December 2022 - dovecot

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

Re: Can't figure out why managesieve (pigeonhole) can't connect
by colin＠colinlikesfood.com 15 Dec '22

15 Dec '22

thank you again. it seems you have seen my paste of config.inc.php. I do not have a config.php: my_user@some_host:/usr/local/www/roundcube/config # ls -l total 67 -rw-r--r-- 1 root wheel 164 Jul 23 15:17 .htaccess -rw-r--r-- 1 root wheel 1867 Nov 22 15:12 config.inc.php -rw-r--r-- 1 root wheel 2943 Jul 23 15:17 config.inc.php.sample -rw-r--r-- 1 root wheel 63790 Oct 29 20:24 defaults.inc.php -rw-r--r-- 1 root wheel 2806 Jul 23 15:17 mimetypes.php my_user@some_host:/usr/local/www/roundcube/config # I have tried changing tls:// to ssl:// and back again (in the line $config['managesieve_host'] = 'tls://obfuscated.domain';) but the error remains the same: roundcube: PHP Error: Connection refused (GET /index.php?_task=settings&_action=plugin.managesieve) roundcube: PHP Error: Unable to connect to managesieve on obfuscated.domain:4190 in /usr/local/www/roundcube/plugins/managesieve/lib/Roundcube/rcube_sieve_engine.php on line 221 (GET /index.php?_task=settings&_action=plugin.managesieve) roundcube: PHP Error: Not currently in AUTHORISATION state (GET /index.php?_task=settings&_action=plugin.managesieve) php: PHP Error: Not currently connected (GET /index.php?_task=settings&_action=plugin.managesieve) roundcube: PHP Error: Connection refused (GET /index.php?_task=settings&_action=plugin.managesieve-action&_framed=1&_nav=hide) roundcube: PHP Error: Unable to connect to managesieve on obfuscated.domain:4190 in /usr/local/www/roundcube/plugins/managesieve/lib/Roundcube/rcube_sieve_engine.php on line 221 (GET /index.php?_task=settings&_action=plugin.managesieve-action&_framed=1&_nav=hide) php: PHP Error: Not currently connected (GET /index.php?_task=settings&_action=plugin.managesieve-action&_framed=1&_nav=hide) roundcube: PHP Error: Connection refused (POST /?_task=settings&_action=plugin.managesieve-save) roundcube: PHP Error: Unable to connect to managesieve on obfuscated.domain:4190 in /usr/local/www/roundcube/plugins/managesieve/lib/Roundcube/rcube_sieve_engine.php on line 221 (POST /?_task=settings&_action=plugin.managesieve-save) i don't understand why it can't connect, this seems to work fine: # gnutls-cli --tofu --starttls -p 4190 10.0.0.91 Processed 142 CA certificate(s). Resolving '10.0.0.91:4190'... Connecting to '10.0.0.91:4190'... - Simple Client Mode: "IMPLEMENTATION" "dovecot" "SIEVE" "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 mime foreverypart extracttext" "NOTIFY" "mailto" "SASL" "CRAM-MD5" "STARTTLS" "VERSION" "1.0" OK "Dovecot ready." STARTTLS OK "Begin TLS negotiation now." *** Starting TLS handshake - Certificate type: X.509 - Got a certificate list of 3 certificates. - Certificate[0] info: - subject `CN=obfuscated.domain.com', issuer `CN=R3,O=Let's Encrypt,C=US', serial xxxxxxxxxxxxxxxxxxxxxx, RSA key 2048 bits, signed using RSA-SHA256, activated `yyyy-mm-dd 17:48:15 UTC', expires `yyyy-mm-dd 17:48:14 UTC', pin-sha256="xxxxxxxxxxxxxxxxxxxxxx" Public Key ID: sha1:xxxxxxxxxxxxxxxxxxxxxx sha256:xxxxxxxxxxxxxxxxxxxxxx Public Key PIN: pin-sha256:xxxxxxxxxxxxxxxxxxxxxx - Certificate[1] info: - subject `CN=R3,O=Let's Encrypt,C=US', issuer `CN=ISRG Root X1,O=Internet Security Research Group,C=US', serial xxxxxxxxxxxxxxxxxxxxxx, RSA key 2048 bits, signed using RSA-SHA256, activated `yyyy-mm-dd 00:00:00 UTC', expires `yyyy-mm-dd 16:00:00 UTC', pin-sha256="xxxxxxxxxxxxxxxxxxxxxx" - Certificate[2] info: - subject `CN=ISRG Root X1,O=Internet Security Research Group,C=US', issuer `CN=DST Root CA X3,O=Digital Signature Trust Co.', serial yyyy-mm-dd, RSA key 4096 bits, signed using RSA-SHA256, activated `yyyy-mm-dd 19:14:03 UTC', expires `yyyy-mm-dd 18:14:03 UTC', pin-sha256="xxxxxxxxxxxxxxxxxxxxxx" - Status: The certificate is NOT trusted. The name in the certificate does not match the expected. *** PKI verification of server certificate failed... Host 10.0.0.91 (sieve) has never been contacted before. Its certificate is valid for 10.0.0.91. Are you sure you want to trust it? (y/N): y - Description: (TLS1.3-X.509)-(ECDHE-SECP256R1)-(RSA-PSS-RSAE-SHA256)-(AES-256-GCM) - Session ID: xx:yy:xx:yy:xx:yy... - Options: "IMPLEMENTATION" "dovecot" "SIEVE" "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 mime foreverypart extracttext" "NOTIFY" "mailto" "SASL" "CRAM-MD5" "VERSION" "1.0" OK "TLS negotiation successful." On 2022-11-23 13:35, Yassine Chaouche wrote: > also make sure your are editing config.php and not config.inc.php > (which you pasted) > > Yassine. > > Le 23 novembre 2022 8:30:36 PM GMT+01:00, Yassine Chaouche > <a.chaouche(a)algerian-radio.dz> a écrit : > >> good. we have established that the problem shouldn't be on dovecot's >> side. i suspect roundcube is misconfigured or can't connect for some >> reason. I believe someone mentioned SSL and TLS support problem in RC >> for a specific version? can you try without? also can you paste RC >> config? >> >> Yassine.

5 6

Kerberos/GSSAPI auth weirdness?
by r3pek 14 Dec '22

14 Dec '22

Hi list! I'm migrating my server into a new REALM (INT.R3PEK.ORG) from an old one (R3PEK.ORG) This is a completely new install and configuration, so no leftovers exits. The machine is correctly register into the REALM and users are able to login without a problem. Now, when I try to login using a Kerberos ticket, for some reason that I can't understand, dovecot is looking for a ticket on the old REALM. Maybe because of the email domain (which stayed the same)? This is the error message I see on the clients: "Failed to authenticate: Server krbtgt/R3PEK.ORG(a)INT.R3PEK.ORG" The one it should be looking for is krbtgt/INT.R3PEK.ORG(a)INT.R3PEK.ORG, but I can't seem to figure out where the problem is. # doveconf | grep -i realm auth_default_realm = INT.R3PEK.ORG auth_realms = INT.R3PEK.ORG What do I need to change to make this work?

1 0

sasl service for other app
by Henry R 13 Dec '22

13 Dec '22

can dovecot run as a general sasl service for other apps? such as webdav. Thanks.

6 10

Backups
by Doug Hardie 11 Dec '22

11 Dec '22

I started to investigate using doveadm backup to backup my mail system. I have a small number of users and the mail store is not large. It uses maildir format. I setup a test system that is not connected to the internet and started up dovecot. I used the following command to backup one user: doveadm backup -u ben remote:test ben is the user is in the mail store. Test is the actual server name. That worked just fine. The maildir was copied completely (as best as I can tell with ls). Then I tried the second user: doveadm backup -u jean remote:test This gives 2 error messages: doveadm(jean)[]<McRYOFI0jGN1MwAAmC+70w>: Error: Mailbox INBOX: Failed to get attribute vendor/vendor.dovecot/pvt/server/sieve/files/.dovecot: Mailbox attributes not enabled doveadm(jean)[]<0IwxIlI0jGMgUwAAZU03Dg>: Error: Remote command returned error 65: ssh test doveadm dsync-server -ujean -U In addition, the maildir directories are created, but there are no emails in any of them (e.g., cur). What is the problem with the 2nd and why does it behave differently from the first? -- Doug

4 8