- dovecot - dovecot.org

problems restarting dovecot from cron: stderr not dup2()ed
by Thomas Klausner 15 Sep '26

15 Sep '26

Hi! I'm using dovecot-2.3.21.1 from pkgsrc on NetBSD 11.99.8/x86_64. I use certbot to handle dovecot's TLS certificate, and since I haven't implemented a nicer method yet, I restart dovecot weekly from cron to make sure it has the newest certificate. The symptom I see is that the restart doesn't finish until the next restart a week later - only then the previous restart's cron job "finishes" and I get an email. In a discussion on the NetBSD mailing list[1], RVP diagnosed it as follows: --- begin quote --- cron is still waiting for output from dovecot: 1. dovecot doesn't dup2() its STDERR_FILENO[2]--prolly to spew out error msgs. 2. when cron runs a crontab command, it dup2(STDOUT, STDERR)s the cmds stdout and stderr. 3. then it reads from this combined output and write()s this to sendmail's stdin. 4. all this works if either a) the program dies (ie. is not a daemon) so that cron can close sendmail's stdin.; or b) the daemon itself closes its stderr so an EOF can be generated when stdout is exhausted. 5. neither a) nor b) are satisfied in this case, so cron is still waiting for output which'll never arrive. 6. a week later, `service dovecot restart' kills the still running dovecot so cron can now "complete" the job and generate the email. --- end quote --- Can you please change dovecot to also dup(2) stderr to fix this? Thank you, Thomas [1] https://mail-index.netbsd.org/current-users/2026/09/13/msg047887.html [2]: https://github.com/dovecot/core/blob/main/src/master/main.c#L931

1 0

Issues after upgrading to the 2.4.* tree
by Kevin N Carpenter 15 Sep '26

15 Sep '26

Hi all. I've been running a Gentoo based mail server for decades - yeah, really! I usually upgrade it about once a year, which I just did. Been using the older dovecot versions until 2.4.5 got pulled in. Under Dovecot 2.3, I simply had the statement: *mail_location = maildir:~/.maildir* In dovecot/conf.d/10-mail and everything has been working. Ungraded my system and along comes 2.4 which changes configuration up. Using their converter, they suggest: *mail_driver = maildir mail_path = ~/.maildir * returns the error: /mail dovecot: imap(ccgllc)<5839><6sozAX5bskysEAQB>: Error: Namespace inbox: mail_storage settings: Failed to parse configuration: Invalid setting mail_path=~/.maildir: mail_path setting used home directory (~/) but there is no mail_home and userdb didn't return it / Setting *mail_home = /home/%{user}* gets closer, but then dovecot has permission errors reading the .maildir directory: /mail dovecot: imap(kevinc)<5541><6HubuX1bV42sEAQB>: Error: maildir: stat(/home/kevinc/.maildir/tmp) failed: Permission denied (euid=5000(<unknown>) egid=5000(<unknown>) missing +x perm: /home/kevinc/.maildir, dir owned by 1000:1000 mode=0700)/ Been going in circles for hours over this... all I want is a nice simply mairdir based imap server. btw - have SSL working just fine now. just added: *passdb pam { driver = pam pam_service_name = * } userdb passwd { driver = passwd } * since passdb and userdb were specified in my old configuration. Now getting: /auth-worker(kevinc,172.16.4.1)<5973></2FNPn5b1YqsEAQB>: request [1]: Error: pam: pam_authenticate() failed: Authentication failure (/etc/pam.d/* missing?/ /etc/pam.d/ exist, including the new dovecot statement. In any case, that last step feels like I am going backwards. Help please. Hi all. I've been running a Gentoo based mail server for decades - yeah, really! I usually upgrade it about once a year, which I just did. Been using the older dovecot versions until 2.4.5 got pulled in. Under Dovecot 2.3, I simply had the statement: mail_location = maildir:~/.maildir In dovecot/conf.d/10-mail and everything has been working. Ungraded my system and along comes 2.4 which changes configuration up. Using their converter, they suggest: mail_driver = maildir mail_path = ~/.maildir returns the error: mail dovecot: imap(ccgllc)<5839><6sozAX5bskysEAQB>: Error: Namespace inbox: mail_storage settings: Failed to parse configuration: Invalid setting mail_path=~/.maildir: mail_path setting used home directory (~/) but there is no mail_home and userdb didn't return it Setting mail_home = /home/%{user} gets closer, but then dovecot has permission errors reading the .maildir directory: mail dovecot: imap(kevinc)<5541><6HubuX1bV42sEAQB>: Error: maildir: stat(/home/kevinc/.maildir/tmp) failed: Permission denied (euid=5000(<unknown>) egid=5000(<unknown>) missing +x perm: /home/kevinc/.maildir, dir owned by 1000:1000 mode=0700) Been going in circles for hours over this... all I want is a nice simply mairdir based imap server. btw - have SSL working just fine now. just added: passdb pam { driver = pam pam_service_name = * } userdb passwd { driver = passwd } since passdb and userdb were specified in my old configuration. Now getting: auth-worker(kevinc,172.16.4.1)<5973></2FNPn5b1YqsEAQB>: request [1]: Error: pam: pam_authenticate() failed: Authentication failure (/etc/pam.d/* missing? /etc/pam.d/ exist, including the new dovecot statement. In any case, that last step feels like I am going backwards. Help please.

1 0

2.4.5 causes an infinite loop in maildir_is_internal_name
by Naofumi Honda 12 Sep '26

12 Sep '26

Dear developers, afte upgrading to the 2.4.5, all the imap-login processes always fall into infinite loops in maildir_is_inernal_name(). This is becuase, when mailbox_list_layout = fs, mlist->module_ctx.super.is_internal_name becomes the jus maidir_is_internal_name() itself, which causes infinite self-recursion loops (by a tail recursion optimization) as was the attached debug info shows.

2 3

2.4.5 build failures on s390x and loong64 architectures
by Noah Meyerhans 10 Sep '26

10 Sep '26

While preparing the Debian packages of Dovecot 2.4.5, I've encountered the following architecture-specific build failures: On s390x, the test suite fails in test-lib with: test-hash-method.c:483: Assert(#28) failed: memcmp(result, test_vectors[i].output, test_vectors[i].olen) == 0 test-hash-method.c:483: Assert(#29) failed: memcmp(result, test_vectors[i].output, test_vectors[i].olen) == 0 hash method xxh64 (test vectors) ..................................... : FAILED Full build log is visible at https://buildd.debian.org/status/fetch.php?pkg=dovecot&arch=s390x&ver=1%3A2… I haven't determined yet whether the problem is with the xxh64 hash method itself or with the test suite, but the failure is reliably reproducible. On loong64, the test suite fails in test-lib with: ostream multiplex stream (backpressure): sub-process ended properly .. : FAILED Warning: test: Sub-process forcibly terminated with signal 9 ostream multiplex stream (backpressure) .............................. : FAILED Complete log is at https://buildd.debian.org/status/fetch.php?pkg=dovecot&arch=loong64&ver=1%3… I'm working on debugging these, and will share whatever I find, but maybe somebody here will beat me to the solution. noah

2 3

Corrupt index.cache file causes looping and out of memory errors
by Evan Booyens 09 Sep '26

09 Sep '26

Dovecot -n and systems report are attached Log file extries: Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****(a)*******.co.za): Error: mremap_anon(/usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: mremap_anon(/usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: mremap_anon(/usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****(a)*******.co.za): Error: mremap_anon(/usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 www23.jnb2.host-h.net dovecot: imap(*****(a)*******.co.za): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Evan Booyens Linux Platform Engineer xneelo (Pty) Ltd SA Contact Centre: 0861 0861 08 International: +27 21 970 2000 Disclaimer: xneelo.co.za/email-disclaimer Formerly known as Hetzner SA Dovecot -n and systems report are attached Log file extries: Sep 9 14:29:32 [1]www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.[2]co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Sep 9 14:29:32 [3]www23.jnb2.host-h.net dovecot: imap(*****(a)*******.[4]co.za): Error: mremap_anon(/usr/www/dovecot_indexes/********.[5]co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 [6]www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: mremap_anon(/usr/www/dovecot_indexes/********.[7]co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 [8]www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.[9]co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Sep 9 14:29:32 [10]www23.jnb2.host-h.net dovecot: imap(*****@*******): Error: mremap_anon(/usr/www/dovecot_indexes/********.[11]co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 [12]www23.jnb2.host-h.net dovecot: imap(*****(a)*******.[13]co.za): Error: mremap_anon(/usr/www/dovecot_indexes/********.[14]co.za/orders/.Client Correspondence/dovecot.index.cache, 967524352) failed: Cannot allocate memory Sep 9 14:29:32 [15]www23.jnb2.host-h.net dovecot: imap(*****(a)*******.[16]co.za): Error: Mailbox INBOX.Client Correspondence: read() failed with file /usr/www/dovecot_indexes/********.[17]co.za/orders/.Client Correspondence/dovecot.index.cache: Cannot allocate memory Evan Booyens Linux Platform Engineer xneelo (Pty) Ltd SA Contact Centre: 0861 0861 08 International: +27 21 970 2000 Disclaimer: [18]xneelo.co.za/email-disclaimer Formerly known as Hetzner SA References Visible links 1. http://www23.jnb2.host-h.net/ 2. http://co.za/orders/.Client 3. http://www23.jnb2.host-h.net/ 4. http://co.za/ 5. http://co.za/orders/.Client 6. http://www23.jnb2.host-h.net/ 7. http://co.za/orders/.Client 8. http://www23.jnb2.host-h.net/ 9. http://co.za/orders/.Client 10. http://www23.jnb2.host-h.net/ 11. http://co.za/orders/.Client 12. http://www23.jnb2.host-h.net/ 13. http://co.za/ 14. http://co.za/orders/.Client 15. http://www23.jnb2.host-h.net/ 16. http://co.za/ 17. http://co.za/orders/.Client 18. http://xneelo.co.za/email-disclaimer

1 0

Crash on FreeBSD 14.4 with dovecot 2.4.5
by Frank Volf 09 Sep '26

09 Sep '26

Hi, On a daily basis I see a anvil crash in my logs for dovecot 2.4.5 on FreeBSD. Not sure if this is a FreeBSD specific issue, a dovecot issue, or a configuration issue (this is the first system converted from 2.3 configuration syntax): Sep 9 03:20:00 mail dovecot[53730]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 8 16:40:00 mail dovecot[53730]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 7 20:47:28 mail dovecot[6215]: anvil: Panic: file anvil-connection.c: line 838 (anvil_connection_destroy): assertion failed: (hash_conn == conn) Sep 6 17:07:24 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 6 20:40:01 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 6 21:53:59 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached The "assertion failed" only happened once, but the "unreached" happen on a daily basis. About 20 minutes (more or less) after each crash I see also the following message: Sep 8 16:56:32 mail dovecot[53730]: anvil: Error: connect limit: disconnection for unknown (pid=59834, user=xxxx, service=imap, ip=x.x.x.x, conn_guid=7925eb292e15a06abae9000050df13ba) Can somebody please give me some pointers on how to solve or troubleshoot this? Kind regards, Frank Hi, On a daily basis I see a anvil crash in my logs for dovecot 2.4.5 on FreeBSD. Not sure if this is a FreeBSD specific issue, a dovecot issue, or a configuration issue (this is the first system converted from 2.3 configuration syntax): Sep 9 03:20:00 mail dovecot[53730]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 8 16:40:00 mail dovecot[53730]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 7 20:47:28 mail dovecot[6215]: anvil: Panic: file anvil-connection.c: line 838 (anvil_connection_destroy): assertion failed: (hash_conn == conn) Sep 6 17:07:24 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 6 20:40:01 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached Sep 6 21:53:59 mail dovecot[96610]: anvil: Panic: file anvil-connection.c: line 837: unreached The "assertion failed" only happened once, but the "unreached" happen on a daily basis. About 20 minutes (more or less) after each crash I see also the following message: Sep 8 16:56:32 mail dovecot[53730]: anvil: Error: connect limit: disconnection for unknown (pid=59834, user=xxxx, service=imap, ip=x.x.x.x, conn_guid=7925eb292e15a06abae9000050df13ba) Can somebody please give me some pointers on how to solve or troubleshoot this? Kind regards, Frank

1 0

iOS email client / imap push
by Jorge Bastos 08 Sep '26

08 Sep '26

Howdy, Does anyway had an issue with latest iOS's, where the native email client put the messages as seen, when the only thing that is done is just "refresh", the message wasn't open, i tested myself. Anyone/is it known, if yes any workarround or wait for apple to solve it? Howdy, Does anyway had an issue with latest iOS's, where the native email client put the messages as seen, when the only thing that is done is just "refresh", the message wasn't open, i tested myself. Anyone/is it known, if yes any workarround or wait for apple to solve it?

1 0

Re: Help with sieve
by Jorge Bastos 08 Sep '26

08 Sep '26

António, Oh, i'm so sorry for the question and the time of you all. Yes, i'm missing the stop on the 1st rule, sieve executes them from top to bottom. Solved, and thank you all for the time, On 2026-09-06 22:14, Antonio Oliveira via dovecot wrote: > On 9/6/2026 9:48 PM, Jorge Bastos via dovecot wrote: > >> Hi all, >> >> Well, my issue is not exacly what you descrived, >> >> I have an account, that i want to forward all incoming emails, so far >> so >> good. >> Before the emails are forward, i want to delete some, based on the >> sender >> domain. >> >> Thing is, i think managesieve roundcube plugin is producing the script >> wrong, here's what i have, i have 1 rule for the delete some senders, >> ok, >> and the 2nd for the redirect >> >> # rule:[blocks] >> if anyof (header :contains "from" "clubefashion.com", header :contains >> "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") >> { >> discard; >> redirect "jorge(a)domain.pt"; >> } >> # rule:[redir] >> if true >> { >> redirect "jorge(a)domain.pt"; >> } > > Hello, > > if you want to forward all of the emails, regardless of the sender > domain and then "delete" them from the original inbox I would do > (option 1): > > # rule:[blocks] > if anyof (header :contains "from" "clubefashion.com", header :contains > "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") > { > redirect "jorge(a)domain.pt"; > discard; > stop; > } > # rule:[redir] > if true > { > redirect "jorge(a)domain.pt"; > stop; > } > > If, on the other hand you do not want do forward them and also do not > want to keep them then simply do (option 2): > > # rule:[blocks] > if anyof (header :contains "from" "clubefashion.com", header :contains > "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") > { > discard; > stop; > } > # rule:[redir] > if true > { > redirect "jorge(a)domain.pt"; > stop; > } > > option 1 - Email is redirected, discarded and rule processing stops > (which inhibits the implicit "keep" from running, so the email is > redirected to the other mailbox but not kept on the original one). > > option 2 - Email from those domains is discarded and rule processing > stops, again inhibiting the implicit keep. For the other domains, email > is redirected and "implicilty" kept on the original mailbox. > > if you're using roundcube you should have an action when creating the > rule to "stop evaluating rules" or something else to that effect when > creating the rule, that will add the "stop;" line on each rule. > you just need to add it in the correct order to make it work. > > In your first message you stated that you wanted to avoid having those > blocked domain emails from being stored in the original mailbox, so > option 2 should work correctly for that. The rules you have in this > message will redirect both the "blocked" and all other emails to being > redirected to the other mailbox, which is what option 1 gives you. With > the rules you sent you would get *two* emails from the "blocked" > domains being redirected to "jorge(a)domain.pt" since both rules would > match. > > Hope this clarifies and helps you achieve the desired outcome. Both are > entirely possible with roundcube generated sieve rules, you just need > to decide on which of the two you want. > > Regards, > >> On 2026-09-06 10:15, Toth Attila via dovecot wrote: >> >> I'm using a filter that moves the mail to another directory. The >> original message disappeared only from the initial location if >> imapsieve_expunge_discarded = yes was set. A duplicate remained >> otherwise. I suggest to take a look at on this configuration option >> depending on the purpose of the filter. >> >> On 2026.09.06 02:30, Stephan Bosch via dovecot wrote: >> >> Op 2-9-2026 om 16:00 schreef Antonio Oliveira via dovecot: >> >> Hello, >> >> You can try to add "stop;" after "discard;", so that rule processing >> stops after the discard and doesn't trickle down to the other rule. >> >> Yes, this will help. Why? The discard action command only cancels the >> implicit keep; nothing else. It means that if and only if no other >> explicit action delivers or preserves (keeps!) the message somehow, it >> is discarded. So, if subsequently redirect is triggered, the message >> is redirected, thereby making the discard action ineffective. Stopping >> the script right after discard will make sure no other actions can be >> triggered, thereby making sure the message is discarded (provided no >> earlier actions already triggered some sort of delivery either). >> Without discard or any other action command that cancels the implicit >> keep (such as redirect, fileinto, and ... keep), the default action is >> to "keep" the message, e.g. file it to INBOX. >> >> Regards, >> >> Stephan. >> >> On 9/2/2026 1:05 PM, Jorge Bastos via dovecot wrote: >> >> Hi guys, >> >> Small help with a sieve script, I have the script below in a >> useraccount, >> in this exact order. >> The redirect works, but the discard doesn't. >> A message that comes from the domain "clubefashing.com" or >> "nutribody.pt", >> is always redirected. >> >> Any idea what is wrong here? >> >> # rule:[blocks] >> if anyof (header :contains "from" "clubefashion.com", header >> :contains >> "from" "nutribody.pt") >> { >> discard; >> } >> # rule:[redir] >> if true >> { >> redirect "[1]jorge(a)decimal.pt"; >> } >> >> _______________________________________________ >> dovecot mailing list -- [2]dovecot(a)dovecot.org >> To unsubscribe send an email to [3]dovecot-leave(a)dovecot.org >> >> _______________________________________________ >> dovecot mailing list -- [4]dovecot(a)dovecot.org >> To unsubscribe send an email to [5]dovecot-leave(a)dovecot.org >> >> References >> >> Visible links >> 1. mailto:jorge@decimal.pt >> 2. mailto:dovecot@dovecot.org >> 3. mailto:dovecot-leave@dovecot.org >> 4. mailto:dovecot@dovecot.org >> 5. mailto:dovecot-leave@dovecot.org >> >> _______________________________________________ >> dovecot mailing list -- dovecot(a)dovecot.org >> To unsubscribe send an email to dovecot-leave(a)dovecot.org Antonio, Oh, i'm so sorry for the question and the time of you all. Yes, i'm missing the stop on the 1st rule, sieve executes them from top to bottom. Solved, and thank you all for the time, On 2026-09-06 22:14, Antonio Oliveira via dovecot wrote: On 9/6/2026 9:48 PM, Jorge Bastos via dovecot wrote: Hi all, Well, my issue is not exacly what you descrived, I have an account, that i want to forward all incoming emails, so far so good. Before the emails are forward, i want to delete some, based on the sender domain. Thing is, i think managesieve roundcube plugin is producing the script wrong, here's what i have, i have 1 rule for the delete some senders, ok, and the 2nd for the redirect # rule:[blocks] if anyof (header :contains "from" "clubefashion.com", header :contains "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") { discard; redirect "[1]jorge(a)domain.pt"; } # rule:[redir] if true { redirect "[2]jorge(a)domain.pt"; } Hello, if you want to forward all of the emails, regardless of the sender domain and then "delete" them from the original inbox I would do (option 1): # rule:[blocks] if anyof (header :contains "from" "clubefashion.com", header :contains "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") { redirect "[3]jorge(a)domain.pt"; discard; stop; } # rule:[redir] if true { redirect "[4]jorge(a)domain.pt"; stop; } If, on the other hand you do not want do forward them and also do not want to keep them then simply do (option 2): # rule:[blocks] if anyof (header :contains "from" "clubefashion.com", header :contains "from" "nutribody.pt", header :contains "from" "mcdonalds.pt") { discard; stop; } # rule:[redir] if true { redirect "[5]jorge(a)domain.pt"; stop; } option 1 - Email is redirected, discarded and rule processing stops (which inhibits the implicit "keep" from running, so the email is redirected to the other mailbox but not kept on the original one). option 2 - Email from those domains is discarded and rule processing stops, again inhibiting the implicit keep. For the other domains, email is redirected and "implicilty" kept on the original mailbox. if you're using roundcube you should have an action when creating the rule to "stop evaluating rules" or something else to that effect when creating the rule, that will add the "stop;" line on each rule. you just need to add it in the correct order to make it work. In your first message you stated that you wanted to avoid having those blocked domain emails from being stored in the original mailbox, so option 2 should work correctly for that. The rules you have in this message will redirect both the "blocked" and all other emails to being redirected to the other mailbox, which is what option 1 gives you. With the rules you sent you would get *two* emails from the "blocked" domains being redirected to "[6]jorge(a)domain.pt" since both rules would match. Hope this clarifies and helps you achieve the desired outcome. Both are entirely possible with roundcube generated sieve rules, you just need to decide on which of the two you want. Regards, On 2026-09-06 10:15, Toth Attila via dovecot wrote: I'm using a filter that moves the mail to another directory. The original message disappeared only from the initial location if imapsieve_expunge_discarded = yes was set. A duplicate remained otherwise. I suggest to take a look at on this configuration option depending on the purpose of the filter. On 2026.09.06 02:30, Stephan Bosch via dovecot wrote: Op 2-9-2026 om 16:00 schreef Antonio Oliveira via dovecot: Hello, You can try to add "stop;" after "discard;", so that rule processing stops after the discard and doesn't trickle down to the other rule. Yes, this will help. Why? The discard action command only cancels the implicit keep; nothing else. It means that if and only if no other explicit action delivers or preserves (keeps!) the message somehow, it is discarded. So, if subsequently redirect is triggered, the message is redirected, thereby making the discard action ineffective. Stopping the script right after discard will make sure no other actions can be triggered, thereby making sure the message is discarded (provided no earlier actions already triggered some sort of delivery either). Without discard or any other action command that cancels the implicit keep (such as redirect, fileinto, and ... keep), the default action is to "keep" the message, e.g. file it to INBOX. Regards, Stephan. On 9/2/2026 1:05 PM, Jorge Bastos via dovecot wrote: Hi guys, Small help with a sieve script, I have the script below in a useraccount, in this exact order. The redirect works, but the discard doesn't. A message that comes from the domain "clubefashing.com" or "nutribody.pt", is always redirected. Any idea what is wrong here? # rule:[blocks] if anyof (header :contains "from" "clubefashion.com", header :contains "from" "nutribody.pt") { discard; } # rule:[redir] if true { redirect "[1][7]jorge(a)decimal.pt"; } _______________________________________________ dovecot mailing list -- [2][8]dovecot(a)dovecot.org To unsubscribe send an email to [3][9]dovecot-leave(a)dovecot.org _______________________________________________ dovecot mailing list -- [4][10]dovecot(a)dovecot.org To unsubscribe send an email to [5][11]dovecot-leave(a)dovecot.org References Visible links 1. mailto:[12]jorge@decimal.pt 2. mailto:[13]dovecot@dovecot.org 3. mailto:[14]dovecot-leave@dovecot.org 4. mailto:[15]dovecot@dovecot.org 5. mailto:[16]dovecot-leave@dovecot.org _______________________________________________ dovecot mailing list -- [17]dovecot(a)dovecot.org To unsubscribe send an email to [18]dovecot-leave(a)dovecot.org References Visible links 1. mailto:jorge@domain.pt 2. mailto:jorge@domain.pt 3. mailto:jorge@domain.pt 4. mailto:jorge@domain.pt 5. mailto:jorge@domain.pt 6. mailto:jorge@domain.pt 7. mailto:jorge@decimal.pt 8. mailto:dovecot@dovecot.org 9. mailto:dovecot-leave@dovecot.org 10. mailto:dovecot@dovecot.org 11. mailto:dovecot-leave@dovecot.org 12. mailto:jorge@decimal.pt 13. mailto:dovecot@dovecot.org 14. mailto:dovecot-leave@dovecot.org 15. mailto:dovecot@dovecot.org 16. mailto:dovecot-leave@dovecot.org 17. mailto:dovecot@dovecot.org 18. mailto:dovecot-leave@dovecot.org

1 0

Problem email client iPhone ios18.2
by andrzej.milewski＠gmail.com 08 Sep '26

08 Sep '26

Hello, I have a fairly new installation of the Dovecot server version 2.3.19.1 on Debian 12.8. In general, the server works correctly. It sends and receives emails without issues. The clients primarily use Thunderbird on desktops and Gmail on Android. However, some clients are using iPhones with various hardware versions. Those with iOS below version 18 work correctly without any issues. Those who have updated their phones to version 18 and now even 18.2 are experiencing problems with mailbox synchronization. On version 18.2, the issue appears as follows: When you open the app, you can read several messages. They load correctly, but, for instance, attachments do not. If I wait for some time, the attachment eventually loads. During this time, I observe the server logs, and it shows that a certain session on the Dovecot server has ended. Dec 17 13:13:01 iredmail0 dovecot[408]: imap(xxx(a)xxx.com)<44785><1QGZCHcp/+3AqM2O>: Disconnected: Connection closed (UID SEARCH finished 92.295 secs ago) in=2972 out=6076956 deleted=0 expunged=0 trashed=0 hdr_count=0 hdr_bytes=0 body_count=36 body_bytes=6059064 Dec 18 08:40:41 iredmail0 dovecot[408]: imap(xxx(a)xxx.com)<177412>: Disconnected: Connection closed (UID FETCH finished 90.079 secs ago) in=5126 out=1304237 deleted=0 expunged=0 trashed=0 hdr_count=94 hdr_bytes=289688 body_count=89 body_bytes=969615 Dec 18 08:42:19 iredmail0 dovecot[408]: imap(xxx(a)xxx.com)<177410><D5mqU4cpZ+nAqM2O>: Disconnected: Connection closed (UID FETCH finished 95.051 secs ago) in=4475 out=1162785 deleted=0 expunged=0 trashed=0 hdr_count=95 hdr_bytes=281674 body_count=61 body_bytes=839545 Dec 18 08:44:17 iredmail0 dovecot[408]: imap(edyta.kowalewska(a)laktopol.com.pl)<177918>: Disconnected: Connection closed (UID FETCH finished 92.132 secs ago) in=29399 out=30266644 deleted=0 expunged=0 trashed=0 hdr_count=565 hdr_bytes=1726228 body_count=628 body_bytes=28109455 After the session is closed, the attachment loads immediately. From the logs and router analysis, it appears that the download happens in another previously opened session. After this incident, I go to another message with an attachment, and the situation repeats. There are also situations where the email body does not load, and the app waits for the server until the session closes and the content is fetched in another session. It does not matter whether I connect to the server via the internet or through a local network. On an iPhone 7 with iOS 15.8.3, everything works smoothly. Attachments open without issues. On iOS 18.1.1, attachments do not load, and the session does not close on its own. It takes forever unless the app is closed. In the Gmail app, everything works correctly without any issues. This is clearly a problem with the mail client on iPhone. However, is there anything I can adjust in my server settings to mitigate this issue for users? My dovecot configuration https://pastebin.com/UPNQ5dZz

26 60

mail-crypt: user key generated with crypt_user_key_password still usable without it
by Twinsen＠die-optimisten.net 08 Sep '26

08 Sep '26

Hi, I please need help. Dovecot 2.4.1-4 (docker-mailserver 16.0.1), Maildir, passwd-file for both passdb and userdb. Config: mail_plugins = mail_crypt mail_attribute { dict file { path = %{home}/Maildir/dovecot-attributes } } crypt_user_key_curve = secp521r1 crypt_user_key_require_encrypted = yes I also added a second passdb in Lua that returns the login password as a sha256 hex string: passdb passdb-default { result_success = continue-ok } passdb lua { lua_file = /etc/dovecot/lua/crypt.lua skip = unauthenticated use_worker = yes } The Lua script returns PASSDB_RESULT_OK with: { noauthenticate = "yes", userdb_crypt_user_key_password = <sha256 hex> } What I have verified so far: doveadm auth login lists crypt_user_key_password under "userdb extra fields", so the field reaches userdb. With crypt_user_key_require_encrypted = no and the attributes file removed, an IMAP login auto-creates a user key. That key is unencrypted. With crypt_user_key_require_encrypted = yes and the attributes file removed, no key is created. IMAP logs: dict(file): stat(.../Maildir/dovecot-attributes) failed: No such file or directory So I generated the key manually: doveadm -o crypt_user_key_password=<sha256 hex> \ mailbox cryptokey generate -u user(a)example.com -Uf This succeeded. The new key is active, the old one shows as inactive, and the attributes file grew from 2280 to 3147 bytes. Now incoming mail is stored encrypted, the files start with CRYPTED. However, this still returns the headers in clear text, without any password and without error: doveadm fetch -u user(a)example.com hdr mailbox INBOX So the key appears to be usable without the password at all. My main question is: Does -o crypt_user_key_password actually encrypt the user key when it is generated, or is it only used to unlock an already encrypted key? Related to that: Can an old inactive user key still be used for decryption in this setup? Is there any marker that indicates in the attributes dict whether a user key is password-encrypted? With passwd-file as userdb (no prefetch), should an extra field from a passdb reach the IMAP process that creates or loads the key? Any pointers apreciated. Best regards, Twinsen

1 0