- dovecot - dovecot.org

Slow replication
by Jonas Plitt 10 Feb '15

10 Feb '15

Hi, I'm running two Dovecot 2.2.9 Servers running replication. Users are the same. My Replication is very slow. Mails from Server one appear on Server two after 1-3 hours. Rarely i can see duplicated mails (Log message "Expunged message reappeared. Setting new UID"). This is my doveconf -n output: > # 2.2.9: /etc/dovecot/dovecot.conf > # OS: Linux 3.13.0-44-generic x86_64 Ubuntu 14.04.1 LTS > auth_cache_size = 10 M > auth_mechanisms = plain login > debug_log_path = /var/log/dovecot-debug.log > doveadm_password = ABC012456789ABC > doveadm_port = 12345 > imap_idle_notify_interval = 30 mins > info_log_path = /var/log/dovecot-info.log > listen = *,[::] > log_timestamp = "%Y-%m-%d %H:%M:%S " > login_greeting = Dovecot (Ubuntu) ready. > mail_fsync = always > mail_gid = vmail > mail_location = mbox:~/mail:INBOX=/var/mail/%u > mail_plugins = " notify replication zlib mail_log" > mail_privileged_group = vmail > mail_uid = vmail > mailbox_idle_check_interval = 1 mins > mbox_lazy_writes = no > namespace inbox { > inbox = yes > location = > prefix = > } > passdb { > args = /etc/dovecot/dovecot-sql.conf > driver = sql > } > passdb { > args = /etc/dovecot/dovecot-sql.conf > driver = sql > } > plugin { > mail_log_events = delete undelete expunge copy mailbox_delete > mailbox_rename > mail_log_fields = uid box msgid size > mail_replica = tcp:server2.mydomain.de:12345 > quota = dict:user::file:/var/vmail/%d/%n/.quotausage > replication_full_sync_interval = 1 hour > sieve = /var/vmail/%d/%n/.sieve > zlib_save = gz > zlib_save_level = 9 > } > postmaster_address = postmaster(a)mydomain.tld > protocols = imap pop3 > replication_max_conns = 30 > service aggregator { > fifo_listener replication-notify-fifo { > mode = 0666 > user = vmail > } > unix_listener replication-notify { > mode = 0666 > user = vmail > } > } > service auth { > unix_listener /var/spool/postfix/private/auth { > group = postfix > mode = 0660 > user = postfix > } > unix_listener auth-userdb { > group = vmail > mode = 0600 > user = vmail > } > user = root > } > service config { > unix_listener config { > user = vmail > } > } > service doveadm { > inet_listener { > port = 12345 > ssl = no > } > user = vmail > } > service imap-login { > client_limit = 1000 > inet_listener imap { > port = 143 > } > inet_listener imaps { > port = 993 > ssl = yes > } > process_limit = 500 > process_min_avail = 5 > service_count = 1 > } > service lmtp { > executable = lmtp -L > process_min_avail = 5 > unix_listener lmtp { > group = postfix > mode = 0666 > user = postfix > } > user = vmail > } > service pop3-login { > inet_listener pop3 { > port = 110 > } > inet_listener pop3s { > port = 995 > ssl = yes > } > } > service replicator { > process_min_avail = 1 > unix_listener replicator-doveadm { > group = vmail > mode = 0666 > user = vmail > } > unix_listener replicator { > group = vmail > mode = 0666 > user = vmail > } > } > ssl_ca = </etc/ssl/certs/startcom-bundle.pem > ssl_cert = </etc/ssl/certs/server2.pem > ssl_cipher_list = > EECDH+ECDSA+AESGCM:EECDH+aRSA+AESGCM:EECDH:EDH:HIGH:+RSA:+SHA: > MEDIUM:+RC4:RC4:!aNULL:!MD5:!eNULL:!LOW:!EXP:!DSS:!PSK:!SRP > ssl_dh_parameters_length = 2048 > ssl_key = </etc/ssl/private/server2.key > ssl_prefer_server_ciphers = yes > ssl_protocols = !SSLv2 !SSLv3 > ssl_require_crl = no > userdb { > args = /etc/dovecot/dovecot-sql.conf > driver = sql > } > userdb { > args = /etc/dovecot/dovecot-sql.conf > driver = sql > } > protocol imap { > imap_client_workarounds = delay-newmail > mail_max_userip_connections = 15 > mail_plugins = " notify replication zlib mail_log quota imap_quota > imap_zlib" > ssl_ca = </etc/ssl/certs/startcom-bundle.pem > ssl_cert = </etc/ssl/certs/imap-domain-de.crt > ssl_key = </etc/ssl/private/imap-domain-de.key > } > protocol pop3 { > mail_max_userip_connections = 10 > mail_plugins = " notify replication zlib mail_log quota" > pop3_uidl_format = %08Xu%08Xv > ssl_ca = </etc/ssl/certs/startcom-bundle.pem > ssl_cert = </etc/ssl/certs/server1.pem > ssl_key = </etc/ssl/private/server1.key > } > protocol lda { > info_log_path = /var/log/dovecot-lda.log > mail_plugins = " notify replication zlib mail_log" > } > Aditionally, I forced a "doveadm sync -f" using crontab twice a day. Can anybody give me a hint, please? -- regards, Jonas

1 0

Re: Uniqueness of dovecot mailbox-guids
by Christian Binder Stadt Freilassing 10 Feb '15

10 Feb '15

Ok, problem solved. Obviously the GUIDs only have to be unique for a specific user account in dovecot. Many thanks to Heinlein Support who helped me with this question!

1 0

Re: Uniqueness of dovecot mailbox-guids
by Christian Binder Stadt Freilassing 10 Feb '15

10 Feb '15

Bump. Can no one give a hint? Thanks a lot! Christian

1 0

Geting mail quota exceeded with plenty of space
by Robert Moskowitz 10 Feb '15

10 Feb '15

I have a user that is getting mail quota exceeded: Feb 9 15:00:21 z9m9z dovecot: lda(dm(a)htt-consult.com): Error: sieve: msgid=<38308773.1704736628308773ywdm(a)htt-consult.com853430>: failed to store into mailbox 'INBOX': Quota exceeded (mailbox for user is full) Yet the quota is set for 1000Mb and the current reported use is 277Mb. There are only 28 messages in the in box and 842 in the spam. Where might the problem becoming from? The server is Redsleeve 6 (Centos6 on arm). I am running a combination of postfix/dovecot/roundcubemail with a mysql virtual domain database maintained with postfix.admin

2 10

Mails Not Deleting in first POP3 Session
by Smith, Michael H. (GSFC-760.0)[NICS] 10 Feb '15

10 Feb '15

The dovecot -n output is attached to this email. Filename dovecot-n_filtered.conf. Dovecot Version: 2.2.10 OS: CentOS 6.6 MTA: Postfix version 2.6.6 The system is using dsync between two mailbox servers. Problem: When a pop3 session deletes a new message and the pop3 session ends, the message is not deleted. On a subsequent pop3 session, the same messages is deleted again and when the session ends, this time the messages is removed. This results in automated scripts getting duplicate messages. Previous mail server utilized postfix and dbmail. The problem did not exist. All of our replication settings are located in the file /etc/dovecot/conf.d/12-replicator.conf. When I move the above file out of /etc/dovecot/conf.d/ on both mailbox servers and restart dovecot, the pop3 deletion problem goes away and the system behaves as expected i.e. when a messages is deleted, it's removed when the session ends (of course its not replicated to the secondary server). Numerous tests using IMAP and deleted messages behaved as expected.with or without replication Here is some additional information that I hope helps: Below I'm running watch 'pwd; ls -l; echo "../cur"; ls -l ../cur | tail -2' On the primary mailbox server: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur | t... Mon Feb 9 22:42:04 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1.nasc om.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, On the backup mailbox server: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur... Mon Feb 9 22:44:12 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1. nascom.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, After I connect to the primary mailbox server using openssl s_client -connect mail1.example.com:995 tls1_2 user test_u pass 123456 output below from the same watch command from above: Primary Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur | t... Mon Feb 9 22:47:50 2015 /home/vmail/nascom.nasa.gov/test_u/new total 0 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.nasco m.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1.nasc om.nasa.gov,S=46223,W=46902:2, Secondary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur... Mon Feb 9 22:48:14 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1. nascom.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, So far, on the primary, the new message is moved to the cur directory while on the backup the new messages is still in new. I perform a list command to get the message I want to delete. In this case the message number is 2402 so it looks like this... list . . . 2401 2402 dele 2402 +OK Marked to be deleted quit Here's the primary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur | t... Mon Feb 9 22:53:16 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1.nasc om.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, Here's the secondary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur... Mon Feb 9 22:53:38 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423522362.M879542P21023.mbox-open-2. nascom.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, See we see the message on the primary is moved back to new while the message on the secindary has a name change. Now for the second session (again using openssl): +OK Dovecot ready. user test_u +OK pass --------------- +OK Logged in. Here's the output on the primary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur | t... Mon Feb 9 22:56:27 2015 /home/vmail/nascom.nasa.gov/test_u/new total 0 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.nasco m.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 46223 Feb 9 22:41 1423521693.M193283P20903.mbox-open-1.nasc om.nasa.gov,S=46223,W=46902:2, Here's the output on the secondary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur... Mon Feb 9 22:57:22 2015 /home/vmail/nascom.nasa.gov/test_u/new total 48 -rw------- 1 vmail mail 46223 Feb 9 22:41 1423522362.M879542P21023.mbox-open-2. nascom.nasa.gov,S=46223,W=46902 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, Now a list and delete... list . . . 2401 2402 dele 2402 +OK Marked to be deleted quit Here's the output of the primary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur | t... Mon Feb 9 22:59:17 2015 /home/vmail/nascom.nasa.gov/test_u/new total 0 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.nascom .nasa.gov,S=3269,W=3980:2, Now the scondary: Every 2.0s: pwd; ls -l; echo "../cur"; ls -l ../cur... Mon Feb 9 22:59:37 2015 /home/vmail/nascom.nasa.gov/test_u/new total 0 ../cur -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191796.M570666P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, -rw------- 1 vmail mail 3269 Nov 28 16:23 1417191797.M813970P7291.mbox-open-1.na scom.nasa.gov,S=3269,W=3980:2, The message is deleted now. Multiple new messages that are all deleted in the same session behave the same way.. However, if the new messages aren't deleted in the first session, they move to the cur directory. On a subsequent session, the messages can be deleted and they are removed. The new messages remain in the new directory after the first session (as previously stated) on the secondary mailbox server until the dele. So second session deletes appear to work fine Any assistance would be greatly appreciated. --Mike

1 0

Postfix , Dovecot & the Spam fight
by Leander Schäfer 09 Feb '15

09 Feb '15

Hi, I'm currently busy with a substiution of my current mail server. I'm currently using * Clam-SMTP and * SpamAssassin to fight Spam. I wonder if it is worth implementing AmaViS with SpamAssassin backend instead and also using AmaViS to speak to clamd directly. But I more and more wonder wether AmaViS is even worth it?! It currently looks to me as if AmaViS is eating LOTS of ressources and it is very uncomfortable for automated installations if you have to do dynamic batch changes on the AmaViS configs - sed(1) is your friend but this hectic escaping and workarrounds is really not sustainable to maintain. So my question is: Does AmaViS have any advantages compared to the current setup? I don't seem to find lots of qualified discussions for this on the net. The AmaViS related articles I found are freaking old. Would be nice the get your best practice as a change. Thanks a lot Best regards Leander

3 3

ssl_cipher_list
by LuKreme 08 Feb '15

08 Feb '15

How do I get a list of the possible ciphers that are installed on the system for use in ssl_cipher_list? -- They all have husbands and wives and children and houses and dogs, and you know, they've all made themselves a part of something and they can talk about what they do. What am I gonna say? "I killed the president of Paraguay with a fork. How've you been?"

2 1

TLS config check
by SW 07 Feb '15

07 Feb '15

Hi All First the essentials: dovecot --version: 2.2.15 /usr/local/etc/dovecot/conf.d/10-ssl.conf: ssl = required ssl_cert = </usr/local/openssl/certs/mail.domain.com.chained.dovecot.ecdsa.crt ssl_key = </usr/local/openssl/certs/mail.domain.com.ecdsa.key ssl_protocols = !SSLv2 !SSLv3 ssl_cipher_list = HIGH:EECDH+ECDSA+AESGCM:EECDH+aRSA+AESGCM:EECDH+ECDSA+SHA384:EECDH+ECDSA+SHA256:EECDH+aRSA+SHA384:EECDH+aRSA+SHA256:EECDH:EDH+aRSA:ECDHE-RSA-AES256-SHA:+DHE-RSA-AES256-SHA:!AES256-SHA256:!AES256-GCM-SHA384:!CAMELLIA256-SHA:!AES128:!CAMELLIA128:!aNULL:!eNULL:!LOW:!3DES:!MD5:!EXP:!PSK:!SRP:!DSS:!RC4:!SEED:+AES256-SHA ssl_prefer_server_ciphers = yes I would really appreciate it if someone could tell me if my config is super secure? I run the following email clients: K9 on Android 4.4.2 Thunderbird 31.4 Outlook 2010 I'm interested to know if the config I have is secure and that my cipher list is acceptable. I'm also keen to hear thoughts on my config in respect of Forward Secrecy and the SSLv3/POODLE attack. Thanks!

5 9

Re: Interface for managing users?
by Thomas Moulia 06 Feb '15

06 Feb '15

Hello, > It's actually the way around: pick any user management and configure > Dovecot to use its database. > Depending on what you want, you need a system, that also removes the > user's home dirs and/or creates them from a skeleton. Makes sense, and is nicely simple. > After doing a change in the user database, make sure to run > doveadm auth cache flush > to force Dovecot to pick up the changes, before you fire up the next test. Ah, good to know -- that would've bit me. Thanks for the help, Steffen! -Thomas

1 0

doveadm quota get doesn't recalculate quota (Maildir backend)
by Teodor Milkov 06 Feb '15

06 Feb '15

Hello, I've noticed that "doveadm quota get" doesn't recalculate quota if maildirsize file is missing for some reason. It just says quota is unknown, usage is 0. In contrast, vpopmail does regenerate maildirsize if it is missing. Reading the Maildir++ specification there is this part: ...If maildirsize does not exist, or if its size is at least 5120 bytes, recalculate it using the procedure defined above, and use the recalculated numbers... One seemingly legitimate reason for maildirsize to be missing is when vpopmail mailbox is set to NOQUOTA (i.e. no limits). My workaround to this is to use a fallback quota: quota_rule = ?:messages=10000000 That is, if there's no backend limit detected (missing maildirsize) then fallback to 10 mln message count quota. In this case doveadm quota get _does_ recalculate usage and even creates maildirsize file. So, my questions are: 1. Isn't it still better to calculate and show usage even if maildirsize is missing? Because calling quota recalc and then quota get is racy - someone else may change configuration or usage etc. between our two calls (forks). 2. If there's good reason against implicit recalculation, then maybe quota get output should be more specific that neither quota nor usage are known? Instead of usage=0 show -1 or "-" or "n/a" or "unkown"?

1 0