- dovecot - dovecot.org

[Dovecot] permissions on some dovecot directories
by Arkadiusz Miskiewicz 19 May '07

19 May '07

Hello, What smallest permissions are needed on directories: /var/run/dovecot /var/run/dovecot/login /var/lib/dovecot dovecot run as dovecot:dovecot. Currently I have 775 dovecot:dovecot /var/run/dovecot 750 root:dovecot /var/run/dovecot/login 755 root:root /var/lib/dovecot -- Arkadiusz Miśkiewicz PLD/Linux Team arekm / maven.pl http://ftp.pld-linux.org/

1 0

[Dovecot] Maildir subfolder conversion from courier-imap
by Remy Blank 19 May '07

19 May '07

Hello, I am trying to convert a maildir that was managed by courier-imap up to now to dovecot. I am using the courier-dovecot-migrate.pl script as follows: courier-dovecot-migrate.pl --convert --recursive ~/Maildir However, this command only seems to convert the base maildir, but not all the subfolders. I thought that that's what the --recursive option was for. Is the command above supposed to convert the base maildir as well as the subfolders, or should I run the command on all subfolders separately (e.g. using a shell script)? Thanks. -- Remy

2 2

Re: [Dovecot] Replication plans
by Bill Boebel 19 May '07

19 May '07

On Fri, May 18, 2007 1:42 am, Troy Benjegerdes <hozer(a)hozed.org> said: > I'm going to throw out a warning that it's my feeling that replication > has ended many otherwise worthwhile projects. Once you go down that > rabbit hole, you end up finding out the hard way that you just can't > avoid the stability, performance, complexity, and whatever problems. > .. > I've found myself pretty much in the same "all roads lead to the > filesystem" situation. I don't want to replicate just imap, I want to > replicate the build directory with my source code, my email, and my MP3 > files. One of the problems with the clustered file system approach seems to be that accessing Dovecot's index, cache and control files are slow over the network. For speed, ideally you want your index, cache and control on local disk... but still replicated to another server. So what about tackling this replication problem from a different angle... Make it Dovecot's job to replicate the index and control files between servers, and make it the file system's job to replicate just the mail data. This would require further disconnecting the index and control files from the mail data, so that there is less syncing required. i.e. remove the need to check directory mtimes and to compare directory listings against the index; and instead assume that the indexes are always correct. Periodically you could still check to see if a sync is needed, but you'd do this must less frequently. I agree that there are already great solutions available for replicated storage, so this would allow us to take advantage of these solutions for the bulk of our storage without impacting the speed of IMAP. Bill

3 2

Re: [Dovecot] dbox redesign
by Bill Boebel 19 May '07

19 May '07

On Sat, May 12, 2007 9:10 am, Timo Sirainen <tss(a)iki.fi> said: > Fast copying > ============ > > Would be nice if copying a message from one mailbox to another wouldn't > require actually reading+writing the whole message contents. But I can't > really figure out how to implement this without requiring that there is > only a single dbox storage which contains the mails for all the > mailboxes, and the mailboxes themselves are just Dovecot's index files > containing pointers to the dbox storage. > > The problem with having everything in one storage is that if the index > files are broken, the messages can't be placed into correct mailboxes > anymore. > > Although one possibility would be treat mailboxes a bit similarly than > keywords. So that when a message is copied to another mailbox, the > message in dbox file is updated to contain information that it exists in > such and such mailboxes. Hmm. Perhaps that would be good enough, yes. > Yes, I think treating mailboxes similary to keywords is ideal. There really is no reason to physically separate mailboxes on disk. All that is needed is this logical separation if it can be done in a reliable way. Or maybe track this in mailbox-specific index files, and also have a corespodning text file that stores a list of messages that are contained in that mailbox... similar to maildir's dovecot-uidlist file. Then if you lose the index you can rebuild the index from the text file. Bill

4 7

[Dovecot] dovecot-shared being ignored?
by bhayden＠umn.edu 19 May '07

19 May '07

So, I'm working on a shared folder. Everything is nice except for one bit. My dovecot-shared file in the Maildir is either being ignored, or doesn't work the way I think it does. Any messages copied into the Maildir using a mail client are mode 660 and owned by the user who did the copy. Perms on the dovecot-shared file are 700 (I'm using Solaris ACLs in addition to the Dovecot ACLs in order to prevent unauthorized access via means that don't pass through Dovecot's IMAP server... ie SFTP/SCP, and so I'm using 700 with a rwx mask and access to the file set per user in the Solaris ACL). The message files being created 660 means they are accessible via SFTP/SCP to anyone in the group, and we have a single one for (nearly) all users. This is not really acceptable, so if anyone can offer guidance on the dovecot-shared that'd be great. I checked the wiki, and it has very minimal information. -Brian

2 1

[Dovecot] Virtual mailbox plans
by Timo Sirainen 18 May '07

18 May '07

Configuration ------------- There could be global and user-specific configuration files, similar to how ACLs work. I think the global virtual mailboxes should be only defaults though, so that users could delete them and create a new mailbox (virtual or non-virtual) with the same name. The global virtual mailboxes could be described in a single file, such as: Trash deleted Work/My Unseen (mailboxes "box1" "box2" "box2/*") or to "tss(a)iki.fi" to "timo.sirainen(a)iki.fi" unseen So the format would be <virtual mailbox name> LF <search condition, allowed to wrap to new line> LF LF <next virtual mailbox, etc.> Per-user virtual mailboxes could also be listed in a similar file. If the search condition is empty, a globally defined virtual mailbox has been deleted. Another way would be to have per-mailbox dovecot-virtual files, but I'm not sure how global virtual mailboxes could then be easily deleted. Also having all of them in a single file makes listing the mailboxes pretty fast. The search condition would be a standard IMAP SEARCH reply, with the exception of the mailboxes parameter to specify what mailboxes are used. Virtual mailboxes would be implemented as a new mail storage backend, so their namespace configuration would be like: namespace private { prefix = "virtual/" location = virtual:~/virtual:GLOBAL=/etc/dovecot/virtual-globals.conf } Syncing ------- ~/virtual directory would contain index files for the virtual mailboxes. If the index files don't exist, a full resync for the virtual mailbox is done. This means opening all the real mailboxes configured for the vbox, running the SEARCH command and adding the returned UIDs to the vbox's indexes. The index would contain at least mailbox ID and message UID numbers. Probably would be a good idea to also save UIDVALIDITY numbers just to be safe, although that's a bit of a waste of space and it could be handled in other ways too, such as changing the mailbox ID whenever its UIDVALIDITY changes.. Mailbox IDs would be mapped to names in mailbox list index. This index already exists, although there still seems to be some bugs in it. At least creating a new list index has a race condition. I should look into making the list index a bit more realiable. Also the list indexes work only with Maildir++ layout currently. After the vbox has initially been synced, it needs to also handle updates. I think optionally the updates could be done the same way as full sync (opening all mailboxes to see if there are changes), but a faster way would be to update the virtual mailboxes immediately when the real mailboxes are being modified. So IMAP, POP3 and deliver all need to be aware of the virtual mailboxes and what real mailboxes are mapped to what virtual ones. Whenever doing a change that causes the message to either be added to or removed from a virtual mailbox according to their rules, the virtual mailbox's index is updated immediately. Being able to determine this change quickly might be the biggest change required to Dovecot's base code. Changes ------- Virtual mailboxes would share their state with the real mailbox, so any flag changes done in the vbox would be immediately changed in the real mailbox as well. The same goes for expunges. Virtual mailboxes' transaction logs would use different records for expunges and removals. Expunges would be handled normally, but removals wouldn't be removed from the in-memory view of the mailbox until EXPUNGE command was given (although they would be removed from dovecot.index immediately, so new sessions wouldn't see them). This makes it possible to have a vbox with only unseen messages, but without the message getting expunged immediately after a client opens the mail. So the EXPUNGE command could act as a "session synchronization" command. And of course it would also expunge the messages that had been marked as \Deleted.

1 0

[Dovecot] deliver exit codes?
by Ralf Hildebrandt 18 May '07

18 May '07

Before switching to dovecot deliver for all local deliveries, I'm calling deliver from maildrop (yes, don't laugh). Snippet from my global /etc/maildroprc: # Store maillist & junk messages directly /^Precedence: ?list/ if ( $MATCH ) { exception { log "Bulk message: $MATCH" #to "$DEFAULT" to "|$DELIVER" } } I replaced to "$DEFAULT" (meaning: delivery to $HOME/Maildir) with to "|$DELIVER" (meaning: pipe the mail into dovecot deliver, and deliver to INBOX = $HOME/Maildir) This works really well, UNLESS a user exceeds his/her quota: In that particular case, deliver fails to deliver the mail and generates a bounce, all on it's own. It's exit status is 0 in that case, so my MTA (postfix) thinks that the mail has been delivered successfully. I'd prefer deliver to return EX_TEMPFAIL, which I can pass through maildrop back to postfix, which will then log a temporary error and will bounce the mail all by itself after 5 days time. But how? -- Ralf Hildebrandt (Ralf.Hildebrandt(a)charite.de) plonk(a)charite.de Postfix - Einrichtung, Betrieb und Wartung Tel. +49 (0)30-450 570-155 http://www.arschkrebs.de The Imperial Vendor or Contractor probably told them that droid armies don't need redundant command centers, nor any kind of high availability ... you're supposed to buy several, and cluster them. Redundant Array of Inexpensive Droids?

2 1

[Dovecot] Disable plaintext authentications
by Alexander Moisseev 18 May '07

18 May '07

Dear Sirs, I have Postfix 2.5 with Dovecot 1.0.0 as authentication server. Can I disable SMTP plaintext authentications for WAN clients, but allow it for LAN ones?

2 1

[Dovecot] UID larger than next_uid in file /hosts/domain/user/dovecot-uidlist
by Chris Jeans 18 May '07

18 May '07

I'm having the following problem and would appreciate any help or tips on how to fix it. None of my users can send email via pop3 or imap clients the only consistent thing that i keep seeing in the logs are the following 2 messages: UID larger than next_uid in file /hosts/domain/user/dovecot-uidlist sync: UID inserted in the middle of mailbox After searching the web all I can find is this link http://dovecot.org/list/dovecot-cvs/2007-May/008819.html which deals with Index: maildir-uidlist.c I'm running dovecot 1.0.0 on FreeBSD 6.2. -- Chris Jeans cmjeans(a)gmail.com

2 1

Re: [Dovecot] Replication plans
by Bill Boebel 18 May '07

18 May '07

On Fri, May 18, 2007 1:10 pm, Timo Sirainen <tss(a)iki.fi> said: > On Fri, 2007-05-18 at 12:20 -0400, Bill Boebel wrote: >> So what about tackling this replication problem from a different >> angle... Make it Dovecot's job to replicate the index and control >> files between servers, and make it the file system's job to replicate >> just the mail data. This would require further disconnecting the >> index and control files from the mail data, so that there is less >> syncing required. i.e. remove the need to check directory mtimes and >> to compare directory listings against the index; and instead assume >> that the indexes are always correct. Periodically you could still >> check to see if a sync is needed, but you'd do this must less >> frequently. > > This would practically mean that you want either cydir or dbox storage. > > This kind of a hybrid replication / clustered filesystem implementation > would simplify the replication a bit, but all the difficult things > related to UID conflicts etc. will still be there. So I wouldn't mind > implementing this, but I think implementing the message content sending > via TCP socket wouldn't add much more code anymore. > > The clustered filesystem could probably be used to simplify some things > though, such as UID allocation could be done by renaming a "uid-<next > uid>" file. If the rename() succeeded, you allocated the UID, otherwise > someone else did and you'll have to find the new filename and try again. > But I'm not sure if this kind of a special-case handling would be good. > Unless of course I decide to use the same thing for non-replicated > cydir/dbox. Dbox definitely sounds promising. I'd avoid putting this uid-<nextuid> file in the same location as the mail storage though, because if you can truly keep mail storage isolated and infrequently accessed, then you can do cool things like store your mail data remotely on Amazon S3 or equivalent.

1 0