December 2022 - dovecot

Happy Holidays From SCOM.CA Internet Services Inc.
by Paul Kudla 23 Dec '22

23 Dec '22

1 0

Pigeonhole v0.5.20 released
by Aki Tuomi 22 Dec '22

22 Dec '22

We are pleased to release v0.5.20 of Pigeonhole. https://pigeonhole.dovecot.org/releases/2.3/dovecot-2.3-pigeonhole-0.5.20.t… https://pigeonhole.dovecot.org/releases/2.3/dovecot-2.3-pigeonhole-0.5.20.t… Binary packages in https://repo.dovecot.org/ Docker images in https://hub.docker.com/r/dovecot/dovecot Regards, Aki Tuomi Open-Xchange oy -- * No changes - release done to keep version numbers synced.

1 0

OT: Two simultaneous POP3 connections from Outlook
by kjohnson＠eclypse.org 21 Dec '22

21 Dec '22

Hello, I recognize this may be somewhat off-topic, but my search-fu has failed to provide useful results, so I ask for pointers to further information. Today, while examining Dovecot logs, I found a case where the Outlook mail client (probably 2016 or later) on a user's computer opened two simultaneous, or at least overlapping, POP3 connections to Dovecot. This did not go well, naturally. Warning: Transaction log file .../mail/.imap/INBOX/dovecot.index.log was locked for 156 seconds (rotating while syncing) If anyone else has found useful information about this peculiar behavior for an email client, I would be pleased to receive a pointer in that direction. Thank you. Ken

2 1

How is URLAUTH broken, and for how long?
by Andy Balholm 21 Dec '22

21 Dec '22

The documentation for imap_urlauth_host says: > Warning > > URLAUTH in current versions of Dovecot is broken in several ways. > This will be fixed in the future, but activating URLAUTH support > on production systems is not recommended. In what ways is it broken? Is anyone actually working on fixing it? Andy

1 0

Re: SPF DNS Error / R_SPF_DNSFAIL
by Jochen Bern 20 Dec '22

20 Dec '22

On 19.12.22 13:20, Carlos Mogas da Silva wrote: > On Mon, 2022-12-19 at 12:38 +0100, Benny Pedersen wrote: >> Carlos Mogas da Silva skrev den 2022-12-19 11:57: >>> What can cause this error and/or how can I debug it? >> >> try avoid to many dns querys > > Even on a caching dns server? Any ideas on why it is failing? I don't remember it being in an actual RFC, but SPF records are supposed to not require more than 10 DNS queries to fully resolve. https://dmarcian.com/spf-survey/ gives cttexpresso.pt an all-green result, but notes that the "+a:email.reditus.pt" part is inoperational ("no A or AAAA RRs found", in fact, I'm getting an NXDOMAIN for the FQDN). Regards, -- Jochen Bern Systemingenieur Binect GmbH

1 0

SPF DNS Error / R_SPF_DNSFAIL
by Carlos Mogas da Silva 20 Dec '22

20 Dec '22

Hi list. My incoming messages are getting R_SPF_DNSFAIL and an error on the Authentication-Results header: spf=temperror (mail01.r3pek.org: error in processing during lookup of no-reply(a)cttexpresso.pt: DNS error) I see the DNS request on my DNS server and a "working" reply going out. What can cause this error and/or how can I debug it?

4 6

mail filters
by mick.crane 20 Dec '22

20 Dec '22

I'm using what ever is in Debian Bookworm, I'm pretty sure. These filters? I collect from different email addresses like gmail and that. If filter for the "Sent to" email address, move to this directory If "From" contains "blah", move to this other directory. If something is Sent to and also From contains "blah" then I get 2 copies of the message in both filter directories. How can I fix this? regards mick

5 6

feature request: add base64 variable modifier
by alex 19 Dec '22

19 Dec '22

Hi, It would be very convenient to have a base64 variable modifier. Main use-case is to base64-encode passwords when changing your hashing scheme. With a base64 encoding modifier (e.g. %B) something like this would safely work: userdb { driver = prefetch default_fields = plain_pass=%Bw } without things breaking when there are %-characters in the password, and without having to send the plaintext password to the database, as current solutions propose, which is never a good idea. alternative solution: adapt the generic hash-modifier to accept a none/plaintext hashing-scheme and use the base64 output functionality that is already available there. Thanks! Alex

1 0

What do I do with the ".ext?"
by James Moe 18 Dec '22

18 Dec '22

dovecot 2.3.15 opensuse LEAP 15.4 I am installing dovecot and have postfix as the SMTP agent. It also needs to use a MySQL database. In </etc/dovecot> there is the main dovecot.conf file, and a number of other files with the extension ".conf.ext." It is not clear how to use these. I could nothing in the documentation that discusses it. Do I copy one, say dovecot-sql.conf.ext, modify it: - rename the copy without the ".ext?" - or move the copy to the "conf.d" directory, with or without the "ext?" -- James Moe moe dot james at sohnen-moe dot com 520.743.3936 Think.

2 2

2.3.14: stale imap-login processes due to bogus user setup tying up process limit
by Volker Stolz 16 Dec '22

16 Dec '22

I just dealt with an interesting Dovecot issue on our small server, and thought I might share it. I also feel that although in the end it was a configuration error, Dovecot should somehow avoid leaving this processes around and hence filling up the process limit quickly. - set up Dovecot IMAP on Debian buster (either with the older 2.3.x Debian package, or 2.3.14 from the dovecot-repo) - config: ... namespace inbox { inbox = yes location = ... default settings ... ... passdb { args = scheme=HMAC-MD5 username_format=%u /etc/dovecot/users driver = passwd-file } protocols = " imap" ... userdb { args = blocking=no driver = passwd } ... - secret sauce: have a user in the passdb that doesn't have a physical account on the system yet (we were migrating to a new machine). The log-file shows successful authentication for that user (after we flipped DNS to point to this new machine), but over time (2-3 times/day) we were left with an increasing number of `imap-login` processes in state `pre-login`, eventually requiring a dovecot restart. Dialling up process_limits didn't seem to help, we'd just get more of these stale processes. After creating the user on the system, the behaviour reverted to normal, no more stale processes. I do not know what error message that user received on their end, but since we announced we were migrating, they probably just ignored it and thought it as part of the transition. Best, Volker (a happy & long Dovecot user)

3 4