soon" or "in about six months." Thanks, Shawn

3 5

Re: Fwd: Dovecot, Load Balancing and SSL
by Joseph Tam 03 Apr '24

03 Apr '24

leon(a)relay.gb.net writes: > I wonder if someone would provide me with some advice. I've been setting > up a couple of Postfix servers just for fun. I've got two Postfix > servers m1.domain.com and m2.domain.com. I can send and receive mail via > both of them. Ive also got Postfixadmin and RoundCube on them and I'm > replicating the database over both servers. > > I introduced a load balancer. Postfixadmin and Roundcube work perfectly. > However. When I send mail from Thunderbird. M1 reports that the > certificate does not match. It's expecting a certificate for > mail.domain.com. Complaints when you're sending mail? This is not Dovecot's problem, but Postfix's. If you're setting your SMTP outgoing to M1, then the certificate M1 uses have M1 as a subject. If you're setting the outgoing mail server as "mail.domain.com" load balanced to M1/M2, then both your Postifx servers need to use the same certificate with "mail.domain.com" as a subject. Simple as that. Joseph Tam <jtam.home(a)gmail.com>

1 0

Dovecot, Load Balancing and SSL
by leon＠relay.gb.net 03 Apr '24

03 Apr '24

Hi Everyone! I wonder if someone would provide me with some advice. I've been setting up a couple of Postfix servers just for fun. I've got two Postfix servers m1.domain.com and m2.domain.com. I can send and receive mail via both of them. Ive also got Postfixadmin and RoundCube on them. I introduced a load balancer. Postfixadmin and Roundcube work perfectly. However. When I send mail from Thunderbird. M1 reports that the certificate does not match. It's expecting a certificate for mail.domain.com. On M1 I thought I would try to specify the location of the SSL when using mail.domain.com. I placed the certificate in the NAS and specified it in the 10-ssl file. I did this as per the docs. mail.domain.com local_name mail.domain.com { ssl_cert = </nas/ssl/fullchain.pem ssl_key = </nas/ssl/privkey.pem } It doesn't work. And I'm not entirely sure why M2 works and M1 doesn't. As far as I can see I haven't specified or placed a certificate for mail.domain.com on M2 and searching my bash history doesn't produce anything related to mail.domain.com. So how can I load balance email connections over my two servers without getting a damn certificate warning? Any help would be much appreciated. Thanks.

2 3

Any way to make a shared mailbox(not a shared folder)
by Maksim Rodin 02 Apr '24

02 Apr '24

Hello I wonder if there is a right way to make a shared mailbox? I do not mean "shared folder" but a whole mailbox. E.g. I have a mailbox info(a)company.com and I have a user mailbox user1(a)company.com. I would like that a user which can already authenticate as user1(a)company.com can setup another account in his Thunderbird as, say, user1@company.com\info@company.com or something like user1*info, enters his own password and can use the mailbox info(a)company.com as his second mailbox. It might be something similar to master user feature but I do not want the user1(a)company.com to have access to all the mailboxes on the dovecot imap server. -- Regards Maksim Rodin

4 6

Implications of LMTP socket being world readable and writable (0666) by default
by da-dovecotlist-15＠abelonline.de 31 Mar '24

31 Mar '24

I was surprised to find that the LMTP socket has permission mode 0666 by default and since configs are merged with defaults, there is no way to disable this AFAICS. # doveconf -d ... service lmtp { unix_listener lmtp { group = mode = 0666 user = } ... } Is this also how it is supposed to be used in production? I understand that LMTP is just for delivering new mails but is there really no need to restrict this further? To me it seems reasonable, to force all services on this machine to go through Postfix and not be able to just put e-mails in the users mailboxes via LMTP. Am I missing something? My use case is a "small" NixOS mail server with Postfix, Nextcloud a.o. installed on the same machine. No chroot jails because of NixOS.

3 4

Re: [Dovecot] [dovecot-2.1.15] mdbox corruption, doveadm force-resync can't repair it (throws segfault)
by Manish Prajapati 31 Mar '24

31 Mar '24

3 2

Separate index get dovecot lost
by Joan Moreau 30 Mar '24

30 Mar '24

Hi I have large number of email (~TB) and want to put the index in a separate, rapid drive Initially, I have mail_location = mdbox:/files/mail/%d/%n If I put mail_location = mdbox:/files/mail/%d/%n:INDEX=/data/mailindexes/%d/%n then dovecot gets totally lost and tries to reach mailboxes content and tree from the INDEX location instead of the original location What is wrong ? Thank you Hi I have large number of email (~TB) and want to put the index in a separate, rapid drive Initially, I have mail_location = mdbox:/files/mail/%d/%n If I put mail_location = mdbox:/files/mail/%d/%n:INDEX=/data/mailindexes/%d/%n then dovecot gets totally lost and tries to reach mailboxes content and tree from the INDEX location instead of the original location What is wrong ? Thank you

2 4

replication timeout
by Andy Balholm 28 Mar '24

28 Mar '24

I have two Dovecot mail servers that replicate to each other. Sometimes there are delays in the synchronization, and I notice that the mail log has entries like this: Error: dsync(spokane): I/O has stalled, no activity for 600 seconds (last sent=mailbox, last recv=mailbox_state) Five minutes seems like a long time to sit there waiting with nothing happening. Is there a way to reduce this timeout so that I don't have so many replicaton connections just sitting around doing nothing? (Of course, a way to prevent the I/O stalls would be great too, but with my limited upload bandwidth, they may be unavoidable.) Andy

2 9

panics
by mgrant＠grant.org 28 Mar '24

28 Mar '24

I am still getting panics. Today I'm seeing this panic only when 'eM Client' connects. Not other imap clients. It's almost as if there's a malformed message in my mailbox which might be causing this. I spent some time trying to do things like delete the last 20 messages or so but no luck. I tried to enable debugging in 10-logging but i get nothing at all in the log file /var/log/dovecot. Yes I've restarted dovecot. I'm running the version from here: https://repo.dovecot.org/ce-2.3-latest/debian/bullseye Any ideas?? Mar 26 19:23:14 strange dovecot: imap(mgrant)<2743779><EQ/3kpgUgvxRaQEP>: Panic: file istream-header-filter.c: line 663 (i_stream_header_filter_snapshot_free): assertion failed: (snapshot->mstream->snapshot_pending) Mar 26 19:23:14 strange dovecot: imap(mgrant)<2743779><EQ/3kpgUgvxRaQEP>: Error: Raw backtrace: /usr/lib/dovecot/libdovecot.so.0(backtrace_append+0x42) [0x7f9d301ed482] -> /usr/lib/dovecot/libdovecot.so.0(backtrace_get+0x1e) [0x7f9d301ed59e] -> /usr/lib/dovecot/libdovecot.so.0(+0x1041fb) [0x7f9d301fa1fb] -> /usr/lib/dovecot/libdovecot.so.0(+0x104291) [0x7f9d301fa291] -> /usr/lib/dovecot/libdovecot.so.0(+0x56670) [0x7f9d3014c670] -> /usr/lib/dovecot/libdovecot.so.0(+0x51ec5) [0x7f9d30147ec5] -> /usr/lib/dovecot/libdovecot.so.0(i_stream_snapshot_free+0x18) [0x7f9d30205c68] -> /usr/lib/dovecot/libdovecot.so.0(i_stream_unref+0x4c) [0x7f9d30205d0c] -> /usr/lib/dovecot/libdovecot.so.0(message_parse_header_deinit+0x15) [0x7f9d301d60e5] -> /usr/lib/dovecot/libdovecot.so.0(+0xe29b1) [0x7f9d301d89b1] -> /usr/lib/dovecot/libdovecot.so.0(message_parser_parse_next_block+0x45) [0x7f9d301d84d5] -> /usr/lib/dovecot/libdovecot.so.0(message_parser_parse_header+0x51) [0x7f9d301d8611] -> /usr/lib/dovecot/libdovecot-storage.so.0(index_mail_parse_headers_internal+0x11a) [0x7f9d3039c8ca] -> /usr/lib/dovecot/libdovecot-storage.so.0(index_mail_parse_headers+0x4c) [0x7f9d3039c9ec] -> /usr/lib/dovecot/libdovecot-storage.so.0(+0xd885a) [0x7f9d303a185a] -> /usr/lib/dovecot/libdovecot-storage.so.0(+0xd8a89) [0x7f9d303a1a89] -> /usr/lib/dovecot/libdovecot-storage.so.0(index_mail_get_special+0x122) [0x7f9d303a1d92] -> /usr/lib/dovecot/libdovecot-storage.so.0(mail_get_special+0xa) [0x7f9d3031cfba] -> dovecot/imap(+0x26808) [0x55f5622cd808] -> dovecot/imap(+0x240a4) [0x55f5622cb0a4] -> dovecot/imap(imap_fetch_more+0x39) [0x55f5622cc4a9] -> dovecot/imap(cmd_fetch+0x39f) [0x55f5622bc8af] -> dovecot/imap(command_exec+0xa4) [0x55f5622c95b4] -> dovecot/imap(+0x2052f) [0x55f5622c752f] -> dovecot/imap(+0x205da) [0x55f5622c75da] -> dovecot/imap(+0x208ed) [0x55f5622c78ed] -> dovecot/imap(client_handle_input+0x1b5) [0x55f5622c7ab5] -> dovecot/imap(client_input+0x70) [0x55f5622c8020] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_call_io+0x69) [0x7f9d30210429] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x132) [0x7f9d30211b12] Mar 26 19:23:14 strange dovecot: imap(mgrant)<2743779><EQ/3kpgUgvxRaQEP>: Fatal: master: service(imap): child 2743779 killed with signal 6 (core dumps disabled - https://dovecot.org/bugreport.html#coredumps)

4 5

SAVEDATE is using ctime instead of btime for Maildir
by Alex 28 Mar '24

28 Mar '24

Dovecot supports SAVEDATE IMAP extension [1]. This is however not implemented correctly for Maildir storage. Compared to "internal date", "save date" is supposed to be the same but well-defined for APPEND and preserved when COPY command is used. However, for Maildir storage INTERNALDATE is implemented using mtime (modification timestamp) and SAVEDATE is implemented using ctime (status change timestamp) [2]. ctime is changed even more often than mtime. For example, chown or chmod operations on the message change its ctime, but preserve its mtime. Dovecot should instead use btime (creation timestamp) when it is available and not implement SAVEDATE at all if there is no btime. Just one example of the message in Maildir: stat 1708031310.M769290P3712350.nine,S=7898,W=8002:2,Sa File: 1708031310.M769290P3712350.nine,S=7898,W=8002:2,Sa Size: 7898 Blocks: 16 IO Block: 4096 regular file Device: 9,2 Inode: 7864351 Links: 1 Access: (0600/-rw-------) Uid: ( 999/ vmail) Gid: ( 996/ vmail) Access: 2024-02-16 08:34:04.374097798 +0100 Modify: 2024-02-15 22:08:30.769813368 +0100 Change: 2024-02-21 13:39:14.878163244 +0100 Birth: 2024-02-15 22:08:30.769813368 +0100 Its Birth date (btime) and Modify date (mtime) are the same, but Change date (ctime) was changed, likely by someone doing chown/chmod manually. You can read more in the issue [3] where we have problems with incorrect INTERNALDATE and investigated the possibility of using SAVEDATE, but found that it is not implemented correctly and is changed even more often than INTERNALDATE. According to the standard it should be the other way round, SAVEDATE should be more stable and survive COPY. [1] <https://www.rfc-editor.org/rfc/rfc8514.html> [2] <https://github.com/dovecot/core/blob/93eb0663336b17f66a67324de43e64bf0bc796…> [3] <https://github.com/deltachat/deltachat-core-rust/issues/5339>

2 1