- dovecot - dovecot.org

Definitive guide to running ObjectiveFS on top of a Clustered File System?
by Roel van Duijnhoven 06 Jun '21

06 Jun '21

Hey there, I am working on migrating a Dovecot system with + 50.000 users / 5TB on mail data to a new set-up that is better scalable. But I found it difficult to find a good solution. In this email I lay out what I learned, in the hope that others can profit. But also others can help me on this design before I actually start migrating. *My aim is to get a clustered file system setup for Dovecot. But with the additional constraint of not having to pay on a per-mailbox basis to some vendor. * Why a clustered file system? I believe this yields the best solution. Your machines are now stateless: consuming services. And scaling or recovery is now relatively easy. Why the constraint of not paying per-mailbox? If you did pay a fee per mailbox; you probably would be better of just using a managed mail service altogether. High-level ======== We are going to use ObjectiveFS as our clustered file system (not involved with them in any way). They use Fuse to make it behave like a POSIX filesystem. Due to the way they compact files that belong together it handles a large number of small files relatively fine. The great thing is: we can mount this filesystem on multiple machines. This piece of software is proprietary, but it's price is very reasonable, and it is a fixed monthly cost. We can use S3, Google Cloud Storage, or others as the actual storage system. We store metadata about the mailboxes in a database that is available to all machines. We can use any managed database service for this. We explicitly set the proxy_host for each user in this database. Thus every user gets routed to the same machine. We make sure to distributed the users among the available machines. All traffic is received by a TCP load balancer. This will forward the traffic to the underlying machines. Ideally this is a managed cloud service. I think there are two valid choices. But I prefer the first one: - Use a TCP pass-through load balancer. This way the source IP is retained without requiring any change to the protocol (an example of this is Google Cloud external TCP/UDP Network Load Balancing <https://cloud.google.com/load-balancing/docs/network>). - Use an TCP SSL terminating load balancer. And make it forward in HaProxy's PROXY format so we retain the source IP. That traffic is now routed to N-machines. Each of those machines runs Dovecot. And Dovecot is configured with Dovecot Proxy (not to be confused with HaProxy's PROXY protocol). Such that Dovecot will proxy the connection to another machine if required (or continue on this machine if you are already on the correct proxy_host). I believe this can be done using proxy_maybe <https://doc.dovecot.org/configuration_manual/authentication/proxies/>. Depending on the choice of load balancer you may need to configure Dovecot to understand HaProxy's PROXY format. We are going to use the maildir format. And store those files on the ObjectiveFS filesystem. The index to Dovecot is stored on a local SSD. Choice motivations =============== ObjectiveFS: - Can probably be another clustered file system as well. - But ObjectiveFS works quite well because it has support for handling many small files. And can be configured with a local SSD disk. - Using this software can get a clustered file system to run quite easy. Regarding the mailbox format: - dbox is said to be more performant. And having less files is an advantage in general. But one big disadvantage: because the index contains crucial information, you will need to place this index on the Clustered File System as well. Maildir does not have this problem, and that is why we prefer it. - obox is only available as Dovecot Pro, and solves a lot of problems. But you really can only get Dovecot Pro if you pay a fee per mailbox. We don't want that! Why don't use Dovecot Director as well? - You can not run Dovecot Director on the same machine as your Dovecot Backend / proxy. Thus we would need to introduce 2 additional hosts to just accept and route the traffic. This would increase the complexity of our solution. - But it could be build on top of this set-up fairly easy. - I don't see too many advantages for Director however currently. - One big advantage *could* be if it would automatically remove failed nodes. However this only works if you use Dovemon <https://doc.dovecot.org/configuration_manual/dovemon/>. And that is packaged with OX/Dovecot and thus requires paying per user. - So in case you want to put a node down you get a slightly easier API to do so. Namely *doveadm director add <backend server ip> 0*. However we can simply imitate this by updating our database and changing the proxy_host to another destination as well. Load balancer - You could also set-up your MX records to directly point to your Dovecot nodes. This should work as well I guess. - But I think the load balancer makes for better control, over changing and managing DNS. Questions ======== In general I'd like to know whether this is a good idea! But some other questions I have at this point that could make or break this setup: - Can a Dovecot backend instance be configured to *also* run Dovecot proxy? - Can a Dovecot Proxy node receive HaProxy PROXY traffic? Details ===== Once I am sure I am on the right track I'd like to post the settings I use for the software as well. But let's first discuss the stuff above! Best, Roel van Duijnhoven

2 1

possible minor bug: lmtp excessice newlines
by Laura Steynes 06 Jun '21

06 Jun '21

Aki, unnecessary newlines in lmtp received header? Received: from mail6.xxxxxxxxxx.com by mail6.xxxxxxx.com with LMTP id xxxxxxxxxxxxxxxxxx (envelope-from <xxxxxx(a)xxxxxxxx.net>) for <xxx(a)xxxxxxxx.net>; Sat, 05 Jun 2021 14:19:07 +0100 is this intentional Aki ? Evident is multiple clients, in widescreen, so is inot client squashing the header # Dovecot version 2.3.14 # Pigeonhole version 0.5.14 # OS: Linux 5.10.41 x86_64

1 0

Panic on indexer-worker and Dovecot stop to search on Solr
by Alessio Cecchi 03 Jun '21

03 Jun '21

Hi, I have setup on a (little busy) Dovecot server FTS with Solr and Virtual folder to enable "search in all folders" for users. All works fine until for some users the indexer-worker process crash. After this crash Dovecot stop to query Solr for new search in BODY, returning SERVERBUG, for all users on the server and only with a dovecot restart users can search in BODY again. Before the dovecot restart I can see that on Solr server no query from dovecot arriving anymore (for SEARCH BODY because SEARCH in FROM/TO/SUBJECT still works fine) Dovecot version is 2.3.14 (from RPM) and Apache Solr si 7.7.3. Here is the error found in the log: May 31 14:58:13 Panic: indexer-worker(user(a)email.com) session=<VA0oxZ/DXtG5Ydlx:0KzOIkLdtGDSawAAxbYlJw>: file http-client-request.c: line 1240 (http_client_request_send_more): assertion failed: (req->payload_input != NULL) May 31 14:58:13 Error: indexer-worker(user(a)email.com) session=<VA0oxZ/DXtG5Ydlx:0KzOIkLdtGDSawAAxbYlJw>: Raw backtrace: /usr/lib64/dovecot/libdovecot.so.0(backtrace_append+0x42) [0x7f3d91b2cac2] -> /usr/lib64/dovecot/libdovecot.so.0(backtrace_get+0x1e) [0x7f3d91b2cbce] -> /usr/lib64/dovecot/libdovecot.so.0(+0xf3cde) [0x7f3d91b39cde] -> /usr/lib64/dovecot/libdovecot.so.0(+0xf3d81) [0x7f3d91b39d81] -> /usr/lib64/dovecot/libdovecot.so.0(i_fatal+0) [0x7f3d91a8b25a] -> /usr/lib64/dovecot/libdovecot.so.0(http_client_request_send_more+0x3dd) [0x7f3d91ad39ad] -> /usr/lib64/dovecot/libdovecot.so.0(http_client_connection_output+0xf1) [0x7f3d91ad8101] -> /usr/lib64/dovecot/libdovecot.so.0(+0x11d2b0) [0x7f3d91b632b0] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_call_io+0x65) [0x7f3d91b51f45] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x12b) [0x7f3d91b538fb] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0x59) [0x7f3d91b52049] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0x38) [0x7f3d91b52288] -> /usr/lib64/dovecot/libdovecot.so.0(http_client_wait+0xcd) [0x7f3d91ae120d] -> /usr/lib64/dovecot/lib21_fts_solr_plugin.so(solr_connection_select+0xe4) [0x7f3d8fca1174] -> /usr/lib64/dovecot/lib21_fts_solr_plugin.so(+0x45d4) [0x7f3d8fc9d5d4] -> /usr/lib64/dovecot/lib20_fts_plugin.so(fts_backend_get_last_uid+0x6e) [0x7f3d90e103fe] -> /usr/lib64/dovecot/lib20_fts_plugin.so(+0xf952) [0x7f3d90e16952] -> /usr/lib64/dovecot/lib20_fts_plugin.so(+0x10ef6) [0x7f3d90e17ef6] -> /usr/lib64/dovecot/lib20_virtual_plugin.so(+0x966a) [0x7f3d90bfc66a] -> /usr/lib64/dovecot/lib20_fts_plugin.so(+0x10ba6) [0x7f3d90e17ba6] -> /usr/lib64/dovecot/libdovecot-storage.so.0(mail_precache+0x2e) [0x7f3d91e3cb0e] -> dovecot/indexer-worker [user(a)email.com Virtual/All - 5300/20832](+0x2924) [0x5648464be924] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_call_io+0x65) [0x7f3d91b51f45] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x12b) [0x7f3d91b538fb] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0x59) [0x7f3d91b52049] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0x38) [0x7f3d91b52288] -> /usr/lib64/dovecot/libdovecot.so.0(master_service_run+0x13) [0x7f3d91abcbb3] -> dovecot/indexer-worker [user(a)email.com Virtual/All - 5300/20832](main+0xd7) [0x5648464be1f7] -> /lib64/libc.so.6(__libc_start_main+0xf5) [0x7f3d9169a555] -> dovecot/indexer-worker [user(a)email.com Virtual/All - 5300/20832](+0x22ba) [0x5648464be2ba] May 31 14:58:13 Fatal: indexer-worker(user(a)email.com) session=<VA0oxZ/DXtG5Ydlx:0KzOIkLdtGDSawAAxbYlJw>: master: service(indexer-worker): child 27602 killed with signal 6 (core dumps disabled - https://dovecot.org/bugreport.html#coredumps) on Solr side I found this error log: org.apache.solr.common.SolrException: Early EOF at org.apache.solr.handler.loader.XMLLoader.load(XMLLoader.java:190) at org.apache.solr.handler.UpdateRequestHandler$1.load(UpdateRequestHandler.java:97) at org.apache.solr.handler.ContentStreamHandlerBase.handleRequestBody(ContentStreamHandlerBase.java:68) at org.apache.solr.handler.RequestHandlerBase.handleRequest(RequestHandlerBase.java:199) at org.apache.solr.core.SolrCore.execute(SolrCore.java:2551) at org.apache.solr.servlet.HttpSolrCall.execute(HttpSolrCall.java:733) at org.apache.solr.servlet.HttpSolrCall.call(HttpSolrCall.java:531) at org.apache.solr.servlet.SolrDispatchFilter.doFilter(SolrDispatchFilter.java:395) [...] and [com.ctc.wstx.exc.WstxLazyException] com.ctc.wstx.exc.WstxIOException: Early EOF at com.ctc.wstx.exc.WstxLazyException.throwLazily(WstxLazyException.java:45) at com.ctc.wstx.sr.StreamScanner.throwLazyError(StreamScanner.java:728) at com.ctc.wstx.sr.BasicStreamReader.safeFinishToken(BasicStreamReader.java:3678) at com.ctc.wstx.sr.BasicStreamReader.getText(BasicStreamReader.java:860) at org.apache.solr.handler.loader.XMLLoader.readDoc(XMLLoader.java:419) at org.apache.solr.handler.loader.XMLLoader.processUpdate(XMLLoader.java:260) at org.apache.solr.handler.loader.XMLLoader.load(XMLLoader.java:188) at org.apache.solr.handler.UpdateRequestHandler$1.load(UpdateRequestHandler.java:97) This is the relevant part of dovecot configuration: plugin { fts = solr fts_autoindex = yes fts_decoder = decode2text fts_enforced = body fts_index_timeout = 5s fts_solr = url=http://10.0.1.3:8983/solr/dovecot/ [...] } Is this an already know bug? Any workaround? Thanks -- Alessio Cecchi Postmaster @ http://www.qboxmail.it https://www.linkedin.com/in/alessice

2 4

Dovecot Stats for Logwatch
by Plutocrat 03 Jun '21

03 Jun '21

Hi all, I run logwatch on a few servers and the script on there which parses the dovecot logs provides waaaaaay too much detail, and contains a bunch of stuff I'm not interested in. There's not much point in having a log summary if its 100Kb long! The one for exim is also terrible, but I've replaced that with eximstats, and a bit of bash scripting and have something I can work with. I might at some point look at replacing that too. Now to my question. I haven't found anything in the dovecot family which compares to eximstats. I was wondering if there possibly is a report tool, or a way of extracting stats from doveadm just for key metrics such as 'number of POP logins in the last 24 hours', number of failed passwords, etc. Things that might be interesting on a logwatch report. Or maybe someone has written a script already that can be used in logwatch in perl, python, bash, whatever. Any pointers, suggestions, or ideas for useful summary metrics to pull out of the logfiles are appreciated. I suspect I might end up writing one myself anyway, but I'm just keen to see what already exists, get the benefit of anyone else's experiences in this matter before I start. P.

3 8

fts_autoindex_exclude for everything in a namespace?
by Daniel J. Luke 03 Jun '21

03 Jun '21

Hi, I've recently moved a system from using procmail to delivering via dovecot-lmtp w/ sieve - and it's great! I've got a copy of incoming mail going into mbox archives (that I compress after months end), and my mail log is full of "Error: fts: Failed to sync mailbox #Archive: Mailbox isn't selectable", though. I /think/ fts_autoindex_exclude = #Archive/* is what I want, but it doesn't seem to silence the error. I don't need the #Archive mailboxes visible to clients, but I don't mind if they are. I tried some setting the archive prefix = AutoArchive and setting list = yes, but that didn't silence the error. I feel like I'm probably missing something obvious :) dovecot 2.3.14 / pigenohole 0.5.14 relevant config from dovecot -n: mail_location = mbox:~/Mail/:INBOX=~/.mbox:INDEX=~/.dovecot-indexes mail_plugins = fts fts_lucene zlib namespace { inbox = yes location = mailbox Drafts { auto = no special_use = \Drafts } mailbox Sent { auto = subscribe special_use = \Sent } mailbox Trash { auto = no special_use = \Trash } mailbox spam { auto = create special_use = \Junk } prefix = separator = / } namespace archive { inbox = no list = no location = mbox:~/MailArchive/ prefix = "#Archive/" separator = / } plugin { fts = lucene fts_autoindex = yes fts_lucene = whitespace_chars=@. recipient_delimiter = + sieve = file:~/.sieve;active=~/.dovecot.sieve sieve_default = /usr/local/etc/dovecot/sieve/default.sieve sieve_default_name = file_spam } and the archive sieve script (which works) looks like this: if currentdate :matches "month" "*" { set "month" "${1}"; } if currentdate :matches "year" "*" { set "year" "${1}"; } if true { fileinto :copy :create "#Archive/${year}-${month}"; } -- Daniel J. Luke

1 0

userdb volatiledir setting not working
by Александер Скобельцын 02 Jun '21

02 Jun '21

Hi all! I just can't deal with VOLATILEDIR userdb setting on my mailservers. For example dovecot-uidlist.lock files are still being created over our network mail store. So, several days ago I changed my user and passwd query setting to include this option and restarted all my dovecot servers. Now it looks like (we use mysql driver): user_query = \ SELECT \ `u`.`pop` AS `home`, \ `u`.`uid`, \ `u`.`gid`, \ CONCAT('maildir:', `u`.`smtp`, ':INDEX=/var/dovecot_indexes/%d/%u:VOLATILEDIR=/var/dovecot_indexes/%d/%u') AS `mail`, \ CONCAT('*:bytes=', `u`.`quota`, 'M') AS `quota_rule` ,\ CONCAT('url=http://', fts_solr, ':8983/solr/dovecot/') AS fts_solr \ FROM `users` AS `u` \ JOIN `domains` AS `d` ON `d`.`domain_id` = `u`.`domain_id` \ WHERE \ `u`.`username` = '%u' \ AND `u`.`enabled` = 1 \ AND `d`.`enabled` = 1 \ AND `u`.`type` IN ('local', 'piped') password_query = \ SELECT \ `u`.`username` AS `user`, \ `u`.`crypt` AS `password`, \ `u`.`pop` AS `userdb_home`, \ `u`.`uid` AS `userdb_uid`, \ `u`.`gid` AS `userdb_gid`, \ CONCAT('maildir:', `u`.`smtp`, ':INDEX=/var/dovecot_indexes/%d/%u:VOLATILEDIR=/var/dovecot_indexes/%d/%u') AS `userdb_mail`, \ CONCAT('*:bytes=', `u`.`quota`, 'M') AS `userdb_quota_rule` \ FROM `users` AS `u` \ JOIN `domains` AS `d` ON `d`.`domain_id` = `u`.`domain_id` \ WHERE \ `u`.`username` = '%u' \ AND `u`.`enabled` = 1 \ AND `d`.`enabled` = 1 \ AND `u`.`type` IN ('local', 'piped') The aim was to move .lock and another temp files out of network storage our mails are stored. We use CephFS for storing user emails (maildir format). In addition we use 3 imap servers, each mount exact same path in cephfs and 2 dovecot directors to handle user=imapserver mapping, so user shouldn't exist on more than one imapserver at a time. Am I must use CONTROL directive in addition to force move control files from network store for VOLATILEDIR setting to work properly or I just misconfigured something? Extra info: dovecot --version 2.3.14 (cee3cbc0d)

1 0

debugging file permissions wrong
by Laura Steynes 02 Jun '21

02 Jun '21

Hi, In trying to debug a strange error where client can't login, I enabled all the usual debug settings, this is all good, it works for imap and pop3 fine, but the problem is when used with dovecot's LDA there is a nasty issue. the file created by debug_log_path in this case /var/log/dovecot/debug.log , this file created as root, again this is fine for nice logging of imap and pop3, but this causes postfix not to deliver mail, for write permission denied, LDA runs as vmail, all my sub sections like *_listener also = vmail and all my uid/gid settings are also to user/group vmail obviously you wont main log files to not be owned by vmail for case of security, so is there a way to set the ownership of the debug file - apart from the obvious of remembering 40 minutes later when you get alert of high mailq level to chown the file :) If there is no way, may the developers take this as a feature request please. Thanks Loz

2 3

New Configuration
by Eric Shields | Mass Transit Honchkrow 29 May '21

29 May '21

Hi, I'm setting up an email server behind a NAT, and I have the necessary DNS records, iptables/ufw rules, and ACLs in place. However, I cannot sign in with my chosen username, which does exist on the same computer I'm hosting the server on. I am using PAM, not a database, to hold credential information, and SSL certificates are active. However, I am getting signal 15 for the following processes: anvil, config, and auth. I can confirm through telnet that the ports I am using are open, and these ports are also open on the WAN side, which I believe is needed. Do you have any pointers for me, or any additional clarifying details? - 🚏

2 1

Different Maildir Locations
by Gabx 28 May '21

28 May '21

Hi, on my vps (debian stable) I have an email system with postfix, dovecot, virtual users and domains. I have 3 users with different mail_locations. the first mail_location = maildir: ~ / Maildir the second mail_location = maildir: ~ / user1 / Maildir the third mail_location = maildir: ~ / user2 / mail In postfix the above schema is represented in virtual_mailbox_maps. In dovecot mail_location is set maildir:~/Maildir. The first user receives regular mail, the other two none. I tried to see if in dovecot there was the possibility to configure the mail_location specifically for each single user, but with namespaces and userdb I am a bit confused and I have not found specific examples for my case. regards Gab -- GPG fingerprint: 5A6A F742 ACEA C9A7 93A8 8D5E 7AEF 5A02 D080 89F7 Contact: xmpp://gabx@jabber.tcpreset.net

3 14

replicator: Panic: data stack: Out of memory when allocating ...
by Jörg Faudin Schulz 28 May '21

28 May '21

frequently, the replicator fails on allocating memory: replicator: Panic: data stack: Out of memory when allocating ... bytes. I already have increased default_vsz_limit=1024M but this only increased the ... value. What can be done here? dovecot_branch=release-2.3.14

2 3