Re: [Dovecot] Embracing .99
Joe Allesi -X (joallesi - Coyote Creek Consulting at Cisco) wrote: high-profile applications. However, does anyone know of a sane method to use maildir on RHEL 4.x without switching over to Postfix? Use Dovecot 1.0x and the default Exim that ships with RHEL4, it all works smooth as glass. As mentioned the only thing you lose support for is Dovecot, and additionally Exim has great per-user filtering capabilities which negate the need to even use procmail. fwiw, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Squirrelmail + Dovecot + Maildir
Vicki Stanfield wrote: Here is what I see in /etc/squirrelmail/config.php $optional_delimiter = '.'; [snip] $default_folder_prefix = ''; $trash_folder = 'Trash'; $sent_folder= 'Sent'; $draft_folder = 'Drafts'; $default_move_to_trash = true; $default_move_to_sent = true; $default_save_as_draft = true; This may or not be related, but I had a lot of trouble getting SM to work in it's "dovecot" mode, I'm not sure it's entirely up to snuff - I haven't spent the time to debug. Instead I switched it back to Courier mode (we migrated from Courier) and it works like a champ - you just need to adjust your $default_folder_prefix if you don't actually run Dovecot in Courier-like mode (namespace), I do believe. Here are what I believe to be the bits out of my config.php that matter: $imapServerAddress = '127.0.0.1'; $imapPort = 143; $imap_server_type = 'courier'; $invert_time= false; $optional_delimiter = '.'; $default_folder_prefix = 'INBOX.'; $trash_folder = 'Trash'; $sent_folder= 'Sent'; $draft_folder = 'Drafts'; $default_move_to_trash = true; $default_move_to_sent = true; $default_save_as_draft = true; $show_prefix_option = false; $auto_expunge = true; $default_sub_of_inbox = false; $show_contain_subfolders_option = false; $auto_create_special= true; $delete_folder = true; $noselect_fix_enable= false; $allow_thread_sort= true; $allow_server_sort= true; $allow_charset_search = true; $uid_support = true; $no_list_for_subscribe = false; $imap_auth_mech = 'login'; $use_imap_tls = false; The matching namespace in dovecot.conf that goes along with this is: # Courier-IMAP friendly namespace private { separator = . prefix = INBOX. inbox = yes } Hope this helps some, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] Multiple private namespace question
(apologies if this has been covered, can't find it on the Wiki/FAQ, etc.) We have a stellar, fantastic, hard working Dovecot 1.0.5 (yes, I need to upgrade) installation running in Courier-IMAP compatibility mode: mail_location = maildir:~/Maildir:INDEX=/var/spool/dovecot/indexes/%1u/%u namespace private { separator = . prefix = INBOX. inbox = yes } A user purchased a Windows Mobile 6 device which is ignoring the namespace (it's the device's bug for sure, other people report it on forums around the 'net) and there's no way on this device to manually set the Prefix; so, his Sent mails are never making it back to the server when replying from the device because of the missing INBOX. issue. Can I add a second hidden namespace with no prefix: namespace private { separator = . prefix = hidden = yes } ...to fix his problem, but without adversely affecting anyone else? The issue that has given me pause is this comment in the config file: "Default namespace is simply done by having a namespace with empty prefix." ..but then there's also: "If namespace is hidden, it's not advertised to clients via NAMESPACE extension or shown in LIST replies." I'm not sure what exactly will happen by adding a hidden namespace with no prefix after an advertised namespace with a prefix; this is a live server, so mucking up the works with a bad config is not an option. :) Anyone have a clue if this is valid/good/correct and Dovecot will play nicely? Many thanks, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Multiple private namespace question
Charles Marcus wrote: On 5/23/2008, Troy Engel ([EMAIL PROTECTED]) wrote: Can I add a second hidden namespace with no prefix: Certainly... http://wiki.dovecot.org/Namespaces See, the wiki page has it backwards though of the way I need to do it; the empty prefix namespace is the inbox = yes, not the hidden = yes. I'm giving pause because I can't find examples of anyone doing it the other way around; empty prefix, hidden = yes as NOT the default namespace (inbox = yes). Have you run the reverse configuration like I need and can confirm it works? thx, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Multiple private namespace question
Charles Marcus wrote: No, but there's no reason it wouldn't... a namespace is a namespace. Easiest is to just try it... ...I think you missed the whole part of my original email that stated this is a live, hard working server and "just trying things" is not an option. I'm normally a just-push-the-button type of guy, but in this case I need to get quantified evidence that it will actually do what I need without breaking other things. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Multiple private namespace question
Timo Sirainen wrote: But I don't think there are any problems if you make INBOX. namespace the hidden one. Most clients don't care about namespaces. Hi Timo, thanks for the reply - I understand what you're saying. Can you theorize what will happen to the already configured/subscribed users (Thunderbird) whose tree looks like: INBOX |_ Sent |_ Drafts ...when I would change the default namespace around and make the empty prefix inbox=yes? Would everything suddenly become "flat" on them and cause confusion? Ala: INBOX Sent Drafts Most Thunderbird instances already configured say "INBOX." under the setting of '(IMAP Account) -> Server Settings -> Advanced -> Personal Namespace'; I think this value is filled in automatically by Thunderbird the first time you configure the account and it queries the server and received the namespace advertisement. Would it be correct to assume that all these existing users would be fine (because of the hidden INBOX prefix namespace, compatibility mode), and only new users would see a flat hierarchy when setting up Thunderbird? Many thanks, -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] tb-negative-fetch workaround
Timo Sirainen wrote: I'm running out of ideas. Check with rawlog what exactly Dovecot and Thunderbird are talking to each others (http://dovecot.org/bugreport.html). Are there any large UIDs either? What is Thunderbird doing just before it sends that command? Maybe its local cache is broken? I don't have actual ideas on the problem, but I can offer up some help on how to debug it. Save the below as "runtbird.bat" and place it on the user's desktop, and tell them to run this instead of the ThunderBird icon. This will log all the debug info from the client so you can go through it: == snip == set mydate=%date:~-4,4%%date:~-7,2%%date:~-10,2% set mytime=%time:~0,2%%time:~+3,2% set NSPR_LOG_MODULES=IMAP:5 set NSPR_LOG_FILE=c:\thunderbird_%mydate%_%mytime%.log start /d "c:\program files\mozilla thunderbird" thunderbird.exe == snip == Obviously adjust the log location and ThunderBird install folder as appropriate. If you want to log all modules instead of just IMAP (SMTP, e.g) then replace "IMAP" above with "all". Note the above will create a date/time stamped logfile for each run, so you won't lose the previous logs. The 'mydate' and 'mytime' are a little bit of smarmy batch-fu, make sure you get all those percent chars and whatnot in your copy/paste. hth, -te ref: http://www.mozilla.org/projects/nspr/reference/html/prlog.html -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] tb-negative-fetch workaround
bofh list wrote: Cool - the user is using linux but I would imagine the changes to accomplish this are minor! (could you *please* turn off sending HTML email to this list? argh.) While I've not done this for a linux person, the changes should be minor; it will vary depending on the distro. For instance, on Red Hat /usr/bin/thunderbird is actually a shell script that launches thunderbird-bin. Something like... == snip == #!/bin/sh MYDATE=`date "+%Y%m%d_%H%M%S"` NSPR_LOG_MODULES=IMAP:5 NSPR_LOG_FILE=/tmp/thunderbird_${MYDATE}.log export NSPR_LOG_MODULES NSPR_LOG_FILE /usr/bin/thunderbird & exit $? == snip == The above is untested but looks good to my eye... :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] tb-negative-fetch workaround
bofh list wrote: I don't think it is unreasonable to send both text/plain and text/html. I do think it is unreasonable to send only text/html. I'm sending both. *sigh* I am not trying to start a flame war here; please Google around and look up mailing list etiquette, it is different than normal emails between two people or a small group. It is generally considered rude and unacceptable on almost every mailing list in existence to post HTML-based email, plain text only is the standard and has been since the dawn of (computer) time. I'm done, nothing more will be said. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Postfix+Dovecot+LDAP
Timo Sirainen wrote: On Wed, 2007-03-14 at 12:20 -0500, Ejay Hire wrote: The general consensus was that nss_ldap has some emotional problems and dovecot seems to annoy it Kind of like poking an angry snapping turtle with a stick. The problem has been found since and it can be worked around: http://wiki.dovecot.org/AuthDatabase/Passwd Should this paragraph be updated in the dovecot-example.conf? == # /etc/passwd or similar, using getpwnam() # In many systems nowadays this uses Name Service Switch, which is # configured in /etc/nsswitch.conf. WARNING: nss_ldap is known to be broken # with Dovecot. Don't use it, or users might log in as each others! # http://wiki.dovecot.org/AuthDatabase/Passwd userdb passwd { } == Maybe that should reflect the "set blocking=yes" statements from the wiki? Or doesn't this apply to userdb {} blocks? I'm a little confused if I should be setting blocking=yes for passdb, userdb, or both... The example conf lists blocking=yes in the passdb {} blocks, but the Wiki specifically gives an example for a userdb{} block. Is blocking=yes applicable to both types of blocks? Should I be setting it both places? Thanks, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Postfix+Dovecot+LDAP
Timo Sirainen wrote: I accidentally added it to passdb instead of to userdb. It doesn't at least currently work in passdb passwd (although it could, but does anyone use it anyway?) OK thanks for that, but I'm still a bit confused -- I'm after the second part of the wiki entry in relation to nss_ldap; I was thinking to set blocking=yes to just avoid any possible problems in the future with PAM/nss_ldap. A) Does blocking=yes still apply to passdb pam {}? B) Do I still add blocking=yes to userdb passwd {} if I'm using passdb pam {} as above? Is there no such block as 'userdb pam {}' since I don't see one in dovecot-example.conf? Thanks for helping, it just doesn't seem clear to me yet what the right combination of settings are. I'll happily update the wiki to be more clear once y'all can get me on track. :) All my user and pass auth is done via PAM/nss_ldap to an external server. Thanks, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] Courier keywords - anyone even tried?
I do realize from the Wiki that the formats are incompatible, and am looking at the Courier spec (http://www.courier-mta.org/imap/?README.imapkeywords.html) and the Dovecot spec (http://wiki.dovecot.org/MailboxFormat/Maildir) trying to see if there's any sane way I could try and write a shellscript. The pain here is the Courier spec, I think I lost a few brain cells trying to understand this convoluted methodology ("When encountering ".N.file" after another ".N.file" was encountered earlier, remove the file with the lesser N, unless the larger of the two Ns is greater than or equals to T." e.g.) Has anyone even started down this path and has some prototype code written that they gave up on? Is this just so painful trying to migrate that I should just give up? Any ideas welcome... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Postfix+Dovecot+LDAP
Timo Sirainen wrote: Yes, http://wiki.dovecot.org/PasswordDatabase/PAM explains what it does and why it might not be a good idea. Although no-one has yet reported any success/failure stories, those are all my own guesses.. I gotcha, makes sense having read it. Now I'm back on the fence, as I'll want to research the status of these PAM/nss_ldap memory leaks that might occur if we set it to blocking=yes. It seems like the choice is to be asynchronous and possibly run into a user collision on the file descriptor (percentage chance, anyone? 2%? 5%?) or to use a synchronous/blocking pipeline at the expense of never releasing the PAM code (sic) and possibly leaking memory. Here comes my "I am not a real programmer" thought -- there really is no way for dovecot-auth's forked child processes to set up their own unique descriptor? Maybe an alternative way of doing the fork? Just pondering... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Postfix+Dovecot+LDAP
Timo Sirainen wrote: blocking=yes doesn't break anything with nss_ldap, since without blocking=yes it'll run in one process anyway. PAM works differently. Thanks for clarifying that. You're somehow mixing up these things. :) Probably because of the "blocking" naming, which actually does the opposite of what it's named.. Yes, thanks for explaining further -- I was completely reading blocking=yes backwards from what you had designed as you pointed out. I was reading it as an instruction to Dovecot what to do, not an explanation to Dovecot what the machine is already doing. blocking=no pam, blocking=yes nss_ldap: No memory leaks leaks. Fixes nss_ldap problems. Each PAM lookup is done in a forked process. NSS lookups are done in auth worker processes, as described above. So again no lookup blocks others. OK this seems like the perfect solution; in dovecot.conf terms for a setup such as mine (nothing in /etc/passwd, 100% LDAP lookups for homedir, password, /etc/nsswitch.conf "passwd: files ldap", etc.) this would then be: passdb pam { args = cache_key=%u dovecot } userdb passwd { args = blocking=yes } This would not block/stall in the pipelines, not cause memory leaks (since underlying code is released each cycle), avoid/fix nss_ldap issues with file descriptor reuse. Do I finally have a good understanding now? (thanks for taking the time to work it out) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] a build/install question
Stewart Dean wrote: Alas, when I actually tried to run dovecot, it was asking for its resource /in the original build tree/ location, which wasn't there. Is there some way I can do what I want? While I haven't done this specifically with Dovecot, you should be able to manage everything in a simple manner like you want; I currently do this with Exim and Courier-IMAP on the old production server solely with switches to ./configure. Here's how it basically works, maybe it'll help you get dovecot working: Source: /opt/build/courer-imap-1.2.3 $ ./configure --prefix=/opt/courier-imap-1.2.3 --with-authchangepwdir=${exec_prefix}/libexec/authlib --with-makedatprog=${exec_prefix}/libexec/makedatprog (more and more switches that don't matter for this) After building, a simple "make install" places everything in /opt/courier-1.2.3. Then there's a symlink like so: /opt/courier-imap -> /opt/courier-1.2.3 ...and finally the /etc/init.d/ script loads the actual courier processes using /opt/courier-imap, not the "standard" /usr based things (so you'll need to customize to fit the default dovecot init.d script). When I upgrade (or downgrade) it's as simple as running the same scenario on the new code (which means a new make install does not overwrite the old), migrate the config files as needed, stop the old daemon, switch the symlink, then start the new daemon up. Users don't see interruption in their email experience and I'm confident of a backup plan in an emergency. This has been working great for many years upgrading Courier and Exim on the same machine; rarely do I have to fall back but when it does happen it's been as simple as pie. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] a build/install question
Stewart Dean wrote: > Thanks Timo, that should do it. As always, (as us boorish Americans > might say), you da man! Your patience and help are always the best. BTW: my instructions work just the same -- compile on a dev machine, tar and untar on a production machine. It doesn't matter where you compile as long as the OSes are compatible (glibc, etc.). What matters is developing a logical symlink infrastructure to accomplish your needs. I kind of thought that would have been obvious, sorry. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] The perfect Courier -> Dovecot conversion tool
Timo Sirainen wrote: I've tested this only with a test user. http://dovecot.org/tools/courier-dovecot-migrate.pl Thanks! The initial runs are looking good; could you beef up the convert_subscriptions() function to spit out debugging info when run in test mode? I have people with really gnarly folder names that I want to be sure look like they'll be converted correctly. So instead of returning, maybe something like: my ($fin, $fout); open ($fin, $in_fname) || die $!; if ($do_conversion) open($fout, ">$out_fname") || die $!; else $fout = STDOUT; while (<$fin>) { chomp $_; (warning, I'm not a perl hacker - it looks good on paper to me. :)) This way it'll spit out to screen what would be going into subscriptions. Or maybe some better variation of my idea... thx, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] 1.0.rc28 released / suse rpmbuild fails what has changed ?
Timo Sirainen wrote: RPM build errors: Installed (but unpackaged) file(s) found: /usr/lib/dovecot/idxview /usr/lib/dovecot/logview No, new stuff was added. Don't know why rpm build has to fail if that happens though.. This is normal -- the packager (Robert) needs to update his spec file to include these two new programs to the binary output RPM. This 'error' is on purpose, it's rpmbuild trying to help you notice changes that have been made. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] 1.0.rc28 released / suse rpmbuild fails what has changed ?
Robert Schetterer wrote: quick grep shows idxview is complete new to 28 Read Timo's changelog, no need to grep: "+ Added idxview and logview utilities to examine Dovecot's index files" so now i have to think how to get out of that rpm warning All you have to do is edit your dovecot.spec and *include* these two files. Don't overthink the situation. :) While you're at it, pay attention to this Changelog entry: "* Added "-example" part to doc/dovecot-sql-example.conf and doc/dovecot-ldap-example.conf. They are now also installed to $sysconfdir with "make install"." You will also need to adjust your spec file as necessary to pull in these two new files which should be throwing errors in your RPM build process, just like the new log utilities did. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] 1.0.rc28 released
Timo Sirainen wrote: + When copying/syncing a lot of mails, send "* OK Hang in there" replies to client every 15 seconds so it doesn't just timeout the connection. This is probably the best news ever. I just had one of my testers beat on this with ~4000 emails-at-once operations which have always timed out and caused us grief. While it still takes forever and a day to accomplish (I blame Thunderbird), it now at least chugs along and completes the ops with no user errors. Thanks. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] idxview: Fatal: file hdr read() 41984 != 1170103198
Playing with the new idxview I got this error running against my INBOX dovecot.index.cache file. What exactly does it mean? Full output: # /usr/libexec/dovecot/idxview dovecot.index.cache -- INDEX: dovecot.index.cache version = 1.8 base header size = 0 header size = 1170103198 record size = 1170103283 compat flags = 7 index id = 0 flags = 26512 uid validity = 0 next uid = 2460778624 messages count = 0 recent messages count = 632 seen messages count = 1 deleted messages count = 1140057508 first recent uid lowwater = 4294966816 first unseen uid lowwater = 3 first deleted uid lowwater = 1836 log file seq = 0 log file int offset = 2 log file ext offset = 1140057513 sync size = 266287972361 sync stamp = 21 day stamp = 0 day first uid[0] = 1378709065 day first uid[1] = 2037149797 day first uid[2] = 980374573 day first uid[3] = 826620960 day first uid[4] = 2019113286 day first uid[5] = 808463731 day first uid[6] = 1279801136 day first uid[7] = 1078085165 Fatal: file hdr read() 41984 != 1170103198 ('idxview dovecot.index' and 'logview dovecot.index.log' come back without any Fatal type errors, just gobs of information) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] idxview: Fatal: file hdr read() 41984 != 1170103198
Timo Sirainen wrote: Use idxview dovecot.index dovecot.index.cache Oooh, gotcha -- now it makes a lot more sense. Thanks. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] New mail not appearing consistently in Outlook 2003
Timo Sirainen wrote: I think your Outlook has just broken its internal cache. Try recreating the account or clearing the cache if you can figure out how. This may help y'all: http://www.slipstick.com/mail1/olinet.htm#imap On many occasions with Outlook+Exchange Calendaring, I've had to delete user's cache files to fix problems with meetings not getting accepted or sent properly; I have no experience with the IMAP portions of it, but I can say those cache files are finicky. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Mailbox isn't a valid mbox file
Jay Chandler wrote: Keep getting this error within Squirrelmail (and other clients) on one particular user's mailbox file: ... Anyone have any ideas? Only one user out of many is having this issue, but he's whiny... Try taking the physical mbox file, putting it on a client machine with Thunderbird in the Mail/Local Folders/ directory, then allow Thunderbird to open/rewrite/save it back out. Then stick it back on your server and see if that fixes it... Or, maybe even do the above, create a *new* folder and drag&drop all the messages from folder A into folder B (so a fresh-write would have to happen), then put folder B back onto the server named as the original. TBird reads/writes regular old normal mbox files. Just some ideas, hth. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Shared mailbox plans
Timo Sirainen wrote: If ACL plugin isn't loaded, I suppose the mailboxes can be accessed if the process has filesystem permissions to access them. Kind of, but it doesn't really work in real life. I tried setting it up initially around... say 1.0rc23 and it runs into permissions problems writing the various dovecot.index files because they're 0600 mode to that specific imap logged in user. The config (commented out right now) looks like: (*) # Spam Training namespace public { separator = . prefix = SpamTraining. location = maildir:/var/spool/imap_shared/SpamTraining } What I was finding with testers is that each person's login process was rewriting permissions on the subscriptions file and the index files didn't work out for the same reason; 1 person would drop an email into a subfolder ("MissedSpam" e.g.), Dovecot would write indexes 0600 to that named user then everyone else would start getting errors. Unfortunately I didn't have the time to do massive debugging right then so just had to comment it out and save it for another day. :-/ -te (*) right now it works with gnarly ugly symlinks in a user's ~/Maildir under Courier-IMAP, and I have two shellscripts that nightly read the contents of two subfolders - MissedSpam and FalsePositives - and passes those to SpamAssassin with 'sa-learn --spam' e.g. then deletes each mail file. -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] rc31 -> "deferring operation: binding"
Ben Beuchler wrote: I installed a virgin install of OpenLDAP 2.3.32 directly on the mail server, copied over my LDAP database, pointed Dovecot at it, and started it up. I immediately started getting the same log entries: Some ideas from the peanut gallery here -- if you run 'ldd dovecot-auth', what openldap library is it linked into? Have you tried compiling dovecot against this new 2.3.32 openldap install? cliff:~ root# egrep -v '(^$|^#)' /opt/dovecot/etc/dovecot-ldap.conf ldap_version = 3 Have you tried binding as version 2, just to see what happens? -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Shared mailbox plans
Timo Sirainen wrote: I think I'll also add a check to compare Maildir, new, cur and tmp directories' permissions and log a warning if they're not the same. Just as a sanity check, mine differ (for whatever reason - I probably used 'maildirmake' from Courier) like so: $ ls -gGld SpamTraining/ drwxrwxr-x 7 117 Feb 2 10:21 SpamTraining/ $ ls -gGld SpamTraining/[new,cur,tmp]* drwxrwx--- 2 6 Sep 9 2004 SpamTraining/cur drwxrwx--- 2 6 Sep 9 2004 SpamTraining/new drwxrwx--- 2 6 Sep 9 2004 SpamTraining/tmp $ ls -gGld SpamTraining/.MissedSpam/ drwxrwxr-t 5 36 Apr 7 09:17 SpamTraining/.MissedSpam/ $ ls -gGld SpamTraining/.MissedSpam/[new,cur,tmp]* drwxrwxrwt 2 6 Feb 3 02:06 SpamTraining/.MissedSpam/cur drwxrwxrwt 2 6 Feb 2 10:26 SpamTraining/.MissedSpam/new drwxrwxrwt 2 6 Feb 2 10:28 SpamTraining/.MissedSpam/tmp So whether or not *my* permissions are correct (they actually look pretty screwed up), implementing that check would start blowing out warnings in the logfile; yet these folders work perfectly fine I suppose. FYI only. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] v1.0.0 released
Tomas Janousek wrote: The Fedora folks are listening and will issue an update soon. They put it into rawhide in today's morning and I suppose they could put it into FC6 within a few days. ;) Hi Tomas, if you want to feed this back upstream based on the work I've done with Axel: dovecot-1.0.beta2-pam-tty.patch: no longer needed, applied upstream in rc22 (slightly different but same result) dovecot-1.0.rc15-default-settings.patch: needs reworked, these files it's patching have changed a lot. The latest rework I did was for rc29, which applies cleanly to 1.0.0 as well (just verified today). Additionally we added a default (commented out) /var/log/dovecot.log setting for the logfile in the default conf file. There's also an added dovecot logrotate.d script (using SIGUSR1) to support the above log; alas, right now I can't connect to the atrpms.net servers to post a link. Feel free to ping me off list and I'll file-attach the patches to save the Fedora folk some work. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] Maildir, INDEX, and folders
I'm reading the wiki et el. regarding indexes, but can't come up with a coherent answer to the setup and what will happen. I want to take the indexes and move them off of NFS (~/Maildir) and place in /var/spool/dovecot/indexes, but not have to hand create everything (user homes, mail folders) So, I create a parent folder /var/spool/dovecot/indexes that has the appropriate permissions for writing by all users/dovecot. If I then set: mail_location = maildir:~/Maildir:INDEX=/var/spool/dovecot/indexes/%1u/%u ...will the dovecot daemon(s) automagically create (A) the parent user folders (i.e. .../indexes/t/tengel/)(1), and (B) the necessary subfolders for all the child folder indexes? (i.e. .../tengel/.Lists.Dovecot/) Most of the doc info seems to revolve around moving your indexes if you use mbox, I'm a little fuzzy on what happens with Maildir when you attempt this. thanks! -te (1) in our setup, initially user's home folders are created on an admin server -> NFS that is not the mail server, so Dovecot needs to create the home & index folders on it's own at runtime. -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Maildir, INDEX, and folders
Timo Sirainen wrote: How about just trying it in a test installation? :) The answer is anyway "yes, they're created". Thanks, I've added a small section to the wiki clarifying this point. This *is* my test installation but I have a dozen users working on it before taking it into production. I needed to have a clue what would happen before I made a change and possibly blew things up. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Maildir, INDEX, and folders
Justin McAleer wrote: It will, but beware... all directories will be created as the user's uid and gid. So, if you use a different uid for each user, you should manually create the /var/spool/dovecot/indexes/%1u directories first with the proper ownership and permissions. Otherwise, when Thanks, I *just* found this out! The UID/GID doesn't matter a lot except for the fact that the perm is 0700 on /a/, as you mention. I reset them to just %u real quick while I manually make all the [a-z] parents, then I'll reset it to %1u/$u as desired. Glad I didn't try this on a production machine. :) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] FYI: TB 2.0.0 + DC 1.0.0 = OK so far
Just a status note - my beta group is upgrading/upgraded to the new Thunderbird 2.0.0 (OS X & Win32 so far, Linux to come) and beating it up against Dovecot 1.0.0 and all is A-OK, nary a problem in sight. Obviously just an initial status report, but good mojo so far for those curious about trying it. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Maillog rotates, but dovecot still writes logs into old logfile?
Mart Pirita wrote: The only solution I did found, is adding into /etc/logrotate.d/syslog command to restart dovecot: /etc/rc.d/init.d/dovecot restart >/dev/null 2>&1 As per the wiki, dovecot listens to SIGUSR1 to reopen a log. You want to simply use a command like: /bin/kill -USR1 `cat /var/run/dovecot/master.pid 2>/dev/null` 2> /dev/null || true See here: http://wiki.dovecot.org/Logging#head-8886f6e09e2f691a4c45448f18f914c124dead50 -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Maillog rotates, but dovecot still writes logs into old logfile?
Mart Pirita wrote: syslog_facility = mail should do the trick and log all via syslog to /var/log/maillog? That's all you need on a modern Linux system, as the syslog facility 'mail' usually is sent to /var/log/maillog; this is the default for all SMTP/IMAP/POP3/etc. on any Red Hat -esque system (RHEL, Fedora, CentOS). Check your /etc/syslog.conf file. Personally I prefer my logs discrete, and use the settings right out of the Wiki page to /var/log/dovecot.log and the included logrotate.d/ script. Works 100%. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] dovecot.spec
J.Palacios wrote: Plis, where can i find an updated version (for recently dovecot v1.0.0) of dovecot.spec, needed to build an rpm for RH 4? Axel has a fully ready spec/RPM on ATrpms.net ready to go for RHEL4; you should be able to just install and go. http://atrpms.net/dist/el4/dovecot/ -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] [courier-dovecot-migrate.pl] - No such file or directory at courier-dovecot-migrate.pl line 230
Wilkinson, Alex wrote: #/usr/bin/perl courier-dovecot-migrate.pl --recursive Finding maildirs under . ./courierimapuiddb: OK No such file or directory at courier-dovecot-migrate.pl line 230. Try running the script under strace (i.e. 'strace -e trace=file perl /script.pl') or truss to watch the perl script and see what "file or directory" it's trying to open. That should provide a quick clue. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] [courier-dovecot-migrate.pl] - No such file or directory at courier-dovecot-migrate.pl line 230
Luca Corti wrote: I was just testing the migration script, encountered the same issue and just added a print before the open || die at line 230. It tries to open the file Maildir/courierimapkeywords/:list which seems to be missing and probably not mandatory. Aha, thanks -- I too encountered the problem this weekend (deployed Dovecot server to production, migrated all our users) and noticed it was with the accounts that were either very new or never used; I suspect the assumption that this file is there is based on the account being in active use with Courier. -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Courier-imap + dovecot simultaneously?
Gunter Ohrner wrote: Am Sonntag, 29. April 2007 schrieb Charles Marcus: outlook express), one accessing his account through courier, and one through dovecot, so he can compare the speed... Doing a whole batch of operations with each server before trying the same with the other server for comparison purposes will probably reflect reality much closer. I posted some ad hoc numbers to the list back in February, one machine Courier and the other Dovecot using Squirrelmail as the frontend. http://dovecot.org/list/dovecot/2007-February/019560.html Now that the server ("Server B") is in production with 1.0.0, the users are seeing fantastic speed gains. Even just emptying the trash feels a whole lot snappier than ever... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] read only maildir
Nagyon Almos wrote: The extra (public) namespace shows up but I can not susbscribe to it (in thunderbird) or If it's read-only, then your user can't write the 'subscriptions' file; as root, edit that file and manually insert the subscription, then your clients should pick it up. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] read only maildir
Nagyon Almos wrote: Is it subscriptions or .subscriptions or in the singular (without the ending s)? Where should it be exactly? If you run dovecot 0.99, it's .subscriptions; if you run dovecot 1.0, it's subscriptions - so adjust as necessary. It is always located in the Maildir/ directory itself - take a look at your own personal maildir folder and it should be pretty obvious. The contents should be the name of the subfolder itself. Again, looking at your personal folder and just copying the idea to the shared one should be all you need to do. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Testing IMAP clients with Dovecot, problems with MacMail
Stewart Dean wrote: It looks like Mail.app treats "non-standard" ports differently. On the standard port (993), the "SSL" option seems to do pure SSL; but with another port specified, it does clean-plus-starttls (and hence fails horribly, because it's talking to a pure SSl service) Is there anything I can do to get MacMail to behave correctly? Or will it clear up when I shut down UWIMAP and give Dovecot the default ports. A Bronx cheer for Mac Mail It's possible it *might* clear up when you move to the standard ports; the server advertises STARTTLS capability (well, at least my Dovecot build does :) ). so it's possible that the Mail app is sniffing the server CAPABILITY flags, seeing a non-default port and trying to do something smart. I found this blurb: "Apple’s Mail application has checkboxes to enable TLS for incoming Internet Message Access Protocol (IMAP) or outgoing SMTP connections. STARTTLS or direct TLS is used automatically, depending on the target port." (http://sial.org/howto/openssl/tls-name/) If you have tcpdump (or similar) on one of your OS X boxen, fire it up and trap the packets between the two servers and check out the net traffic and what ports it's trying to use. That'd be where I start... Also, check out this blog post: http://blog.scottlowe.org/2005/07/02/starttls-and-imap-in-mailapp/ "So, it appears that Mail.app does indeed support STARTTLS for IMAP, but only if you set the port number back to 143 after checking the “Use SSL” checkbox." -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Dovecot Startup error
kamran arshad wrote: /var/log/maillog:May 8 09:33:13 lhr dovecot: auth-worker(default): mysql: Connected to 127.0.0.1 (exim) This looks a little strange - why would there be an (exim) in parens, are you running dovecot as your exim userid? Regardless, my first guess is that your MySQL instance has not finished starting and is not accepting connections before Dovecot is initiated, so the auth worker has no socket to connect to; what's the timing and order of your /etc/rc* tree look like? If you have the services starting in parallel, then you might have to move Dovecot further down the startup tree, or add a little shell voodoo to check for the MySQL socket before starting it... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] .imap files
L. Scott Loftin wrote: Dovecot version: 1.0.rc7 I have recently upgraded from Fedora Core 4 to Fedora Core 6 (the FC6 was a ... Is there a way to either convert the .imap* files into the new dovecot.* files or any way to recover the email that anyone knows about? FC4 used the older 0.99 dovecot installs (0.99.14 it looks like), you will need to read the Wiki and docs, and migrate your Dovecot config to the newer style -- once you do that, I bet everything will start working as intended. http://wiki.dovecot.org/UpgradingDovecot Before even starting, the first thing I'd do is upgrade to the actual 1.0.0 release, there were... 25(?) further release candidates beyond rc7. Axel has an RPM ready for you: http://atrpms.net/dist/fc6/dovecot/ hth, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] .imap files
L. Scott Loftin wrote: dovecot config file is "new". I guess I'm trying to see how to convert or somehow migrate he actual contents of the old .imap* (user email) files to the new dovcot.* files. Any ideas? This unfortunately is an mbox thing, and I've never used anything except Maildir; it sounds like the files from .imap/ are usable, you just have to use the location settings to set it up. I'll have to let an mbox expert give better info than mine. :) Try looking at these two pages: http://wiki.dovecot.org/MailLocation http://wiki.dovecot.org/MailLocation/Mbox -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] Thunderbird or Dovecot bug? (multiple keywords)
Here's a strange one. I have a user who's using Thunderbird 2.0.0 and assigning multiple tags (keywords) to emails, which are getting "lost" (I'll explain) when he leaves the folder and comes back. I can reproduce this on *his* account myself, but cannot reproduce it on *my* account (separate new TBird 2.0.0 profiles on Windows XP). Use the two TBird default tags 'Important' and 'Work' for all these tests, which it refers to as $Label1 and $Label2 in IMAP-keywords land. If I go into my account, label an email with both tags, it sticks -- I can click on Drafts then back to Inbox, and both tags/keywords still show in the client. If I go into his account, add the two tags, then click on Drafts and back on Inbox, only the *second* tag (Work, $Label2) shows. If I remove that tag, then the first tag (Important, $Label1) immediately pops up! On the backend the actual Maildir files match the dovecot-keywords and have the correct letter appended in both cases - so, in his account when the email only shows one tag, the physical mail file has the letters for two tags. So Thunderbird should be showing both tags, like it does in my account where everything is working correctly. This leads me to believe that the problem has something to do with the difference of dovecot-keywords content, and the interaction TBird has with Dovecot. My account, working correctly, looks like: dovecot-keywords: 0 $Forwarded 1 $MDNSent 2 $Label2 3 $Label1 file: Maildir/cur/1178712608.H529732P11686.yak.fluid.com:2,RSdc ...which matches, d=$Label1 (Important), c=$Label2 (Work) and TBird is displaying the email as having both tags - all is well and good. Now, his account looks like: dovecot-keywords: 0 Junk 1 NonJunk 2 $Label4 3 $Label1 4 $Label2 5 $Label3 6 $Label5 file: Maildir/cur/1178728930.P28392Q0M960111.yak.fluid.com:2,RSbde ...again, everything matches on the server; yet, it Thunderbird is only displaying $Label2 (d,3). I also noticed something interesting, even if I have the TB 'Junk Controls' completely disabled on his account, and remove all tags using the menu option, that file still retains the 'b' (NonJunk) tag in the filename: Maildir/cur/1178728930.P28392Q0M960111.yak.fluid.com:2,RSb All this long-winded explanation leads me to believe that somehow the difference in dovecot-keywords is causing this problem, but what I can't tell is who's problem it is -- Dovecot, or Thunderbird? Thoughts welcome, -te PS: all the normal things tried like deleting the dovecot index files, Thunderbird cache files and profiles, etc. -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Thunderbird or Dovecot bug? (multiple keywords)
Timo Sirainen wrote: A lot easier way is to look at the IMAP traffic and see what's different in there. Then you can see if Dovecot is giving wrong replies somewhere or if it's Thunderbird that gets confused from something. I just used the TB debugging to trap all the IMAP traffic capturing a good session (my account) and the troubled session (his account) using the same WindowsXP 2.0.0 client. Session: a) add $Label1 b) add $Label2 c) select INBOX.Drafts d) select INBOX e) remove $Label2 f) remove $Label1 Alas, I can't spot where the problem is in the conversation; can I zip up both logs and send to you? (they're pretty large and contain some private info) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Question: contention with blackberry
Stewart Dean wrote: a) a VIP has multiple secretaries accessing a single mailbox (actually, they are professional enough to have figured the realities of conflicting access and rarely have a problem), b) somebody leaving their machine on at home and coming in and firing up their work computer c) important people with Blackberriesand the BB service polls the mailbox every so often and breaks the lock. In any case, the maillog Honestly, moving to Maildir will solve all your problems with (a) and (b), and most likely (c) as well. Especially with (a), you would be doing your VIP's people a huge favor by converting to Maildir during your IMAP daemon move... $0.02. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] May 21 09:13:14 mail dovecot: imap-login: No authentication sockets found
[EMAIL PROTECTED] wrote: Ive been runing dovecot on a fc6 dell poweredge 2950 for about 4 weeks now with no issues. Im authenticating against openldap 2.3.27. It appears ok. I restarted ldap, and dovecot, and everything looked fine As Timo mentioned, it sounds like the auth socket is being deleted because it's in a place that the 'tmpwatch' tool will clean out; the default on FC6 is anything more than 720hrs old (30 days), which fits right in to your 4-week observation. Compare the location of your sockets/cache (dovecot -n) with the list of directories cleaned by tmpwatch (look in /etc/cron.daily/tmpwatch) and make sure the two aren't stepping on each other's toes. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] May 21 09:13:14 mail dovecot: imap-login: No authentication sockets found
Timo Sirainen wrote: As Timo mentioned, it sounds like the auth socket is being deleted because it's in a place that the 'tmpwatch' tool will clean out; Does some system really do this? I'd think it would break other software as well. I see at least that I have /var/run/mysqld/mysqld.sock which Only *certain* directories (ddh pasted the output of a default file). By default the /var/run/dovecot should be left alone, but it never hurts to have people check their system... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] May 21 09:13:14 mail dovecot: imap-login: No authentication sockets found
[EMAIL PROTECTED] wrote: Ive been runing dovecot on a fc6 dell poweredge 2950 for about 4 weeks now with no issues. Im authenticating against openldap 2.3.27. It appears ok. I restarted ldap, and dovecot, and everything looked fine login_dir: /var/run/dovecot/login auth default: passdb: driver: pam userdb: driver: passwd Just FYI, you are not authenticating [dovecot] against OpenLDAP; you're authenticating using PAM, and it's PAM who is doing the LDAP lookups via nss_ldap. (http://wiki.dovecot.org/AuthDatabase/Passwd) Did you have all the other 'old' (30d+) files in /var/run/dovecot/ before having restarted the daemon(s)? Any chance you saved a ls -l of that dir somehow? One thing to check is if the socket alone disappeared, or if others (the PID file, for example) were also gone. Since they'd typically have the same datetime stamp, it would provide a clue... -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] May 21 09:13:14 mail dovecot: imap-login: No authentication sockets found
[EMAIL PROTECTED] wrote: Yes I knew I was authenticating against pam. Didnt do an ls -l of the /var/run/dovecot. Wouldnt a restart of dovecot fix that though? Correct, the concept was *before* you restarted the daemon, to try and capture as much info first. It's hard when you're under the gun and need to restore services, but if it happens again I'd suggest scraping the system for clues for a few minutes. (ls -laR important dirs like /var/run/dovecot, ps -ef, maybe some lsof and lslk action, etc.) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Courier migrating issues: indexes, maildirsize, update query
Jan van den Berg wrote: mail_location = maildir:/var/spool/mail/%1u/%2u/%u:INDEX=MEMORY Of interest to you might be using a scratch disk space to store the indexes on a/the local mail server; I just did a bit of math for you (well ok, a 'bc' script did the math) and with 90.5gigs of email in Maildir folders via NFS, we have 372megs of indexes on the local disks for performance gains. mail_location = maildir:~/Maildir:INDEX=/var/spool/dovecot/indexes/%1u/%u I'm not sure the space your 100K+ mailboxes take, but maybe you can use the above numbers to do some math and see if you can do indexes like that. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] testing needed: log file concurrency
OS: CentOS 4.5 (Final) (RHEL4 clone) $ /usr/bin/time -f "total time: %E\ni/o waits: %w\n" ./concurrency writing, page size = 4096 Command terminated by signal 2 total time: 10:41.53 i/o waits: 312177 $ /usr/bin/time -f "total time: %E\ni/o waits: %w\n" ./concurrency 1 reading, page size = 4096 page size cut Command terminated by signal 2 total time: 10:39.67 i/o waits: 314930 Machine: $ uname -srvmpoi Linux 2.6.9-42.0.10.ELsmp #1 SMP Tue Feb 27 10:11:19 EST 2007 i686 i686 i386 GNU/Linux $ cat /proc/cpuinfo | egrep "(processor|model name)" processor : 0 model name : Intel(R) Xeon(R) CPU5120 @ 1.86GHz processor : 1 model name : Intel(R) Xeon(R) CPU5120 @ 1.86GHz processor : 2 model name : Intel(R) Xeon(R) CPU5120 @ 1.86GHz processor : 3 model name : Intel(R) Xeon(R) CPU5120 @ 1.86GHz Timo Sirainen wrote: http://dovecot.org/tmp/concurrency.c I'd want to know what results this program gives with different systems. Please test and reply (but don't bother if someone already replied with the same OS+result). I expect it to print: - SMP kernels: "page size cut" once in a while - UP (uniprocessor) kernels: Nothing - The most important thing is that it never prints "broken data" It might take a while for it to print anything. With my computer it takes anything from a few seconds to a minute or so. See the file itself for compiling/running instructions. So far I've tested only with Linux 2.6.21 x86-64/SMP and a slow Solaris/Sparc/UP. If you're interested in knowing what this is about: Dovecot writes to dovecot.index.log files by first writing the transaction with its size being 0. After that it writes the 4 size bytes again (using a bit special format with all bytes ORed with 0x80). I expected that when another process is read()ing the file and it notices the size being valid (all bytes having 0x80) that the whole transaction could always be read. But looks like if the size happens to be just before a memory page boundary, it's possible that the updated size is read, but the rest of the transaction isn't. -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] LogicMail for Blackberry and Dovecot
Rick Romero wrote: Unfortunately I'm not a Blackberry user, but I know IMAP clients are hard to find. So I can't comment on it's usability or features, but I thought I'd give it a little press for his quick response :) I'm a LogicMail user and it's great, I second your interactions with Derek (ps: thanks for the Dovecot bug report :) ). Currently it's lacking a few abilities but it's got the basic reading/writing/replying down. Hopefully we'll bring it up to the level of ChatterMail and it's ilk and fill a huge gap in the BlackBerry world. -te PS: for the curious, Rick's thread: http://sourceforge.net/forum/forum.php?thread_id=1755262&forum_id=589235 -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Monitoring Recomendations
Dehnert James Sr wrote: The performance issues are a) slow response on the lan, and b) unusable low response from remote users. This is always with IMAP, of we switch As a possible aide, have your user(s) run TCPView (http://www.microsoft.com/technet/sysinternals/Networking/TcpView.mspx) in a corner of their screen before starting Outlook, then when Outlook is launched to watch the view for "red bars" (which indicate a client attempt to do something and getting blocked/denied). I had a problem with setting up Outlook + Exchange + self-signed SSL + Remote awhile back; using TCPView helped show me that even though I was telling Outlook to use SSL/HTTP for the connection, the initial Outlook setup was still trying to use NetBIOS calls for the server, which showed up in TCPView very clearly. Perhaps the util will help you identify something Outlook is doing that it's not supposed to, as was my case. hth, -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Dovecot with iPhone
Timo Sirainen wrote: On Thu, 2007-07-05 at 15:54 -0400, Erik Dasque wrote: Ah, oddly, got it working with 'INBOX'. Mail.app was fine, it's only the iPhone that required it. I guess you had configured Dovecot to use "INBOX." namespace. It's probably a bug in Dovecot then: http://hg.dovecot.org/dovecot-1.0/rev/f92ea3eed616 I'll second the need/bug/fix, an employee handed me an iPhone on Monday to get hooked up to Dovecot 1.0.1 over IMAP/SSL. We run Dovecot in Courier-IMAP namespace mode, I had to find the advanced section of the iPhone Mail (labeled IMAP Prefix?) and add INBOX. to it. On a good note, it works very well - sent, drafts, folders and SSL are all humming along for him. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Virtual servers
John Hedges wrote: but I can't get ithis to work. Dovecot starts but authorisation fails. Is this kind of setup possible - is it possible to configure different passwd-files for connections on different IPs, or am I going to have to run separate instances of Dovecot for each virtual host? There have been several fixes since 1.0rc15 which deals with multiple dovecots, auth socket accidental stomping, auth caches and things like that. I would suggest you upgrade to the latest 1.0.1 first and see if that has any impact on your problem. $0.02 USD, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
[Dovecot] 1.0.1: corrupt index on fresh mailbox, ideas?
A new employee started today, when logging into his new mailbox this morning during a walk-through I got a nasty server error response in Thunderbird - something I never see, DC runs smooth as butter for us. Checking the server log I see this: == snip == dovecot: Jul 09 09:38:54 Error: IMAP(xxx): Maildir /home/x/xxx/Maildir sync: UIDVALIDITY changed (1183406107 -> 1183999134) dovecot: Jul 09 09:39:06 Error: IMAP(xxx): Corrupted transaction log file /var/spool/dovecot/indexes/x/xxx/.INBOX/dovecot.index.log.2: invalid indexid (1183406107 != 1183999146) dovecot: Jul 09 09:39:06 Error: IMAP(xxx): Corrupted index cache file /var/spool/dovecot/indexes/x/xxx/.INBOX/dovecot.index.cache: indexid changed == snip == Some facts that make this a strange occurrence: 1) new IMAP mailbox, only one email was in it that I had sent last week (boilerplate Welcome email) 2) was working perfectly fine on Friday when I set things up (Thunderbird and SquirrelMail) and tested with the same client laptop and same Thunderbird 3) nothing bad happened with the servers over the weekend, all is tip-top - no restarts, reboots, nothing. 4) DC does use NFS mounted maildirs, but there is only one DC accessing any maildir - no other servers touch ~/Maildir/. mmap is disabled. What could be a reason for this index to have gotten corrupted? This is the first time I've had an index go corrupt out of all the users, so I'd like to try and track down the cause of this... thx, -te == dovecot -n == # 1.0.1: /etc/dovecot.conf log_path: /var/log/dovecot.log protocols: imap imaps listen: * ssl_cert_file: /opt/SSL/server.pem ssl_key_file: /opt/SSL/server.pem login_dir: /var/run/dovecot/login login_executable: /usr/libexec/dovecot/imap-login login_process_per_connection: no mail_location: maildir:~/Maildir:INDEX=/var/spool/dovecot/indexes/%1u/%u mmap_disable: yes namespace: type: private separator: . prefix: INBOX. inbox: yes auth default: cache_size: 4096 cache_ttl: 7200 passdb: driver: pam args: cache_key=%u dovecot userdb: driver: passwd args: blocking=yes == end == -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] 1.0.1: corrupt index on fresh mailbox, ideas?
Timo Sirainen wrote: So the real reasons for these could include temporarily using different index file paths, restoring indexes/uidlists from backups, temporarily using index files for accessing other mailbox. OK I understand your 3 points about how it actually could happen, and the ideas above as to possible causes; unfortunately we are in none of those situations. :( It was a mailbox simply created, accessed via Thunderbird and Squirrelmail all of... maybe 3 times. Nothing at all changed between the time it was working and the time it wasn't (no other mailboxes were affected). I'll keep an eye out - I created another new mailbox yesterday which will get the exact same treatment for an employee starting next Monday. If it doesn't happen again I'll just resign to calling it a random happenstance. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] 1.0.1: corrupt index on fresh mailbox, ideas?
Timo Sirainen wrote: On Tue, 2007-07-10 at 13:28 -0400, Stewart Dean wrote: when will there be a respin for a V1.0.2? This week. Thank you! I'm not sure how I would test this for you, prior to 1.0.2 - do you need it tested somehow? regards, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] NFS cache flush tester
Timo Sirainen wrote: http://dovecot.org/tools/nfstest.c Latest version used (2007-07-12 ~11am PST) for tests. Machine 1: CentOS 4 (RHEL 4 clone) Linux 2.6.9-42.0.8.ELsmp #1 SMP Tue Jan 30 12:33:47 EST 2007 i686 i686 i386 GNU/Linux Machine 2: CentOS 4 (RHEL 4 clone) Linux 2.6.9-42.0.10.ELsmp #1 SMP Tue Feb 27 10:11:19 EST 2007 i686 i686 i386 GNU/Linux NFS Server: SnapServer 4500 (linux software RAID based appliance) Linux 2.4.19-gos-up #1 Wed Jul 12 23:00:50 PDT 2006 i686 i686 i386 GNU/Linux machine 1: server machine 2: client = Info: Connected: client Info: O_EXCL works Info: Testing attribute cache.. Info: Attr cache flush fchown(-1, -1): failed Info: Attr cache flush fchown(uid, -1): OK Info: Attr cache flush fchmod(mode): OK Info: Attr cache flush chown(-1, -1): failed Info: Attr cache flush chown(uid, -1): OK Info: Attr cache flush chmod(mode): OK Info: Attr cache flush dup+close: OK Info: Testing write flushing.. Info: Write flush no caching: failed Info: Write flush fcntl(shared): OK Info: Write flush fcntl(exclusive): OK Info: Write flush flock(shared): failed Info: Write flush flock(exclusive): failed Info: Write flush reopen: OK Info: Write flush dup+close: OK Info: Write flush attr cache: failed Info: Testing data cache.. Info: data cache: Reading EOF requires attribute cache flush Info: Data cache flush no caching: failed Info: Data cache flush attr cache: failed Info: Data cache flush fcntl(shared): OK Info: Data cache flush fcntl(exclusive): OK Info: Data cache flush flock(shared): failed Info: Data cache flush flock(exclusive): failed Info: Data cache flush dotlock: failed Info: Data cache flush O_DIRECT: OK Info: Data cache flush dup+close: OK machine 1: client machine 2: server = Info: Connected: client Info: O_EXCL works Info: Testing attribute cache.. Info: Attr cache flush fchown(-1, -1): failed Info: Attr cache flush fchown(uid, -1): OK Info: Attr cache flush fchmod(mode): OK Info: Attr cache flush chown(-1, -1): failed Info: Attr cache flush chown(uid, -1): OK Info: Attr cache flush chmod(mode): OK Info: Attr cache flush dup+close: OK Info: Testing write flushing.. Info: Write flush no caching: failed Info: Write flush fcntl(shared): OK Info: Write flush fcntl(exclusive): OK Info: Write flush flock(shared): failed Info: Write flush flock(exclusive): failed Info: Write flush reopen: OK Info: Write flush dup+close: OK Info: Write flush attr cache: failed Info: Testing data cache.. Info: data cache: Reading EOF requires attribute cache flush Info: Data cache flush no caching: failed Info: Data cache flush attr cache: failed Info: Data cache flush fcntl(shared): OK Info: Data cache flush fcntl(exclusive): OK Info: Data cache flush flock(shared): failed Info: Data cache flush flock(exclusive): failed Info: Data cache flush dotlock: failed Info: Data cache flush O_DIRECT: OK Info: Data cache flush dup+close: OK -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Proxy problems with 1.0.5 !
Ed W wrote: All I'm getting from Thunderbird is an "Invalid Command" response. Is there an easy way to get a look inside the imap stream to try and see what's failing? From the client end of things: http://wiki.dovecot.org/Debugging/Thunderbird -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Running a dovecot IMAPS server
Timothy Murphy wrote: It could therefore be described as a "normal" maildir format; and if dovecot does not like this format, I think this should be explained clearly in the dovecot documentation. [I didn't find the Maildir vs Maildir++ account very illuminating.] I'm jumping in late to the game (sorry missed your original post somehow); my personal opinion here is that kmail is doing what I would call non-standard Maildir-ing, and it's causing you confusion with what most of us learn about Maildir in our travels. From what I understand via the spec and all Timo's posts, the actual folder format is never laid down in stone and is open to each implementation. However the default/accepted Maildir format as used by all the big boys (courier, dovecot, exim, postfix, etc.) is: ~/Maildir/.Family/ ~/Maildir/.Friends/ The Maildir hierarchy that kmail is giving you is not the ad-hoc accepted norm; they chose to do away with the leading '.' character it appears (I don't use kmail). The dot-prefix format is what I call normal, kmail seems abnormal (in the big picture, but kmail is doing nothing wrong per spec!) From your above statement you believe it to be the reverse and that dovecot doesn't like the format -- it's not that, it's that dovecot was built to work out of the box with the accepted global norm which is a dot-prefix Maildir hierarchy. With v1.1 you can do this with: mail_location = maildir:~/Maildir:LAYOUT=fs I'm not clear what this means. I believe Timo means that in the latest code (1.1 is alpha status) he has added new features, and one of those features is to use the filesystem separator for folders. Since '/' is the standard on *nix, then your kmail format would magically be recognized and used without it's dot-prefix. 1.0 dot-prefix: ~/Maildir/.Family.Marge/ 1.1 LAYOUT=fs: ~/Maildir/Family/Marge/ Why do you want it to work like that? Can't you just use the Maildir++ layout and use the email only via IMAP? I would be quite happy to do this, but when I tried briefly re-naming ~/Maildir/Family/ to ~/Maildir/.Family/ on my server and re-started dovecot it did not seem to work - I did not see the Family folder on my client (using IMAPS). You went a little too fast. :) After renaming the folder to .Family, use your email client to subscribe to the folder; I don't use kmail but in Thunderbird you rightmouse click on Inbox and choose Subscribe from the popup menu. Technically this places the name of the subscribed folder in a file named 'subscriptions' in the ~/Maildir/ directory on the server. Dovecot only presents the folders listed in this subscriptions file to the client when the client asks for a list of folders. IMAP allows folders present on the server which are not presented to the client when it logs in, hence the idea of subscribing. Unsubscribed folders are great for archiving old stuff that you don't really need to see but need to keep around. By not subscribing after you renamed the folder to include a dot it remained invisible to your client. hth, -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Running a dovecot IMAPS server
Charles Marcus wrote: Does KMail actually work this way? This would be enough to make me try it out... although my day to day WS is a windows box, so I'd most likely wait until KDE4 is available on Windows... Offtopic to the left -- before I switched to using GMail for my personal mail, I used to use Sylpheed ( http://sylpheed.sraoss.jp/en/ ) and it stores messages in a MH format (a file-per-message design somewhat like Maildir) and it can import mbox files from Thunderbird. The client itself is very nice too, I remember really liking it a lot. Bonus: Win32 binaries are also available. hth, -te -- Troy Engel | Systems Engineer Fluid Inc. | http://www.fluid.com
Re: [Dovecot] Running a dovecot IMAPS server
FiL @ Kpoxa wrote: I know this is totally unrelated to this topic, but... Is there any extension to create and use templates? I don't like "name> wrote:" in my reply messages and would prefer some custom templates, that would be different for different accounts. This might not be exactly what you need but then again it might be - I use an extension named Clippings: https://addons.mozilla.org/en-US/thunderbird/addon/1347 When you're in the compose window, rightclick on an empty space in the text editing area and choose Clippings. I have templates for all sorts of things (new hire welcome message with company info, for instance) that save me a lot of time. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] filtering when mail is moved to IMAP Inbox
Scott Silva wrote: Might be easier to have fetchmail get the mails from gmail and then you can process them any way you want. You can write procmail rules, or sieve scripts, or whatever you want. +1 - you're using a hammer when you need a screwdriver. Use something fetchmail-esque to POP out your email instead of TBird. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] --enable-headers-install problem in 1.0.5
Trever L. Adams wrote: > I am trying to get a SPEC file for an RPM to build a devel package. I > have it working, except for one thing. 1.0.5 doesn't actually install > the headers with that option to configure found in the subject. See Axel's SPEC file, he installs by hand: http://dl.atrpms.net/all/dovecot.spec (look for "# devel files" for the specific section) -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] Dovecot hanging up with many defunct processes
Mayank Joshi wrote: passdb: driver: pam userdb: driver: passwd 4) Dovecot's PAM file (/etc/pam.d/dovecot) #%PAM-1.0 authrequired pam_ldap.so account required pam_ldap.so session required pam_mkhomedir.so skel=/etc/skel umask=0077 session required pam_ldap.so Our situation is similar (but not exact) to yours - I'm authenticating against LDAP though and not AD2003. Perhaps one of these settings I use may help: dovecot.conf: login_process_per_connection: no auth default: cache_size: 4096 cache_ttl: 7200 passdb: driver: pam args: cache_key=%u dovecot userdb: driver: passwd args: blocking=yes The blocking=yes in the userdb might be the first thing you try, see these wiki pages for the reasons I have it like this: http://wiki.dovecot.org/UserDatabase/NSS http://wiki.dovecot.org/AuthDatabase/Passwd Additionally my pam.d dovecot uses the system stack, which *then* uses nss_ldap on it's own; I don't specifically bind pam_ldap into the dovecot pam.d file: # cat /etc/pam.d/dovecot #%PAM-1.0 auth required pam_nologin.so auth required pam_stack.so service=system-auth accountrequired pam_stack.so service=system-auth sessionrequired pam_stack.so service=system-auth In this case the file /etc/nsswitch.conf is controlling my authentication mechanism (LDAP), and the pam_ldap.so stack entry is located in /etc/pam.d/system-auth (these are all Red Hat defaults out of the box). Hope some of this helps, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com
Re: [Dovecot] umask not applied
Thomas wrote: I searched the wiki and googled but couldn't find anything helpful. Do you have any ideas? I'll be the first to mention that 1.0rc15 is ancient by dovecot standards, upgrade to 1.0x first and see if it's still broken. -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com