- dovecot - dovecot.org

[Dovecot] Dovecot and stale nfs-locks hanging processes
by Søren Schrøder 05 Apr '08

05 Apr '08

Greetings dovecot mailing list. I have implemented a relatively big dovecot setup (250k users) and overall I am very pleased with dovecot functionality and performance. Setup description: * dovecot 1.0.x * FreeBSD 6.3 * Postfix (using dovecot deliver as LDA). * OpenLdap backend * Storage is NFS (Clariion EMC NFSd for Maildir, and FreeBSD NFSd for Indexes). * Locking is fcntl using RPC.lockd. * Users are accessing mail using POP3 and IMAP (IMAP mainly via Squirrelmail, but also direct) * 3 frontends for POP/SMTP and 2 frontends for IMAP (webmail). Round Robin DNS My problem: I am having issues where POP3, IMAP and DELIVER processes gets stuck, apparently waiting for device. fstat shows: bash# fstat -p 93522 USER CMD PID FD MOUNT INUM MODE SZ|DV R/W 302870 pop3 93522 root / 2 drwxr-xr-x 512 r 302870 pop3 93522 wd /home/mnt5 51592 drwxr-xr-x 80 r 302870 pop3 93522 text /usr 121619 -r-xr-xr-x 436616 r 302870 pop3 93522 0* internet stream tcp 302870 pop3 93522 1* internet stream tcp 302870 pop3 93522 2* pipe c778aa48 <-> c778a990 0 rw 302870 pop3 93522 3 /dev 24 crw-rw-rw- random r 302870 pop3 93522 5* pipe ce440b28 <-> ce440be0 0 rw 302870 pop3 93522 6* pipe ce440be0 <-> ce440b28 0 rw 302870 pop3 93522 7 /home/mnt5 9010290 -rw------- 1493 rw 302870 pop3 93522 8 - - bad - 302870 pop3 93522 9 - - bad - 302870 pop3 93522 10 - - bad And the inode in question on /home/mnt5 is a dot-nfs file, indicating stale lock: bash# ls -li | grep 9010290 9010290 -rw------- 1 302870 42 1493 Apr 3 18:05 .nfs.0668c236.6d524.4 ktrace on the pid shows absolutely no activity. The pop3 process is un-killable, and I end up stacking up pop3 processes from the user, as well as deliver to the user. Not healthy.. I was under the impression that POP3 would exit when a lock is set, preventing more than one pop3 processes pr. user, but it doesn't seem to be the case. Stopping dovecot entirely, leaves these stale pop3/imap/deliver processes hanging, even with shutdown_clients = yes The windows-problem-solution (reboot) seems to be the only way to get rid of the locked processes. So: Has anyone else observed this behavior, and eventually found the magic cure ? I wonder if there was a way to implement a "max wall-clock time" per dovecot process type (i.e.. terminate process after for example 120 sec. delivery, 600 sec pop3 etc...), as a crude "garbage-collector". Any hints/suggestions is welcome. -- Søren Schrøder

3 4

[Dovecot] LDAP auth_bind hangs and times out
by Jack McKinney 04 Apr '08

04 Apr '08

I have _almost_ got Dovecot working! One little snag... My users login using their email address as username. Each domain has their own LDAP subtree. Each user has an entry in the ou=users subtree of the domain subtree, and has a mail: field (inetOrgPerson) listing their email address/login name. I am trying to use auth_bind: when I login with jackmc(a)lorentz.com, dovecot should search for mail=jackmc(a)lorentz.com in the onelevel below ou=users,dc=lorentz,dc=com and find me as "cn=Jack McKinney,ou=users,dc=lorentz,dc=com". I have created an entry in LDAP (varmail) that should be able to do this query. Indeed, from the command line, it works: ldapsearch -h ldap.lrtz -b 'ou=users, dc=lorentz, dc=com' -D 'cn=varmail,ou=users,dc=lorentz,dc=com' -x -W -s onelevel '(&(objectClass=inetOrgPerson)(mail=jackmc(a)lorentz.com))' Enter LDAP Password: # extended LDIF # # LDAPv3 # base <ou=users, dc=lorentz, dc=com> with scope oneLevel # filter: (&(objectClass=inetOrgPerson)(mail=jackmc(a)lorentz.com)) # requesting: ALL # # Jack McKinney, users, lorentz.com dn: cn=Jack McKinney,ou=users,dc=lorentz,dc=com objectClass: top objectClass: person objectClass: organizationalPerson objectClass: inetOrgPerson cn: Jack McKinney givenName: Jack McKinney sn: McKinney mail: jackmc(a)lorentz.com # search result search: 2 result: 0 Success # numResponses: 2 # numEntries: 1 ' However, it appears that dovecot performs the above query successfully, but then never uses the password and retrieved DN to attempt to bind and authenticate the user. Instead, it just times out. My dovecot-ldap.conf is: hosts = ldap.lrtz dn = cn=varmail,ou=users,dc=lorentz,dc=com dnpass = ********* ldap_version = 3 auth_bind = yes pass_filter = (&(objectClass=inetOrgPerson)(mail=%Lu)) base = ou=users, dc=%Dd scope = onelevel The dovecot log shows: Apr 3 08:13:21 fourier dovecot: auth(default): new auth connection: pid=15774 Apr 3 08:13:30 fourier dovecot: auth(default): client in: AUTH^I1^IPLAIN^Iservice=IMAP^Isecured^Ilip=x.x.x.x^Irip=y.y.y.y^Iresp=<hidden> Apr 3 08:13:30 fourier dovecot: auth(default): ldap(jackmc(a)lorentz.com,y.y.y.y): bind search: base=ou=users, dc=lorentz,dc=com filter=(&(objectClass=inetOrgPerson)(mail=jackmc(a)lorentz.com)) Apr 3 08:16:30 fourier dovecot: imap-login: Disconnected: Inactivity: method=PLAIN, rip=y.y.y.y, lip=x.x.x.x, TLS The OpenLDAP log shows that the query is received and that it returns a match: Apr 3 08:13:30 fourier slapd[14039]: conn=7 op=3 SRCH base="ou=users,dc=lorentz,dc=com" scope=1 deref=0 filter="(&(objectClass=inetOrgPerson)(mail=jackmc(a)lorentz.com))" Apr 3 08:13:30 fourier slapd[14039]: conn=7 op=3 SRCH attr=uid Apr 3 08:13:30 fourier slapd[14039]: conn=7 op=3 SEARCH RESULT tag=101 err=0 nentries=1 text= When I do the query from the command line, there is an additional pair of lines: Apr 3 08:18:10 fourier slapd[16106]: conn=22 op=2 UNBIND Apr 3 08:18:10 fourier slapd[16106]: conn=22 fd=28 closed I thus conclude that Dovecot is not closing the connection. -- Jack McKinney GPG 1024D/99C6A174 jackmc(a)lorentz.com YM:lfaatsnat2006 AIM:jackmclorentz "There is no parameter that makes it impossible for you to perform still more excellently." -Mario Cuomo, on the lack of a clock in baseball

2 11

[Dovecot] "Dumb" proxying?
by Dan Bongert 04 Apr '08

04 Apr '08

I'm a database noob, and it really seems like it would be overkill for my setup: I just want to proxy all connections from my DMZ to my internal mail server -- same internal server for all users. I used to use perdition for this set up, but am having issues getting it to play nicely with my new servers. Can I bypass all of the SQL stuff and just have dovecot forward all POP/IMAP connections to a specific server? Or perhaps set up a wild card lookup in the database (preferably something serverless like sqlite) so that no matter the query, it'll return one server? I don't really want to add "database management" to my list of jobs I have to do, especially for ~1000 users, and when I don't handle user creation/deletion. -- Dan Bongert dbongert(a)ssc.wisc.edu

4 3

[Dovecot] How to force working of a redirect features in sieve plugin?
by Vitalij Satanivskij 04 Apr '08

04 Apr '08

Hello Trying to setup dovecot installation with sieve plugin, can't configure how to make redirect working. My setup 1 Exim + dovecot lda begin transports dovecot_virtual_delivery: driver = pipe command = /usr/local/dovecot/libexec/dovecot/deliver -d $local_part message_prefix = message_suffix = delivery_date_add envelope_to_add return_path_add log_output user = exim 2 Dovecot protocol lda { postmaster_address = postmaster@******* hostname = ********** mail_plugins = cmusieve mail_plugin_dir = /usr/local/dovecot/lib/dovecot/lda # Binary to use for sending mails. sendmail_path = "/usr/sbin/sendmail" # UNIX socket path to master authentication server to find users. auth_socket_path = /var/run/dovecot/auth-master } /usr/sbin/sendmail it's realy exim binary but I.m try full path to real exim - /usr/local/sbi/exim too. That's I.m getting in log files Mar 28 22:23:10 scorpio deliver(satan): Sending a forward to <*****@**********> with return path <<postmaster@******>> Mar 28 22:23:10 scorpio deliver(satan): Sendmail process terminated abnormally, signal 11 Mar 28 22:23:10 scorpio deliver(satan): sieve runtime error: Redirect: Error sending mail and in exim log 2008-03-28 22:23:10 1JfL6U-000LBj-24 <satan@********>: dovecot_virtual_delivery transport output: /libexec/ld-elf.so.1: environment corrupt; mi Question is - how deliver sending mail thru sendmail_path = "/usr/sbin/sendmail"? Via pipe? -- Vitalij Satanivskij

2 2

[Dovecot] outlook2007 shows frequent imap disconnect no matter what outlook-idle setting in dovecot.conf
by Kielbasiewicz, Peter 04 Apr '08

04 Apr '08

Hello, I saw the problem quite often in various posts but could not find a real solution . Problem is that I frequently get annoying popups from outlook 2007 about imap idle timeouts which block working with outlook until the popup is acknowledged. The popup says: Your IMAP server closed the connection. This can occur if you leave the connection idle for too long. Protocol: IMAP Server: testhost Port: 143 Error Code: 0x800CCCDD I tried to run dovecot with or without "imap_client_workarounds = outlook-idle" but do not see any difference. I am running dovecot on Red Hat Enterprise Linux Client release 5.1 (Tikanga). Below is the output of some commands to get a clue about my config. Background info: Our primary mailboxes are on exchange servers. As there is a size limit on the server mailboxes I am evaluating a local storage solution for our users. Thus I set up an additional mailbox in outlook using IMAP. Local .pst files can not be backuped easily (especially when outlook is running) and I am investigating a local IMAP server where users can move their mails to. This server can easily be backuped (even incrementally when using Maildir) and users do always have access to their mails. The drawbacks are that users must have an extra login to the imap server because I have no ldap access to the company AD server from our local site domain. The solution will not be accepted though if I can not prevent the frequent disconnect popups. Peter #dovecot --version 1.0.rc15 #dovecot -n # /etc/dovecot.conf protocols: imap imaps login_dir: /var/run/dovecot/login login_executable: /usr/libexec/dovecot/imap-login mail_location: maildir:~/Maildir maildir_copy_with_hardlinks: yes auth default: passdb: driver: pam userdb: driver: passwd # tail /var/log/messages Mar 26 06:45:28 testhost dovecot: imap-login: Login: user=<peter>, method=PLAIN, rip=::ffff:130.143.81.176, lip=::ffff:130.143.83.92 Mar 26 06:46:01 testhost dovecot: IMAP(peter): Disconnected: Logged out Mar 26 06:46:01 testhost dovecot: IMAP(peter): Disconnected Mar 26 07:15:31 testhost dovecot: imap-login: Login: user=<peter>, method=PLAIN, rip=::ffff:130.143.81.176, lip=::ffff:130.143.83.92 Mar 26 07:15:31 testhost dovecot: imap-login: Login: user=<peter>, method=PLAIN, rip=::ffff:130.143.81.176, lip=::ffff:130.143.83.92 Mar 26 07:16:04 testhost dovecot: IMAP(peter): Disconnected Mar 26 07:16:04 testhost dovecot: IMAP(peter): Disconnected Mar 26 07:40:13 testhost dovecot: imap-login: Login: user=<peter>, method=PLAIN, rip=::ffff:130.143.81.176, lip=::ffff:130.143.83.92 Mar 26 07:40:13 testhost dovecot: imap-login: Login: user=<peter>, method=PLAIN, rip=::ffff:130.143.81.176, lip=::ffff:130.143.83.92 Mar 26 07:40:39 testhost dovecot: IMAP(peter): Disconnected

6 6

[Dovecot] Confusion of Sieve filter scripts and the file separator. (Amatuer Siever)
by Jack 04 Apr '08

04 Apr '08

Hello, I have just been playing around with sieve filters to split up all my incoming mail into the correct directories. I have created some subdirectories to put mailing list emails in and just want to check if my thinking is correct. I am using namespaces with the '/' character as the separator. When creating the fileinto line on my sieve scripts I thought I would have to use the '/' character to separate the folders (as horde mail also does), but that doesn't work (and produces a "Invalid folder" error) instead I must use the '.' character. I am using the dovecot deliver module and managesieve for my sieve stuff (dovecot 1.0.13). I am just wondering if this is the correct behaviour or if it is a bug of sorts. Cheers. From, Jack

2 1

[Dovecot] Dovecot config
by Dien Phan 04 Apr '08

04 Apr '08

Hi, I've installed dovecot 1.0.13 on FreeBSD 6.3 RELEASE and now trying to configure it. Below is my config file: $ dovecot -n # 1.0.13: /usr/local/etc/dovecot.conf log_path: /var/log/dovecot.log info_log_path: /var/log/dovecot-info.log protocols: imap ssl_disable: yes disable_plaintext_auth: no login_dir: /var/run/dovecot/login login_executable: /usr/local/libexec/dovecot/imap-login mail_location: maildir:~/Maildir dotlock_use_excl: yes maildir_copy_with_hardlinks: yes mail_plugins: antispam imap_client_workarounds: delay-newmail outlook-idle netscape-eoh tb-extra-mailbox-sep auth default: verbose: yes passdb: driver: checkpassword args: /var/MailRoot/bin/dovecot_xmail userdb: driver: static args: uid=vxmail gid=vxmail home=/var/MailRoot/domains/%d/%n allow_all_users=yes socket: type: listen master: path: /var/run/dovecot/auth-master mode: 384 user: vxmail plugin: antispam_signature: X-DSPAM-Signature antispam_trash: Trash antispam_spam: SPAM antispam_dspam_binary: /usr/local/bin/dspam antispam_dspam_args: --user;%u We are using a web mail software that allows users to login either as us 'abc' or 'abc@domain'. The problem is: when a user logs in as 'abc' (without domain part) dovecots creates a directory /var/MailRoot/domains/abc and check mail there (as I understand because %d is blank). My question is how to config dovecot so it automatically add default domain %d is blank. Thank you for help. Dien Phan

3 2

Re: [Dovecot] shared Sent and Junk folders {Scanned}
by Tolga 04 Apr '08

04 Apr '08

Scott Silva wrote: > on 4-2-2008 11:51 PM Tolga spake the following: >> Scott Silva wrote: >> >>> on 4-2-2008 4:01 AM Tolga spake the following: >>> >>>> Charles Marcus wrote: >>>> >>>>> On 4/2/2008 3:38 AM, Tolga wrote: >>>>> >>>>>> For some odd reason, two of my users share the same Sent and Junk >>>>>> folders, therefore I keep getting errors about Sent folder reading >>>>>> permissions and one user reading the other user's Junk folder. What >>>>>> can I do about this odd problem? >>>>>> >>>>> Do you really think anyone can even guess what might cause something >>>>> like this based on the information given? >>>>> >>>>> How about some hints, like: >>>>> >>>>> dovecot version, platform, output of dovecot -n, and maybe even - >>>>> gasp - some log entries given when these users are experiencing the >>>>> problem... >>>>> >>>> Dovecot version is 1.0.7, on Linux gentoo, below is dovecot -n >>>> >>>> log_path: /var/log/dovecot.log >>>> info_log_path: /var/log/dovecot.log >>>> listen: [::] >>>> ssl_cert_file: /etc/ssl/dovecot/server.pem >>>> ssl_key_file: /etc/ssl/dovecot/server.key >>>> disable_plaintext_auth: no >>>> login_dir: /var/run/dovecot/login >>>> login_executable: /usr/libexec/dovecot/imap-login >>>> mail_extra_groups: mail >>>> mail_location: >>>> mbox:/var/mail/:INBOX=/var/mail/%u:INDEX=/var/mail/folders/%u >>>> >>> Here you need something like mbox:/var/mail/%u/ or mbox:/var/%u/mail/ >>> to keep all the users other mailboxes separate. >>> >> I did like you said, and now I get IMAP(tolga): open() failed with >> subscription file /var/mail/tolga/.subscriptions: Not a directory. btw, >> /var/mail/tolga is a file, but what I want is to have all the other >> folders in /var/mail/folders/tolga. Should I modify it like >> mbox:/var/mail/folders/%u/ ? >> >> > That should work. It indeed work, thanks :) > > -- > > Scott Silva > Network Administrator > San Gabriel Valley Water Company > 11142 Garvey Ave. El Monte CA 91733 > 626.448.6183 x.296 > ssilva(a)sgvwater.com > > > -- > This Message has been scanned for > viruses and dangerous content by > the San Gabriel Valley Water Co.MailScanner <http://www.mailscanner.info> > and is believed to be clean. > -- Stay with *nix :)

1 0

[Dovecot] Icedove (Thunderbird) crashes when reading IMAP messages
by Cam Ellison 04 Apr '08

04 Apr '08

This is a Debian lenny/sid installation. Dovecot has been installed on it for over a year, with no problems. Lately (i.e. the past few months), if I attempt to read a message, Icedove crashes. It does not do this when I read a message downloaded from the ISP's POP3 server (not that I do this very often - almost all my mail goes through Dovecot). I am now having to read email with KMail, but cannot send from it - that's another story. Googling for the past month or so reveals nothing: there seems to be no bug filed for this, and indeed I think it is an odd communication problem with dovecot that has not been seen before. Has anyone any experience with this, or any ideas how I should go about resolving it? I have not done so (because I don't know how to use the data, not being a coder), but can run strace and/or valgrind on icedove if that will help. Thanks in advance Cam -- Cam Ellison Ph.D. R.Psych. #01417 Cam Ellison & Associates Ltd. Management Psychology 3446 Beach Avenue Roberts Creek BC V0N 2W2 Phone: 604.885.4806 Fax: 604.885.4809 Cell: 604.989.0635

5 9

[Dovecot] shared Sent and Junk folders
by Tolga 03 Apr '08

03 Apr '08

Hi, For some odd reason, two of my users share the same Sent and Junk folders, therefore I keep getting errors about Sent folder reading permissions and one user reading the other user's Junk folder. What can I do about this odd problem? Regards, Tolga -- Stay with *nix :)

4 5