- dovecot - dovecot.org

"Group doesn't exist: dovecot"
by James Brown 19 Nov '18

19 Nov '18

No matter what I do, I always get this error when trying to start Dovecot: $ sudo /usr/local/opt/dovecot/sbin/dovecot Warning: fd limit (ulimit -n) is lower than required under max. load (256 < 1000), because of default_client_limit Fatal: service(stats) Group doesn't exist: dovecot (See service stats { unix_listener /usr/local/var/run/dovecot/stats-writer { group } } setting) The user and group it should be using are _dovecot. Config: $ sudo /usr/local/opt/dovecot/sbin/dovecot -n # 2.3.2.1 (0719df592): /usr/local/etc/dovecot/dovecot.conf # OS: Darwin 18.2.0 x86_64 apfs # Hostname: mailbordocomau-dr.local doveconf: Error: t_readlink(/usr/local/var/run/dovecot/dovecot.conf) failed: readlink() failed: No such file or directory default_internal_user = _dovenull default_login_user = _dovecot first_valid_gid = 0 first_valid_uid = 102 last_valid_gid = 500 last_valid_uid = 105 mail_gid = 102 mail_location = maildir:/usr/local/virtual/%d/%u mail_uid = 102 namespace inbox { inbox = yes location = mailbox Drafts { special_use = \Drafts } mailbox Junk { special_use = \Junk } mailbox Sent { special_use = \Sent } mailbox "Sent Messages" { special_use = \Sent } mailbox Trash { special_use = \Trash } prefix = } passdb { args = /usr/local/etc/dovecot/dovecot-sql.conf.ext driver = sql } protocols = imap pop3 service auth { unix_listener /var/spool/postfix/private/auth { group = _postfix mode = 0660 user = _postfix } unix_listener auth-userdb { mode = 0660 user = _vmail } } ssl_cert = </usr/local/etc/dovecot/cert.pem ssl_dh = # hidden, use -P to show it ssl_key = # hidden, use -P to show it userdb { driver = passwd } userdb { driver = prefetch } userdb { args = /usr/local/etc/dovecot/dovecot-sql.conf.ext driver = sql } Not sure why it’s looking for dovecot.conf in the wrong place. Any suggestions? thanks, James.

1 0

Solr 7.5.0 managed-schema
by Daniel Miller 19 Nov '18

19 Nov '18

The attached is the Dovecot schema I am now using for Solr 7.5.0 - no "deprecated" warnings! -- Daniel

2 1

macOS Mojave: setgroups(501) failed: Too many extra groups
by Heiko W. Rupp 18 Nov '18

18 Nov '18

Hello, I have recently upgraded to macOS 10.14 (Mojave) and am running into an issue where one use can no longer log into dovecot via imap. Log shows Oct 11 08:10:27 imap(hwr)<12659><YpC/0+133+EAAAAAAAAAAAAAAAAAAAAB>: Fatal: setgroups(501) failed: Too many extra groups and indeed, the user is in 17 groups, which is more than NGROUPS_MAX (16). Another user with << 16 groups can log in fine. Unfortunately it is not (easily) doable to reduce the number of groups, as macOS seems to set them internally. Is there a config option that I am missing to work around this? Looking at the source, I see this is handled in src/lib/restrict-access.c::fix_groups_list(), where above the call to setgroups() a gid_list2 is constructed. I wonder if one could have a config option to prevent adding all those extra groups, which then make the call to setgroups() fail Any help appreciated Heiko -- hwr(a)pilhuhn.de m:0179/207 4919 b:http://pilhuhn.blogspot.com

4 6

Re: Issue sharing folders with Thunderbird
by Jason Perry 17 Nov '18

17 Nov '18

Hi Aki I see what you are saying, and no I don’t believe so. I’m using typical /etc/passwd for the user db. I reviewed https://wiki.dovecot.org/UserDatabase/ExtraFields <https://wiki.dovecot.org/UserDatabase/ExtraFields> but I’m not clear on how to actually return system_user userdb attribute. I also have set mail_user_groups = sharedusers in dovecot.conf and then in /etc/group put all the users with mailboxes into the sharedusers group. I can see why that wouldn’t work if the above is not set. Can you point me in the right direction on how to pass that information? Thanks! > On Aug 28, 2018, at 12:40 AM, Aki Tuomi <aki.tuomi(a)dovecot.fi> wrote: > > Are you returning system_user userdb attribute? Extra groups from /etc/group are not applied without this. > > --- > Aki Tuomi > Dovecot oy > > -------- Original message -------- > From: Jason Perry <jason.perry(a)dtainc.us> > Date: 28/08/2018 04:27 (GMT+02:00) > To: Aki Tuomi <aki.tuomi(a)dovecot.fi> > Cc: dovecot(a)dovecot.org > Subject: Re: Issue sharing folders with Thunderbird > > Hi Aki, they are. > >> 3. The permissions for the file are: >> root@dal-notify-01 Maildir]# ls -la /var/lib/dovecot/db/shared-mailboxes.db >> -rw-rw-r--. 1 operations sharedusers 197 Aug 26 20:33 /var/lib/dovecot/db/shared-mailboxes.db >> 4. Users I'm testing with are part of the sharedusers group in /etc/group > > The dir /var/lib/dovecot/db has this for permissions: > drwsrwsrwx. 2 root sharedusers 33 Aug 27 21:12 db > > the user operations is a member of “sharedusers” in /etc/group. I even chmod’ed the permissions after to rw for global and I get the same error/result. > > The file does appear to get written to. For example, lets say user u2.name creates a folder called “starbucks” and uses the IMAP commands to share it out to user operations. It seems to write to the acl_shared_dict file in /var/lib/dovecot/db/shared-mailboxes.db with an entry like >> shared/shared-boxes/user/u2.name/operations >> 1 > > And it writes to the ~/Maildir/.starbucks/dovecot-acl file for u2.user with an entry like: > user=operations lrw > > So it appears to be working. However, if I log into Thunderbird with the user operations account and go to subscribe, I do not see “starbucks” in the list of available folders. And I still get the nfs_flush_chown error. > > Is there anything else I can be doing with the file/dir permissions? > > Thanks. > > >> On Aug 27, 2018, at 10:50 AM, Aki Tuomi <aki.tuomi(a)dovecot.fi <mailto:aki.tuomi@dovecot.fi>> wrote: >> >> acl_shared_dict file & folder must be readwritable by user performing the sharing as stated in wiki. >> >> https://wiki.dovecot.org/SharedMailboxes/Shared <https://wiki.dovecot.org/SharedMailboxes/Shared> >> --- >> Aki Tuomi >> Dovecot oy >> >> -------- Original message -------- >> From: Jason Perry <jason.perry(a)dtainc.us <mailto:jason.perry@dtainc.us>> >> Date: 27/08/2018 17:01 (GMT+02:00) >> To: dovecot(a)dovecot.org <mailto:dovecot@dovecot.org> >> Subject: Issue sharing folders with Thunderbird >> >> Trying to share folder "JasonAlerts" within mailbox operations to user u1.name >> >> Issues: >> 1. Shared mailboxes do not appear under subscriptions in Thunderbird for another user >> 2. When I try to set permissions via IMAP commands I get in /var/log/dovecot.log >> Aug 26 20:33:24 imap(operations): Error: nfs_flush_chown_uid: chown(/var/lib/dovecot/db) failed: Permission denied >> 3. The permissions for the file are: >> root@dal-notify-01 Maildir]# ls -la /var/lib/dovecot/db/shared-mailboxes.db >> -rw-rw-r--. 1 operations sharedusers 197 Aug 26 20:33 /var/lib/dovecot/db/shared-mailboxes.db >> 4. Users I'm testing with are part of the sharedusers group in /etc/group >> 5. IMAP commands DO write to the file: >> [root@dal-notify-01 Maildir]# cat /var/lib/dovecot/db/shared-mailboxes.db >> shared/shared-boxes/user/u1.name/u2.name >> 1 >> shared/shared-boxes/user/operations/u2.name >> 1 >> shared/shared-boxes/user/u1.name/operations >> 1 >> shared/shared-boxes/user/u1.name/operations >> 1 >> 6. It will also write to the user's /home/operations/Maildir/.JasonAlert/dovecot-acl file >> user=u1.name ilrw >> >> However if I got into Thunderbird and try to subscribe, none of the shared folders are listed. Can anyone help?? Thanks!! >> >> Dovecot version = 2.2.10 >> >> dovecot -n >> [root@server]# dovecot -n >> # 2.2.10: /etc/dovecot/dovecot.conf >> # OS: Linux 3.10.0-693.17.1.el7.x86_64 x86_64 CentOS Linux release 7.4.1708 (Core) >> first_valid_uid = 100 >> listen = * >> log_path = /var/log/dovecot.log >> mail_debug = yes >> mail_location = maildir:~/Maildir >> mail_plugins = acl >> mail_privileged_group = mail >> mbox_write_locks = fcntl >> namespace Private { >> hidden = no >> inbox = no >> list = no >> location = >> prefix = Private/ >> separator = / >> type = private >> } >> namespace Public { >> hidden = no >> inbox = no >> list = no >> location = maildir:/var/mail/public >> prefix = Public/ >> separator = / >> subscriptions = yes >> type = public >> } >> namespace Shared { >> hidden = no >> inbox = no >> list = yes >> location = maildir:/var/mail/%u:INDEXPVT=~/Maildir/%u >> prefix = shared/%u/ >> separator = / >> subscriptions = yes >> type = shared >> } >> namespace inbox { >> inbox = yes >> location = >> mailbox Drafts { >> special_use = \Drafts >> } >> mailbox Junk { >> special_use = \Junk >> } >> mailbox Sent { >> special_use = \Sent >> } >> mailbox "Sent Messages" { >> special_use = \Sent >> } >> mailbox Trash { >> special_use = \Trash >> } >> prefix = >> separator = / >> } >> passdb { >> driver = pam >> } >> passdb { >> driver = pam >> } >> plugin { >> acl = vfile >> acl_shared_dict = file:/var/lib/dovecot/db/shared-mailboxes.db <file:///var/lib/dovecot/db/shared-mailboxes.db> >> } >> protocols = imap >> ssl_cert = </etc/pki/dovecot/certs/dovecot.pem >> ssl_key = </etc/pki/dovecot/private/dovecot.pem >> syslog_facility = local2 >> userdb { >> driver = passwd >> } >> userdb { >> driver = passwd >> } >> protocol imap { >> mail_plugins = acl imap_acl >> } >

1 1

RSA+ECC certificates
by Alex Hautequest 17 Nov '18

17 Nov '18

Any way of using both a RSA and an ECC certificates, letting the client pick the one to use? Alex

2 1

Mail slowed down to a crawl...
by StarionTech (IMAP) 16 Nov '18

16 Nov '18

Been moving along just fine for a couple years now, then in the last two days, email has slowed to a crawl. Retrieving email via IMAP is very slow, progress bar in the mail client shows that it is downloading messages almost constantly, like it never closes the connection, or it’s getting the messages very slowly. The server that dovecot/postfix is on has also bogged down to a crawl. The postfix queue is sitting with hundreds of queued up good customer email. It’s almost acting like dovecot is not getting the messages from postfix, or they’re being delayed for some reason. How to approach fixing this? Jeff J.

4 4

huge increase in storage activity afther dovecot upgrade
by Adrian M 16 Nov '18

16 Nov '18

Hi, we upgraded our servers from version 2.2.13 to 2.2.27. After the upgrade we notice a 10x increase in traffic with the nfs storage an errors like this in the logfile: Nov 12 09:48:16 mail dovecot: imap(...): Error: Corrupted index cache file /.../dovecot.index.cache: invalid record size Nov 12 09:48:16 mail dovecot: imap(...): Error: unlink(/.../dovecot.index.cache) failed: No such file or directory (in mail-cache.c:29) Nov 12 09:48:16 mail dovecot: imap(...): Error: Corrupted index cache file /.../dovecot.index.cache: invalid record size Nov 12 09:48:16 mail dovecot: imap(...): Error: Broken file /.../dovecot-uidlist line 8: Invalid data: Is this normal ? Will the activity wind down ? Can we do something, like deleting the old dovecot.index* or dovecot-uidlist files from maildirs, or doing an doveadm force-refresh for all inboxes ? Thank you ! Here's my configuration: # 2.2.27 (c0f36b0): /etc/dovecot/dovecot.conf # Pigeonhole version 0.4.16 (fed8554) # OS: Linux 4.9.0-8-amd64 x86_64 Debian 9.6 nfs auth_failure_delay = 15 secs auth_mechanisms = plain login auth_verbose = yes auth_worker_max_count = 256 default_client_limit = 40000 default_process_limit = 512 dict { lastlogin = mysql:/etc/dovecot/mysql/dovecot-dict-lastlogin.conf quotadict = mysql:/etc/dovecot/mysql/dovecot-dict-quota.conf } disable_plaintext_auth = no first_valid_uid = 100 log_timestamp = "%Y-%m-%d %H:%M:%S " login_log_format_elements = user=<%u> method=%m rip=%r lip=%l pid=%p %c mail_fsync = always mail_location = maildir:*/home/virtual/*%d/%u mail_max_userip_connections = 16 mail_nfs_index = yes mail_nfs_storage = yes mail_plugins = zlib quota mail_log notify mail_privileged_group = mail mailbox_list_index = yes maildir_stat_dirs = yes managesieve_notify_capability = mailto managesieve_sieve_capability = fileinto reject envelope encoded-character vacation subaddress comparator-i;ascii-numeric relational regex imap4flags copy include variables body enotify environment mailbox date ihave mmap_disable = yes namespace inbox { inbox = yes location = mailbox Drafts { special_use = \Drafts } mailbox Junk { special_use = \Junk } mailbox Sent { auto = subscribe special_use = \Sent } mailbox "Sent Messages" { special_use = \Sent } mailbox Spam { auto = subscribe special_use = \Junk } mailbox Trash { auto = subscribe special_use = \Trash } mailbox virtual/All { special_use = \All } prefix = } passdb { args = /etc/dovecot/mysql/dovecot-sql.conf driver = sql } plugin { last_login_dict = proxy::lastlogin last_login_key = last-login/%u mail_log_events = delete undelete expunge copy mailbox_delete mailbox_rename mail_log_fields = uid box msgid size quota = dict:user::proxy::quotadict quota_rule2 = Trash:ignore quota_rule3 = Spam:ignore quota_status_nouser = DUNNO quota_status_overquota = 552 5.2.2 The email account that you tried to reach is over quota quota_status_success = DUNNO quota_warning = storage=95%% quota-warning 95 %u quota_warning2 = storage=80%% quota-warning 80 %u sieve = ~/.dovecot.sieve sieve_before = /etc/dovecot/sieve/default.sieve sieve_dir = ~/sieve stats_refresh = 30 secs stats_track_cmds = yes trash = /etc/dovecot/dovecot-trash.conf zlib_save = gz zlib_save_level = 6 } protocols = imap sieve pop3 lmtp service anvil { client_limit = 20000 unix_listener anvil-auth-penalty { mode = 00 } } service auth { client_limit = 20000 unix_listener /var/spool/postfix/private/auth { group = postfix mode = 0660 user = postfix } unix_listener auth-master { group = mail mode = 0660 user = vmail } user = root vsz_limit = 1 G } service dict { unix_listener dict { mode = 0660 user = vmail } } service imap-login { client_limit = 4096 executable = /usr/lib/dovecot/imap-login inet_listener imap { address = * port = 143 } inet_listener imaps { address = * port = 993 } process_limit = 4096 process_min_avail = 24 service_count = 100 } service imap { executable = /usr/lib/dovecot/imap process_limit = 4096 process_min_avail = 24 vsz_limit = 1 G } service lmtp { process_min_avail = 24 unix_listener /var/spool/postfix/private/dovecot-lmtp { group = postfix mode = 0600 user = postfix } vsz_limit = 1 G } service managesieve-login { client_limit = 4096 inet_listener sieve { port = 4190 } process_limit = 4096 process_min_avail = 24 service_count = 1 } service managesieve { process_limit = 4096 vsz_limit = 1 G } service pop3-login { client_limit = 4096 executable = /usr/lib/dovecot/pop3-login inet_listener pop3 { address = * port = 0 } inet_listener pop3s { address = * port = 995 } process_limit = 4096 process_min_avail = 24 service_count = 100 } service pop3 { executable = /usr/lib/dovecot/pop3 process_limit = 4096 process_min_avail = 24 vsz_limit = 1 G } service quota-status { client_limit = 4096 executable = /usr/lib/dovecot/quota-status -p postfix process_min_avail = 24 unix_listener /var/spool/postfix/private/quota-status { group = postfix mode = 0660 user = postfix } } service quota-warning { executable = script /usr/local/bin/quota-warning.sh unix_listener quota-warning { mode = 0660 user = vmail } user = vmail } userdb { args = /etc/dovecot/mysql/dovecot-sql.conf driver = sql } protocol imap { imap_idle_notify_interval = 4 mins imap_max_line_length = 64 k mail_max_userip_connections = 16 mail_plugins = zlib quota mail_log notify trash imap_quota } protocol pop3 { mail_max_userip_connections = 16 mail_plugins = zlib quota mail_log notify pop3_client_workarounds = outlook-no-nuls oe-ns-eoh pop3_fast_size_lookups = yes pop3_no_flag_updates = yes pop3_uidl_format = %08Xu%08Xv } protocol lda { auth_socket_path = /var/run/dovecot/auth-master mail_plugins = zlib quota mail_log notify sieve sendmail_path = /usr/lib/sendmail syslog_facility = mail } protocol sieve { mail_max_userip_connections = 16 managesieve_max_compile_errors = 5 } protocol lmtp { mail_plugins = zlib quota mail_log notify sieve syslog_facility = mail }

4 8

failed: Invalid cross-device link
by KES 15 Nov '18

15 Nov '18

Hi. I mount ~/Maildir/new via s3fs s3fs mybucket /home/kes/Maildir/new -f -o use_cache=/tmp/S3 -o noatime -o endpoint=eu-west-1 -o profile=mail -o allow_other -o uid=1000,gid=1000 I configure dovecot: maildir_copy_with_hardlinks = no Nov 15 19:30:48 work dovecot: imap(kes): Error: rename(/home/kes/Maildir/new/ddd3, /home/kes/Maildir/cur/ddd3:2,) failed: Invalid cross-device link Tell me please, why I still get this error and how to fix it?

2 1

dovecot 2.2/openssl 1.0 vs dovecot 2.3/openssl 1.1.1 ssl regression
by Arkadiusz Miśkiewicz 15 Nov '18

15 Nov '18

Hi. I'm considering dovecot migration from 2.2.36 run with openssl 1.0.2o to dovecot 2.3.3 run with openssl 1.1.1. Currently I have both variants running with identical configs and certs (the only differences are due to config syntax changes in dovecot 2.3), so for example on both I have: ssl_ca = </etc/openssl/certs/wildcard_ca.pem (this file contains single intermediate certificate of my CA) ssl_cert = </etc/openssl/certs/wildcard_crt.pem (this contains single cerificate for my *.example.com domain) ssl_key = # hidden, use -P to show it (and one key) No alt certs in use. Chain is: - CA trusted by clients (this certificate isn't provided by my dovecot, it's not needed) - wildcard_ca.pem - intermediate CA - wildcard_crt.pem - wildcard certificate for my *.example.com domain dovecot 2.2.36 behaviour is to provide wildcard_ca.pem and wildcard_crt.pem to the client - that behaviour is OK. Client has full trust chain. dovecot 2.3.3 provides only wildcard_crt.pem certificate to the client which is a big problem because missing wildcard_ca.pem (intermediate certificate) breaks chain and client is not able to verify trust chain. Testing is done with simple: openssl s_client -connect my.example.com:143 -starttls imap -servername my.example.com -showcerts 2.3.x announcements and upgrade wiki mention no such behaviour change, so I assume it is a regression. Now doing cat wildcard_ca.pem >> wildcard_crt.pem solves the problem and dovecot starts providing both certs to clients but if that's the proper way of solving this issue then what's the point of having ssl_ca config setting? Ideas? -- Arkadiusz Miśkiewicz, arekm / ( maven.pl | pld-linux.org )

3 3

FETCH BODY vs FETCH BODYSTRUCTURE + FETCH BODY.PEEK
by Alessio Cecchi 15 Nov '18

15 Nov '18

Hello, we are developing a web page, for a ticket system, to show the email messages ordered by the biggest attachments size present in an IMAP folder, with a pagination of 100 messages per page. Each raw represents a message with "DATE, FROM, SUBJECT, the preview (the first 500 chars of the text/html of the messsage) and the list of the attachments (name + size) it contains. The problem is that the page load is VERY slow because we fetch the entire BODY of all 100 messages, in order to get the preview and the attachments name and size. Now, considering that each messages can contain up to 50 MB of attachments, we risk to download 5000MB from IMAP for every page. An option might be to perform two FETCH for each message: - one fetch with just the BODYSTRUCTURE (where we can get the attachments name and size and the text/html parts). - one fetch with the body parts (BODY.PEEK[]) we need to build the preview (text + html). We are wondering if there is a single FETCH command to get the text and html parts (needed to build the preview) with only the name and the size of the attachments without downloading the entire message. Thanks. -- Alessio Cecchi Postmaster @ http://www.qboxmail.it https://www.linkedin.com/in/alessice

2 1