- dovecot - dovecot.org

[Dovecot] Dovecot reports zero messages erroneously
by Chris Spiegel 29 Oct '05

29 Oct '05

On October 11, Mark Farmer reported a problem about not being able to see new mails. I'm running into the same issue. Even though cur/ and new/ contain messages, Dovecot reports that there aren't any. This problem seemed to start happening in alpha3 (and persists in alpha4). When I revert back to alpha2, the messages show up again. It looks like a problem with the dovecot.index* files. When I delete them, alpha4 can serve up the mails just fine, and will in fact keep working for a little while. Eventually (within a few hours, although I'm not sure what causes it), dovecot thinks things are empty again. Like Mark, I'm using Postfix.. I'm not sure if that has anything to do with the problem but I thought it'd be worth pointing out. If there are any specific things you'd like me to do to try to pin down this problem, let me know. Chris

2 4

[Dovecot] Some errors with 1.0.alpha4
by Nicolas STRANSKY 29 Oct '05

29 Oct '05

Hi, I'm getting three types of errors with latest Dovecot compiled for x86_64 GNU/Linux Debian. Oct 29 12:47:57 rouge imap-login: unable to dlopen /usr/lib/sasl2/libntlm.so.2: /usr/lib/sasl2/libntlm.so.2: failed to map segment from shared object: Cannot allocate memory Oct 29 12:47:57 rouge pop3-login: unable to dlopen /usr/lib/sasl2/libntlm.so.2: /usr/lib/sasl2/libntlm.so.2: failed to map segment from shared object: Cannot allocate memory I have no idea why there is this error, but it comes each time a user logs in, with no other severe effect. Oct 29 01:28:54 rouge dovecot: imap(glondu): pool_system_malloc(): Out of memory Oct 29 01:28:54 rouge dovecot: child 5642 (imap) returned error 83 (Out of memory) Oct 29 01:38:10 rouge dovecot: imap(glondu): pool_system_malloc(): Out of memory Oct 29 01:38:10 rouge dovecot: child 5639 (imap) returned error 83 (Out of memory) Oct 29 01:49:22 rouge dovecot: imap(glondu): pool_system_malloc(): Out of memory Oct 29 01:49:22 rouge dovecot: child 8839 (imap) returned error 83 (Out of memory) The mailbox for this user is not bigger that 10MB and dovecot mail_process_size is set to default setting so I don't see why there could be a "Out of memory" error. Oct 28 18:00:45 rouge dovecot: imap(vidal): Lost transaction log file /var/dovecot-indexes/vidal/.INBOX/dovecot.index.log seq 1 Oct 28 18:00:45 rouge dovecot: imap(vidal): Lost transaction log file /var/dovecot-indexes/vidal/.INBOX/dovecot.index.log seq 1 Oct 28 18:00:56 rouge dovecot: imap(vidal): Lost transaction log file /var/dovecot-indexes/vidal/.INBOX/dovecot.index.log seq 1 Oct 28 18:00:56 rouge dovecot: imap(vidal): Lost transaction log file /var/dovecot-indexes/vidal/.INBOX/dovecot.index.log seq 1 This error caused the user's mailbox to be unreadable, and erasing the indexes resolved the problem, but I would have expected Dovecot to have done it itself if the indexes were corrupted ? Hope this feedback helps, -- Nico

1 0

[Dovecot] Basic Configuration
by Ashvin Meetoo 29 Oct '05

29 Oct '05

Hi, Been setting up a backup mail server just in case... using FC4... basic configuration... postfix/dovecot... imap pop3... When I send an external email, example(a)gmail.com, everything works well. When I send an internal email, i.e. to another user, user receives a blank email? Everytime he clicks on get mail a new empty email is added... checked /<user>/Mailbox/new folder email is there and header is ok. What can be the matter? Did this thousand times before and even reinstalled FC4... same problem. Also, I have tried: echo test | mail root but mail gives me : No mail for root Help please. __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com

1 0

[Dovecot] dovecot-lda cvs
by Rene Luria 28 Oct '05

28 Oct '05

I am using dovecot-lda from cvs (checked out on september the 21st) and dovecot 1.0alpha2 (not yet upgraded ok). I have a strange feeling about "redirect" feature. Let's see two sieve scripts examples: # ex1 require ["fileinto", "vacation"]; if header :is ["blabla"] "blabla" { fileinto "blabla"; } redirect "user(a)domain.com"; keep; # ex2 require ["fileinto", "vacation"]; redirect "user(a)domain.com"; keep; In both cases, I redirect mail to an external address, keeping a copy. (this would emulate redirect :copy "user(a)domain.com"; which is still not implemented). In first example I run a test on headers, but not on the second. The effect is that in the second example, the forwarded mail is "complete" (exactly the original one with routing header modified), but in the first one, the forwared mail is completely messed up: no Subject header, and so on. This could be because as tests are run on the incoming mail, the working in memory of the mail is changed, and then, forwarded. So when no tests are run, the mail remains intact. How can I deal with this ? Is there some nice hotfix that would prevent a few (like 100'000) users to kill me ? Best regards. -- Rene Luria

3 5

[Dovecot] 1.0alpha3: fcntl() complaints?
by Jeff A. Earickson 28 Oct '05

28 Oct '05

Hi, Since installing 1.0alpha3 this morning, I've seen this syslog complaint a few times. Anything to worry about? A "server too busy" issue? imap(user): fcntl() failed with mbox file /var/mail/user: Resource temporarily unavailable Jeff Earickson Colby College

2 4

[Dovecot] DoveCot 1.0a3 hangs at login
by Justin Finkelstein 28 Oct '05

28 Oct '05

I've just upgraded to 1.0a3 via a freebsd port, and I've begun to notice odd behaviour from Dovecot. After converting my config file settings over to the new format, everything runs fine for an hour or so, then dovecot stops responding: oot@toejamfootball# telnet localhost 143 Trying ::1... telnet: connect to address ::1: Connection refused Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK Dovecot ready. 1 login justin(a)redwiredesign.com password [dovecot then just stops doing anything] * BYE Disconnected for inactivity. Connection closed by foreign host. in the logs, it says: Oct 13 09:43:22 toejamfootball dovecot: imap-login: Disconnected: Inactivity: method=PLAIN, rip=192.168.2.11, lip=192.168.2.43 Oct 13 09:43:22 toejamfootball dovecot: auth(default): client in: AUTH 1 PLAIN service=IMAP lip=192.168.2.43 rip=192.168.2.11 resp=AGp1c3RpbkByZWR3aXJlZGVzaWduLmNvLnVrAHBhc3N3b3Jk Oct 13 09:43:22 toejamfootball dovecot: auth-worker(default): sql(x,192.168.2.11): query: SELECT username AS user, MD5(clear) AS password FROM users LEFT JOIN domains ON domains.domain_id = users.domain_id WHERE domain="x" AND username="x" AND users.enabled = 1 AND domains.enabled = 1 Oct 13 09:43:22 toejamfootball dovecot: auth(default): pam(justin(a)x,192.168.2.11): pam_authenticate() failed: authentication error Any clues? I don't use pam for auth anyway, so I just ignore that. When I restart the server and login again it's fine. j.

2 1

[Dovecot] 1.0.alpha3 plaintext authentication problem
by Sergey Redin 28 Oct '05

28 Oct '05

Hello everyone, Since the upgrade to 1.0.alpha3 I experience a strange problem with plaintext authentication. I successfully login first few (2-5) times after the dovecot restart, and then on a reason I cannot login any more (both via POP3 and IMAP). I tried to find out what could be the reason for this but I failed. Nothing is changed inside the user's home directory or /var/run/dovecot, but I cannot login any more. I guess there is something wrong inside dovecot-auth od dovecot process itself. I just start to receive "01 NO Authentication failed." or "-ERR Authentication failed." from dovecot and "Authenticate PLAIN failed: Unsupported authentication mechanism.: method=PLAIN" in my logs, even though I used the same PLAIN two minutes ago without a problem. Here is my dovecot.conf: nevoservice# egrep -v '^[ ]*(\#.*)?$' /usr/local/etc/dovecot.conf protocols = imap pop3 ssl_disable = no ssl_cert_file = /etc/mail/certs/server.crt ssl_key_file = /etc/mail/certs/server.key disable_plaintext_auth = no log_path = /var/log/dovecot/log info_log_path = /var/log/dovecot/info login_user = dovecot login_processes_count = 1 verbose_proctitle = no first_valid_gid = 0 mail_extra_groups = mail mail_debug=yes default_mail_env = mbox:/var/mail/%u maildir_copy_with_hardlinks = yes protocol imap { imap_client_workarounds = delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep } protocol pop3 { pop3_uidl_format = %08Xu%08Xv pop3_client_workarounds = outlook-no-nuls oe-ns-eoh } auth_verbose = yes auth_debug = yes auth default { mechanisms = plain passdb passwd-file { args = /usr/local/etc/dovecot.passwd } userdb passwd-file { args = /usr/local/etc/dovecot.passwd } user = root } Here are two lines in dovecot.passwd which define the user that I use: wail:password-hash:1050:1032:Administrator wail.spb.ru:/home/wail.spb.ru:::/var/mail/wail wail@nevoservice.ispvds.com:password-hash:1050:1032:Administrator wail.spb.ru:/home/wail.spb.ru:::/var/mail/wail A few lines from my /var/log/dovecot/info file: dovecot: Oct 10 02:03:11 Info: auth(default): client in: AUTH 1 PLAIN service=POP3 lip=82.146.50.34 rip=82.146.63.23 resp=AHdhaWwAVGxNSFRuUlI2 dovecot: Oct 10 02:03:11 Info: auth(default): client out: OK 1 user=wail dovecot: Oct 10 02:03:11 Info: auth(default): master in: REQUEST 3 53919 1 dovecot: Oct 10 02:03:11 Info: auth(default): master out: USER 3 wail uid=1050 gid=1032 home=/home/wail.spb.ru mail=/var/mail/wail dovecot: Oct 10 02:03:11 Info: pop3(wail): Effective uid=1050, gid=1032 dovecot: Oct 10 02:03:11 Info: pop3(wail): maildir autodetect: stat(/var/mail/wail/cur) failed: Not a directory dovecot: Oct 10 02:03:11 Info: pop3(wail): mbox autodetect: data=/var/mail/wail dovecot: Oct 10 02:03:11 Info: pop3(wail): mbox autodetect: INBOX file: yes (/var/mail/wail) dovecot: Oct 10 02:03:11 Info: pop3(wail): mbox: data=/var/mail/wail dovecot: Oct 10 02:03:11 Info: pop3(wail): mbox: root exists (/home/wail.spb.ru/mail) dovecot: Oct 10 02:03:11 Info: pop3(wail): mbox: root=/home/wail.spb.ru/mail, index=/home/wail.spb.ru/mail, inbox=/var/mail/wail dovecot: Oct 10 02:03:11 Info: pop3-login: Login: user=<wail>, method=PLAIN, rip=82.146.63.23, lip=82.146.50.34 dovecot: Oct 10 02:03:13 Info: pop3(wail): Logout. top=0/0, retr=0/ del=0/3, size=5372 dovecot: Oct 10 02:03:20 Info: auth(default): client in: AUTH 1 PLAIN service=POP3 lip=82.146.50.34 rip=82.146.63.23 resp=AHdhaWwAVGxNSFRuUlI2 dovecot: Oct 10 02:03:20 Info: auth(default): client out: OK 1 user=wail dovecot: Oct 10 02:03:20 Info: auth(default): master in: REQUEST 4 53918 1 dovecot: Oct 10 02:03:20 Info: auth(default): master out: USER 4 wail uid=1050 gid=1032 home=/home/wail.spb.ru mail=/var/mail/wail dovecot: Oct 10 02:03:20 Info: pop3(wail): Effective uid=1050, gid=1032 dovecot: Oct 10 02:03:20 Info: pop3(wail): maildir autodetect: stat(/var/mail/wail/cur) failed: Not a directory dovecot: Oct 10 02:03:20 Info: pop3(wail): mbox autodetect: data=/var/mail/wail dovecot: Oct 10 02:03:20 Info: pop3(wail): mbox autodetect: INBOX file: yes (/var/mail/wail) dovecot: Oct 10 02:03:20 Info: pop3(wail): mbox: data=/var/mail/wail dovecot: Oct 10 02:03:20 Info: pop3(wail): mbox: root exists (/home/wail.spb.ru/mail) dovecot: Oct 10 02:03:20 Info: pop3(wail): mbox: root=/home/wail.spb.ru/mail, index=/home/wail.spb.ru/mail, inbox=/var/mail/wail dovecot: Oct 10 02:03:20 Info: pop3-login: Login: user=<wail>, method=PLAIN, rip=82.146.63.23, lip=82.146.50.34 dovecot: Oct 10 02:03:21 Info: pop3(wail): Logout. top=0/0, retr=0/ del=0/3, size=5372 dovecot: Oct 10 02:03:28 Info: pop3-login: Authenticate PLAIN failed: Unsupported authentication mechanism.: method=PLAIN, rip=82.146.63.23, lip=82.146.50.34 dovecot: Oct 10 02:03:30 Info: pop3-login: Aborted login: method=PLAIN, rip=82.146.63.23, lip=82.146.50.34 We see here two successfull logins with PLAIN and then suddenly PLAIN becomes unsupported. I run dovecot 1.0.alpha3 on FreeBSD 4.11, it was installed from ports and then I edited dovecot.conf (and created dovecot.passwd). Any help would be really appreciated. Sergey Redin.

2 1

Re: [Dovecot] Dovecot Alpha 4
by Tomi Hakala 28 Oct '05

28 Oct '05

Marcus Don wrote: > How confident are you of the CVS? Any idea when the next (perhaps beta?) > release will be? I run CVS code most of the time, but only with handful of users and with them it has been working OK. I wouldn't use CVS code in production without thorough stress testing in a lab first. >>Have you thought of clustered file system to go with your Dovecot >>hosts and CLARiiON device? > I'm not sure why this would benefit us at the moment. Before we switched to > maildirs, the bottleneck we were experiencing at busy times was the ethernet > speed, not the storage device speed. > Is there another reason you suggested this? No, I'm just not very fond of NFS myself but it seems that for you it is doing job pretty well. Your CLARiiON is a normal SAN device and you have some singe host providing NFS? It didn't quite open to me how your mail spool is set up, is it some meta LUN distributed over two disk arrays or do you just have one huge LUN and mirror it between two arrays? -- Tomi Hakala

1 0

[Dovecot] problem with thunderbird
by Alfonso Sparano 28 Oct '05

28 Oct '05

Hi all, I've a problem with thunderbird 1.0.6/7/1.5 and dovecot pop3 I've tried dovecot alpha 3 and 1.0 stable but neither one is capable of downloading mails. Nothing is written into the log. I've tried also to delete dovecot.index* and dovecot-uidlist into the Maildir directory. The cur directory contains more than 8100 messages (single files) Anyone has founded this problem? Alfonso

3 3

[Dovecot] Corrupted index cache file [...] invalid record size
by Lorens 28 Oct '05

28 Oct '05

I get this error every single time some of my users log in: Error: POP3(${user}): Corrupted index cache file /${userdir}/dovecot.index.cache: invalid record size The named file seems to be deleted as soon as corruption is found. This only affects some users. I thought the affected users had in common that they are using POP3 in "leave-messages-on-server" mode, and that their mailboxes are big ( > 100 MB). However I've found a few counterexamples, few and far between (not every login) but still. Google's no help, for once :-) Could this be a locking problem, a bug with big Maildirs (counterexamples being few), or what? I know some of my users have had problems logging several IMAP clients into the smae account at the same time, wouldn't put it past some of those with problems to be trying the same with POP. I know that a detected cache corruption is merely that, but I find the error message annoying :-) Configuration: v1.0-stable (0.99.20050712-2 from debian unstable) Maildir only Default locking (nothing mentioned in conf file) Access over NFS mmap_disable = yes (wiki says to disable mmap to use indexes on NFS). Thanks for any pointers.

2 1