- dovecot - dovecot.org

[Dovecot] v2.0: tcp-wrappers support
by Timo Sirainen 06 Feb '10

06 Feb '10

Some people have been requesting this forever. So finally this works: service tcpwrap { unix_listener login/tcpwrap { mode = 0666 } } login_access_sockets = tcpwrap There can also be more sockets listed, in case someone wants to build their own access checking service.

1 0

[Dovecot] pop3 quick question
by Nikita Koshikov 06 Feb '10

06 Feb '10

Hello list, I have found imap_allowed option in dovecot which do exactly what I need, for imap. Is there something similar for pop3 protocol? The main goal for it - is to enable pop3 access to specific users. Or maybe dovecot have some 'pop3=yes' or 'imap=no' args to userdb/passdb sections ? I know future dovecot 2.* versions will support rewriting for protocols, but this is required for 1.2 branch.

2 2

Re: [Dovecot] mdbox compression
by Damon Atkins 06 Feb '10

06 Feb '10

ZFS has support for compression on the file system ( lzjb | gzip | gzip-N | zle ). gzip eats CPU even at levels as low as 3. From the zfs man page The lzjb compression algorithm is optimized for performance while providing decent data compression. You can specify the gzip level by using the value gzip-N where N is an integer from 1 (fastest) to 9 (best compression ratio). The cost of CPU to support compression (gzip) vs the Cost of Disk if mboxes are not compressed? The other interesting feature ZFS (OpenSolaris) has is de-duplication, it detects if the block which is about to be written, is the same as another block and instead of writing the block it puts a pointer to the already existing block, saving disk space. Only problem with this is you would need to filter out headers which are different even when the to/body is the same. Cheers

1 0

[Dovecot] mdbox compression
by Timo Sirainen 06 Feb '10

06 Feb '10

I was wondering if I should add compression support to mdbox one mail at a time or one file (~2MB) at a time. The tradeoffs are: * one mail at a time allows quickly seeking to wanted mail inside the file, but it can't compress mails as well * one file at a time compresses better, but seeking is slow because it can only be done by uncompressing all the data until the wanted offset is reached I did a quick test for this with 27 MB of my old INBOX mails: (note the -b option, so it doesn't count wasted fs space) mdbox/storage% du -sb . 15120350 . Maildir/cur% du -sb . 16517320 . % echo 1-15120350/16517320|bc -l .08457606924125705623 So, compressed mdboxes take 8.5% less space. This was with regular gzip compression with default level. With bzip2 -9 compression the difference was 10%. Any thoughts on if 8-10% is significant enough improvement to make seeking performance worse? Or perhaps I should just implement both ways.. :)

4 6

[Dovecot] Dovecote does'nt start - no log ?
by Jean-Francois 06 Feb '10

06 Feb '10

Hi All, Actually, I installed Dovecot on an OpenBSD server, however, after properly setting (assumed) the confs, Dovecot does'nt start. I don't know the reason but don't know how to debug. Where is the log about reason for not starting ? Thank you for helping. Regards

4 11

[Dovecot] Forcibly emptying a POP3 mailbox
by Răzvan Sandu 05 Feb '10

05 Feb '10

Hello, I've recently saw a discussion on this list, about a standardised way of emptying POP3 mailboxes on the server - but no solution came up. Please provide a suggestion for the following situation: - I am running a stock POP3 dovecot server, on Fedora 12 (dovecot version 1.2.9) - passwd-file authentication via /etc/imap.passwd - users are virtual, sharing the same UID/GID; - storage is in MailDir system; - for remote users, server is POP3-only, since the administrative policy says that every user should download his messages on the local workstation, then delete them from server immediately after succesful download. - IMAP is allowed only locally, in order to be used together via stock squirrelmail Web interface (for emergencies only) However, there are users that check the infamous „Leave a copy on the server” checkbox (or equivalent) in their mail clients, eventually adding „Delete messages after X days”. Since I have no direct administrative control on the clients, I absolutely need a way to forbid the above behaviour at server level, in order to enforce the policy. The server's HDD is meant to store received messages from reception to the first access only. Filesystem quotas are not an option in this scenario, since the UID/GID is the same for all virtual users. What I need is way to make sure that "cur" and "tmp" subdirectories are perfectly empty after the user downloads his mail. (I will also need a way to make sure that users using squirrelmail leave no populated "Sent" and "Trash" folders on server's HDD, but that's offtopic here). Could you please suggest a way to solve this ? Thanks a lot, Răzvan

5 5

Re: [Dovecot] Forcibly emptying a POP3 mailbox
by Răzvan Sandu 05 Feb '10

05 Feb '10

Hello, Thanks a lot, but the problem here are not the IMAP users accessing the server via squirrelmail. Few people use this. What I mainly want to avoid are *POP3* users checking the „Leave a copy on the server” checkmark. I need a way of automatically performing the DELE command at server level, just after a POP3 user *succesfully* downloaded his messages via POP3... Thanks again, Răzvan

5 4

[Dovecot] postfix and dovecot on different systems - sasl request
by thomas polnik 05 Feb '10

05 Feb '10

Hello, until now I have use postfix and dovecot on the same server. Now I want/must split this both systems (server1 - postfix, server2 - dovecot). I used dovecot auth system for postfix ## snip ## smtpd_sasl_type = dovecot smtpd_sasl_path = private/auth ## snap ## but this does not work, if dovecot is not running on the same server. I did not find any information in postfix/dovecot documentation, how I could solve this problem smart. My only idea, but it looks like quick and dirty solution: On the postfix system a small dovecot systems runs only for postfix sasl via /var/spool/postfix/private/auth. (Or I use saslauthd with sql or imap) The postfix system (on server 1) delivers mails via transport map to the dovecot system. On the dovecot system postfix delivers this mails i.e. via ## snip ## dovecot unix - n n - - pipe flags=DRhu user=vmail:vmail argv=/usr/lib/dovecot/deliver -d ${recipient} ## snip ## to the mail boxes. Are there other ways, how I could solve this problem? My wish configuration: postfix sends the auth request via tcp/ip to the dovecot auth system and not via a socket. I use postfix 2.6 and dovecot 1.2.6. Best regards, thomas polnik.

2 1

[Dovecot] dovecot-auth stops responding
by Jonathan Siegle 05 Feb '10

05 Feb '10

Running dovecot 1.2.4(imap only) on AIX out of inetd. I accept passwords(PAM for auth) and kerberos tickets. Occasionally we see dovecot-auth stop responding. I just kill it off and another process is spawned which works fine. It doesn't look like it is hitting any ulimits(11MB for memory). I issued a kill -11 and ran it through the debugger. Here is what it says: dbx dovecot-auth core Type 'help' for help. [using memory image in core] reading symbolic information ... Segmentation fault in __fd_poll at 0x900000000117634 ($t1) 0x900000000117634 (__fd_poll+0x98) e8410028 ld r2,0x28(r1) (dbx) where __fd_poll(??, ??, ??) at 0x900000000117634 ioloop-poll.poll(__listptr = 0x00000001104dd330, __nfds = 145, __timeout = 1051), line 121 in "poll.h" io_loop_handler_run(ioloop = 0x000000011001bfb0), line 157 in "ioloop- poll.c" io_loop_run(ioloop = 0x000000011001bfb0), line 335 in "ioloop.c" main(argc = 1, argv = 0x0ffffffffffffad8), line 347 in "main.c" Here are my ulimits: ulimit -a time(seconds) unlimited file(blocks) 2097151 data(kbytes) 1048576 stack(kbytes) 65536 memory(kbytes) 65536 coredump(blocks) 2097151 nofiles(descriptors) 32768 Any help would be appreciated. Thanks, Jonathan

3 12

[Dovecot] Fatal: chdir( ) failed: Permission denied
by Vallo Kallaste 05 Feb '10

05 Feb '10

Hi Dovecot fails to chdir() into users home directory giving the following error: Oct 9 16:43:49 jura dovecot: dovecot: Fatal: chdir(/a/home/mina) failed: Permission denied (euid=1118 0(mina) egid=11332(devel) missing +x perm: /a/home) Oct 9 16:43:49 jura dovecot: dovecot: child 14441 (imap) returned error 89 (Fatal failure) The reason seems to be that user primary group is "devel" and directory /a/home has ACL which gives execute permission only to the users in "domain users" group. Now the catch is that the user _is_ member of "domain users" group and is happily accessing his home directory over CIFS(Samba) for long time. Also, the user has no problem entering his home directory when using shell access. It just seems like the user lost membership of a bunch of supplementary groups. The group membership info is coming from M$ PDC via winbind and NSS. We have few users and can change the users primary group and/or relax the /a/home directory ACL, but overall this is not right. Shall I shut up or can we hope for a fix or at least explanation? Thanks jura:/a # ls -la drwxrws--T+ 3 root domain admins 18 Oct 8 15:10 home jura:/a # getfacl home/ # file: home # owner: root # group: domain\040admins user::rwx group::rwx group:domain\040users:--x mask::rwx other::--- -- Vallo

4 8