alex84

Regarding sieve: this works now on my server: Note the protocol wrapping around mail_plugins (and I am using the old path of the scripts) protocol lda { mail_plugins { sieve = yes } } sieve_script personal { driver = file # path = ~/sieve # active_path = ~/.dovecot.sieve path = /var/customers/mail/sieve/%{user | domain}/%{user | username} active_path = /var/customers/mail/sieve/%{user | domain}/%{user | username}/.dovecot.sieve }

The main.cf from Postfix and an internet search brought me back to an old topic of mine 🧐 -> quota... does not show any obvious errors, but still does not work again (as it shoudl) I removed it and that's good enough for now for me. I'll investigate about sieve and quote later. Thank you very much for your time and input!

Ah I forgot to mention that: I upgraded over the last years from Buster (with Froxlor 1. or 0.x) via Bullseye and Bookworm (somewhere in between came Froxlor 2.x) now to Trixie. So the installation is running since a couple of years. The /var/customers/mail and its subfolders are (still) owned by vmail:vmail. I also suspect some too old configs or config/values. So now, I moved the exsting Dovecot configuration files away and reinstalled Dovecot apt-get -o DPkg::options::="--force-confmiss" install dovecot-core dovecot-imapd dovecot-managesieved dovecot-mysql dovecot-pop3d dovecot-sieve Then, I moved the new /etc/dovecot/conf.d -> /etc/dovecot/conf.d.frx and added the 99-froxlor.conf to a clean /etc/dovecot/conf.d folder. Still have to disable sieve, though: Error: Couldn't load required plugin /usr/lib/dovecot/modules/lib90_sieve_plugin.so: dlopen() failed: /usr/lib/dovecot/modules/lib90_sieve_plugin.so: undefined symbol: mail_deliver_ctx_get_log_var_expand_table -> This brought me one step (or two) ahaed: Login looks good, no error messages. -> I can send emails again to the outside world. -> I can receive emails originating from my server (either cronjobs or any other email from an emailadress configured on this server) BUT unfortunately, receiving from the outside world does still not work yet. Any idea where I can look next? A typical dialog looks like this: 2025-08-14T12:38:54.764069+02:00 v220240687080274156 postfix/smtpd[81709]: connect from anoubis.emsgrid.com[89.25.175.25] 2025-08-14T12:38:54.797207+02:00 v220240687080274156 postfix/smtpd[81709]: discarding EHLO keywords: CHUNKING 2025-08-14T12:38:54.931708+02:00 v220240687080274156 postfix/smtpd[81709]: discarding EHLO keywords: CHUNKING 2025-08-14T12:38:55.027763+02:00 v220240687080274156 policyd-spf[81711]: : prepend Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=89.25.175.25; helo=anoubis.emsgrid.com; envelope-from=sm.22324005868.ot0fgncnhj3ts44401-info=asdf.com@smgrid.com; receiver=myserver.com 2025-08-14T12:38:55.027941+02:00 v220240687080274156 postfix/smtpd[81709]: warning: connect to 127.0.0.1:12340: Connection refused 2025-08-14T12:38:55.028001+02:00 v220240687080274156 postfix/smtpd[81709]: warning: problem talking to server 127.0.0.1:12340: Connection refused 2025-08-14T12:38:56.028308+02:00 v220240687080274156 postfix/smtpd[81709]: warning: connect to 127.0.0.1:12340: Connection refused 2025-08-14T12:38:56.028570+02:00 v220240687080274156 postfix/smtpd[81709]: warning: problem talking to server 127.0.0.1:12340: Connection refused 2025-08-14T12:38:56.028623+02:00 v220240687080274156 postfix/smtpd[81709]: NOQUEUE: reject: RCPT from anoubis.emsgrid.com[89.25.175.25]: 451 4.3.5 <x.y@myserver.com>: Recipient address rejected: Server configuration problem; from=<sm.22324005868.ot0fgncnhj3ts44401-info=asdf.com@smgrid.com> to=<x.y@myserver.com> proto=ESMTP helo=<anoubis.emsgrid.com> 2025-08-14T12:38:56.028657+02:00 v220240687080274156 postfix/smtpd[81709]: using backwards-compatible default setting smtpd_relay_before_recipient_restrictions=no to reject recipient "x.y@myserver.com" from client "anoubis.emsgrid.com[89.25.175.25]" 2025-08-14T12:38:56.061914+02:00 v220240687080274156 postfix/smtpd[81709]: disconnect from anoubis.emsgrid.com[89.25.175.25] ehlo=2 starttls=1 mail=1 rcpt=0/1 data=0/1 rset=1 quit=1 commands=6/8 "server configuration problem" is not so helpful and just a "connection refused" also not. What should be running on 12340?

I appreciate any feedback (and your and your teams work) and please, don't feel pressured. (And I don't blame your or your team for my mistakes/carelessness) I compared the settings but it still does not work. The error remains the same. I am thinking about backing up the whole current config in /etc and remove/reinstall the dovecot packages. Do you see any problems with that? Mails, MySQL-Data should all remain in place, right? Do I then just add the 99-froxlor.conf or do I have to move other config files away?

Hi, so.... Trixie is released.... and... Dovecot is broken (I was a bit reckless yesterday and after a successful desktop upgrade i did the server upgrade) However, .... I adjusted the configs according to @MeinerEiner´ s hints. Those helped a great deal. But it's still not working: quota throws da(59714): Fatal: Couldn't load required plugin /usr/lib/dovecot/modules/lib11_imap_quota_plugin.so: dlopen() failed: /usr/lib/dovecot/modules/lib11_imap_quota_plugin.so: undefined symbol: command_unregister so i disabled it - I don't care for the moment Login is not working. I adjusted the sql part to userdb sql { iterate_query = "SELECT username AS user FROM mail_users WHERE (imap = 1 OR pop3 = 1)" query = SELECT CONCAT(homedir, maildir) AS home, 'maildir' as mail_driver, CONCAT(homedir, maildir) AS mail_path, uid, gid, CONCAT('*:storage=', quota, 'M') as quota_rule FROM mail_users WHERE (username = '%{user}' OR email = '%{user}') } #password_query = SELECT username AS user, password_enc AS password, CONCAT(homedir, maildir) AS userdb_home, uid AS userdb_uid, gid AS userdb_gid, CONCAT('maildir:', homedir, maildir) AS userdb_mail, CONCAT('*:storage=', quota, 'M') as userdb_quota_rule FROM mail_users WHERE (username = '%u' OR email = '%u') AND ((imap = 1 AND '%Ls' = 'imap') OR (pop3 = 1 AND '%Ls' = 'pop3') OR ((postfix = 'Y' AND '%Ls' = 'smtp') OR (> passdb sql { # default_password_scheme = CRYPT query = SELECT username AS user, password_enc AS password FROM mail_users WHERE (username = '%{user}' OR email = '%{user}') AND ((imap = 1 AND '%{protocol | lower}' = 'imap') OR (pop3 = 1 AND '%{protocol | lower}' = 'pop3') OR ((postfix = 'Y' AND '%{protocol | lower}' = 'smtp') OR (postfix = 'Y' AND '%{protocol | lower}' = 'sieve'))) } Now, seems to work, but not really completely. I still get dovecot: imap(mail@address.com)<60163><d5PoI088l/FbAjZ0>: Error: Mailbox INBOX: Failed to autocreate mailbox: Permission denied Any hint what's still missing? /Edit: Note: the mailboxes etc are already (or still) available on the filesystem. So, why the permissions problem and, more important, why the autocreate?

Ach klar, so kann man's auch machen.. Danke

Hallo, ich nutze Froxlor 2.1.8 (in Bulls-Eye/oldstable) und würde gerne PHP7.4 löschen. In den Froxlor-Einstellungen ist das allerdings als "System default" gesetzt und lässt sich nicht löschen - zumindest der Button fehlt in dieser Zeile: Wie bekomme ich die da raus? Kann ich die Version trotzdem bedenkenlos löschen (apt-get --purge remove) ? Alle anderen könnte ich löschen, obwohl in Benutzung, zumindest ist der Button da. Viele Grüße Alex

Hallo, ich nutze Lets Encrypt für SSL Zertifikate. Ich stolpere gerade darüber, das Dovecot (und evlt Postfix) nicht mitbekommen, wenn diese von Froxlor für die Domain erneuert werden und liefert weiterhin die alten Zertifikate aus. systemctl restart postfix systemctl restart dovecot beheben das zwar,aber sollte das nicht automatisch passieren? Übersehe ich permanent was in den Einstellungen?

Danke

Hallo zusammen, vielleicht bin ich blind, aber ich habe fast nichts dazu gefunden, unter welchen PHP-Versionen das Panel selbst lauffähig ist. Mir geht's vor allem um die aktuelle Begrenzung nach oben (PHP7.4 - vermeintliches Minimum - nutzt ja keiner mehr ernsthaft *scnr ) Meines läuft derzeit mit 8.0. Laut einem User hier läuft's wohl lauch unter 8.1. Wie schaut's mit 8.2 aus? Gibt's schon Erfahrungen mit 8.3? Danke und schönen Sonntag noch.

thanks

Hi, is there a way to reset the admin password e.g. from cli and without password-email? Or does the old way (add a new MD5 string to the db table) still work? In my table, the hash has a (example, not the real value) $2$10$ prefix... (its on a local testserver, without email. but i'd hate to start from scratch just because of the forgotten admin password) Kind regards