October 2017 - dovecot

Trash plugin unexpected results
by Stephan Herker 16 Oct '17

16 Oct '17

I have the trash plugin enabled and testing it out I had an account with a 30MB quota. In the accounts trash it had an email with a large attachment. I sent the same email again to the account expecting the trash plugin to purge the message from trash to make space for the new message in the inbox. However I got an error saying it couldn't remove enough trash to make space. This behavior seems unexpected because it seems like the trash plugin didn't account for the quota and size available. Debug logs and dovecot -n output below. Oct 07 17:33:13 lmtp(31543): Debug: none: root=, index=, indexpvt=, control=, inbox=, alt= Oct 07 17:33:13 lmtp(31543): Info: Connect from local Oct 07 17:33:13 lmtp(31543): Debug: Loading modules from directory: /usr/lib64/dovecot Oct 07 17:33:13 lmtp(31543): Debug: Module loaded: /usr/lib64/dovecot/lib10_quota_plugin.so Oct 07 17:33:13 lmtp(31543): Debug: Module loaded: /usr/lib64/dovecot/lib11_trash_plugin.so Oct 07 17:33:13 lmtp(31543): Debug: Module loaded: /usr/lib64/dovecot/lib20_autocreate_plugin.so Oct 07 17:33:13 lmtp(31543): Debug: Module loaded: /usr/lib64/dovecot/lib90_sieve_plugin.so Oct 07 17:33:13 lmtp(31543): Debug: auth input: 123(a)testdo.com home=/home/vmail/testdo.com/123 mail=maildir:/home/vmail/testdo.com/123 uid=2000 gid=2000 quota_rule=*:storage=30000 Oct 07 17:33:13 lmtp(31543): Debug: Added userdb setting: mail=maildir:/home/vmail/testdo.com/123 Oct 07 17:33:13 lmtp(31543): Debug: Added userdb setting: plugin/quota_rule=*:storage=30000 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Effective uid=2000, gid=2000, home=/home/vmail/testdo.com/123 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Quota root: name=User quota backend=dict args=:proxy::quotadict Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Quota rule: root=User quota mailbox=* bytes=30720000 messages=0 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Quota rule: root=User quota mailbox=Trash bytes=+104857600 messages=0 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Quota grace: root=User quota bytes=3072000 (10%) Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: dict quota: user=123(a)testdo.com, uri=proxy::quotadict, noenforcing=0 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: Namespace inbox: type=private, prefix=, sep=, inbox=yes, hidden=no, list=yes, subscriptions=yes location=maildir:/home/vmail/testdo.com/123 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: maildir++: root=/home/vmail/testdo.com/123, index=, indexpvt=, control=, inbox=/home/vmail/testdo.com/123, alt= Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: trash plugin: Added 'Spam' with priority 1 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: trash plugin: Added 'Trash' with priority 2 Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: sieve: Pigeonhole version 0.4.2 initializing Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: uFHMBEly2Vk3ewAAobcTNA: sieve: using the following location for user's Sieve script: /home/vmail/testdo.com/123/.dovecot.sieve;name=main script Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: uFHMBEly2Vk3ewAAobcTNA: sieve: loading script /home/vmail/testdo.com/123/.dovecot.sieve;name=main script Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: uFHMBEly2Vk3ewAAobcTNA: sieve: script binary /home/vmail/testdo.com/123/.dovecot.svbin successfully loaded Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: uFHMBEly2Vk3ewAAobcTNA: sieve: binary save: not saving binary /home/vmail/testdo.com/123/.dovecot.svbin, because it is already stored Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: uFHMBEly2Vk3ewAAobcTNA: sieve: executing script from /home/vmail/testdo.com/123/.dovecot.svbin Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Debug: trash plugin: Failed to remove enough messages (needed 17035766 bytes, expunged only 16803594 bytes) Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Info: uFHMBEly2Vk3ewAAobcTNA: sieve: msgid=<c5189e27-b2e3-bcc2-b34f-ecffa9d614cc(a)keagaming.com>: failed to store into mailbox 'INBOX': Quota exceeded (mailbox for user is full) Oct 07 17:33:13 lmtp(31543, 123(a)testdo.com) Info: uFHMBEly2Vk3ewAAobcTNA: sieve: execution of script /home/vmail/testdo.com/123/.dovecot.sieve;name=main script failed with unsuccessful implicit keep (user logfile /home/vmail/testdo.com/123/.dovecot.sieve.log may reveal additional details) Oct 07 17:33:13 lmtp(31543): Info: Disconnect from local: Successful quit # 2.2.10: /etc/dovecot/dovecot.conf # OS: Linux 3.10.0-514.26.2.el7.x86_64 x86_64 CentOS Linux release 7.4.1708 (Core) xfs auth_mechanisms = plain login dict { quotadict = mysql:/etc/dovecot/dovecot-dict-quota.conf } first_valid_uid = 2000 hostname = testdo.com last_valid_uid = 2000 log_path = /var/log/dovecot.log mail_debug = yes mail_gid = vmail mail_location = maildir:/home/vmail/%d/%n mail_plugins = " quota trash expire" mail_privileged_group = mail mail_uid = vmail managesieve_notify_capability = mailto managesieve_sieve_capability = fileinto reject envelope encoded-character vacation subaddress comparator-i;ascii-numeric relational regex imap4flags copy include variables body enotify environment mailbox date ihave mbox_write_locks = fcntl namespace inbox { inbox = yes location = mailbox Drafts { auto = subscribe special_use = \Drafts } mailbox Junk { auto = subscribe special_use = \Junk } mailbox Sent { auto = subscribe special_use = \Sent } mailbox "Sent Messages" { special_use = \Sent } mailbox Trash { auto = subscribe special_use = \Trash } prefix = } passdb { args = /etc/dovecot/dovecot-sql.conf.ext driver = sql } plugin { quota = dict:User quota::proxy::quotadict quota_grace = 10%% quota_rule2 = Trash:storage=+100M quota_status_nouser = DUNNO quota_status_overquota = 552 5.2.2 Mailbox is full quota_status_success = DUNNO sieve_dir = ~/sieve sieve_global_dir = /etc/dovecot/sieve/global/ sieve_global_path = /etc/dovecot/sieve/default.sieve trash = /etc/dovecot/dovecot-trash.conf.ext } postmaster_address = administrator(a)testdo.com protocols = imap lmtp sieve service auth-worker { user = vmail } service auth { unix_listener /var/spool/postfix/private/auth { group = postfix mode = 0666 user = postfix } unix_listener auth-client { group = postfix mode = 0660 user = postfix } unix_listener auth-master { group = vmail mode = 0660 user = vmail } unix_listener auth-userdb { mode = 0600 user = vmail } user = dovecot } service dict { unix_listener dict { group = vmail mode = 0600 user = vmail } } service imap-login { inet_listener imap { port = 143 } } service lmtp { executable = lmtp -L unix_listener /var/spool/postfix/private/dovecot-lmtp { group = postfix mode = 0600 user = postfix } } service managesieve-login { inet_listener sieve { port = 4190 } } service pop3-login { inet_listener pop3 { port = 110 } } ssl_cert = </etc/pki/dovecot/certs/dovecot.pem ssl_key = </etc/pki/dovecot/private/dovecot.pem userdb { args = /etc/dovecot/dovecot-sql.conf.ext driver = sql } protocol lmtp { info_log_path = /var/log/dovecot-lmtp.log log_path = /var/log/dovecot-lmtp-errors.log mail_plugins = " quota trash expire autocreate sieve quota" postmaster_address = postmaster(a)testdo.com } protocol sieve { info_log_path = /var/log/dovecot-sieve.log log_path = /var/log/dovecot-sieve-errors.log managesieve_implementation_string = dovecot managesieve_max_line_length = 65536 } protocol imap { mail_plugins = " quota trash expire imap_quota notify quota" } protocol pop3 { mail_plugins = " quota trash expire quota" } Any help is appreciated, thanks.

3 5

SSL overview...
by SH Development 16 Oct '17

16 Oct '17

Can someone help me understand the overall picture of SSL certificates in this scenario? I have a working dovecot/postfix/mysql server. It has a certificate. I now want to create a second, essentially duplicate configured server for use with replication. What is the relationship between the certificate and the hostname, or the DNS entry since the certs are created using the server’s domain name? mail.serverA.mydomain has a certificate that was created using mail.serverA.mydomain. The server's hostname is mail.serverA.mydomain. Now enter the new server. It would probably be: mailserverB.mydomain and a certificate created using mail.serverB.mydomain. The server’s hostname would be mail.serverB.mydomain. My questions: 1. Can I use the same certificate on both servers since they are serving email for the same domain? 2. Does the hostname have to be the same as the domain name, and thus the name used to create the cert? Sorry if I’m muddled about this. It was never really an issue until I wanted to add a secondary server into the mix. Ethon B.

2 3

Convert from mbox to Maildir
by Josef Wolf 16 Oct '17

16 Oct '17

Hello everybody, I am in the process to install dovecot as an IMAP server. Currently, this system runs postfix and uses procmail for local (mbox style) delivery. Procmail ist started by postfix like this: /etc/postfix/main.cf: mailbox_command = /usr/bin/procmail ~/.forward: "|IFS=' ' && exec /usr/bin/procmail -f- || exit 75 # jw" Before going into IMAP details, I decided to convert the system to Maildir as a first step. So I created a procmail-entryrecipe to populate a test folder: ~/.procmailrc: :0 * ^Subject:.*dovecot-test { ORGMAIL="${HOME}/Maildir/" DEFAULT=${ORGMAIL} DELIVER="/usr/lib/dovecot/deliver -o mail_location=maildir:~/Maildir -o lda_mailbox_autocreate=yes" :0 | $DELIVER -m dovecot-test } This setup seems to work. But there is one point that bugs me: in the first tests, dovecot was mis-configured (the two -o options were missing) and the test mails were silently lost. No errors in the procmail logfile. Only dovecot reports errors in its logfile. But the error seems not to be noticed by procmail. Therefore procmail assumes deliver could store the mail successfully and won't keep a backup of it. I am sure, I'm doing something stupid here. I have never seen such behavior from procmail before. Normally procmail would report failure of sub-commands. Now I hesitate to go forward in this conversion process. I don't want to loose mails if something is wrong with the dovcot setup. Besides: I wonder whether the setting DEFAULT=${ORGMAIL} is a good idea. Wouldn't it be better to have DEFAULT=/var/spool/mail/$USER so there would be a fallback in case there are delivery problems to ${ORGMAIL} Any thougths? -- Josef Wolf jw(a)raven.inka.de

4 5

Question about ITERINDEX in 2.2.32
by María Arrea 16 Oct '17

16 Oct '17

Hello I have a question about the new setting ITERINDEX. This is my filesystem layout /indices_dovecot : 8 x Fast SSD drives /buzones : 32x SAS 10k rpm disks /buzones/alternativo : 24 x SAS 7.2k rpm slow disks This is my mail_location: mail_location = mdbox:/buzones/domain/%2.26Hn/%2.200Hn/%n:INDEX=/indices_dovecot/indices/%2.26Hn/%2.200Hn/%n:ALT=/buzones/alternativo/domain/%2.26Hn/%2.200Hn/%n What should I write in mail_location with the new ITERINDEX setting? Regards María

2 6

Locking Errors
by List 16 Oct '17

16 Oct '17

We recently upgraded to version 2.2.10-8 on Centos 7 (from 2.2.10-7) and immediately started experiencing errors in the log file that look like this: Error: Broken file /mail/user/Maildir/dovecot-uidlist line 891: Invalid data: Error: Timeout (180s) while waiting for lock for transaction log file /mail/user/Maildir/dovecot.index.log Error: /mail/user/Maildir/.Trash/dovecot-uidlist: next_uid was lowered (421 -> 419, hdr=418) Error: Corrupted index cache file /mail/user/Maildir/dovecot.index.cache: invalid record size Error: Broken file /mail/user/Maildir/.Trash/dovecot-uidlist line 156: Invalid data: Error: Corrupted index cache file /mail/user/Maildir/dovecot.index.cache: invalid record size Error: Index /mail/user/Maildir/dovecot.index: Lost log for seq=2 offset=2392 Error: fcntl(write-lock) locking failed for file /mail/user/Maildir/dovecot.index.log: Stale file handle We are not sure what might be causing this, can anyone help illuminate what’s going on?

3 5

Setting up replication - First steps...
by SH Development 16 Oct '17

16 Oct '17

Just had my first (non-Dovecot) server failure due to dried out caps on the motherboard. Got me to thinking about my single-point-of-failure mail server. Currently running Dovecot and Postfix with no issues, but want to start taking steps just to be safe. I currently run a filesystem backup every 24 hours to a tar file over NFS to another server in our rack. I am backing up: /home/vmail /etc/dovecot /etc/postfix Unfortunately, the vmail directory has grown to 27GB and takes around 7 hours now to backup as described above. Which leads me to start thinking about how quickly I could restore the server from a backup if need be, and that time is at least 7 hours just to copy and untar the files onto another hard drive. I’m sure I could hook up a HD up directly to the backup server, then I could considerably reduce the time, but I’m making the assumption that I won’t always have quick physical access to the location. So I believe my first step is to set up another server, on another IP, different hostname, with Dovecot and Postfix, and simply use the files from the /etc directories of the existing server to configure it (changing the IP and hostname of course). Am I on the right track so far? Next steps involve setting up replication with dsync? If I have successfully setup replication between the two servers, does this mean users can then actually log into either server and have their “stuff” intact? So I could set up DNS failover in case the primary server fails? Would this make the setup acceptable for secondary MX as everything should sync to the primary server when it comes back online? Sorry for thinking out loud, but I want to make sure I’m understanding the bigger picture here. Jeff

4 3

Re: Newbie question...
by SH Development 16 Oct '17

16 Oct '17

I see you haven’t received an answer. Sorry people here don’t seem to have time or inclination to answer basic questions. I’ve tried. You’re better to search elsewhere. Ethon > On Oct 11, 2017, at 12:17 PM, Michael Segel <dovecot_mjs(a)hotmail.com <mailto:dovecot_mjs@hotmail.com>> wrote: > > Hi, > > I’ve downloaded and built dovecot from the source. > Now I’ve run in to a couple of issues. > Like namespaces. > > Is there an example of the namespaces settings? Nothing fancy, just the basics… > > These are private mailboxes, but still need to know what I have to do to set this correctly. > > Thanks In Advance, > > -Mike >

3 4

Filtering by country
by Anvar Kuchkartaev 16 Oct '17

16 Oct '17

Is it possible to filter out logins by country (I would like to limit dovecot instance users to log in only from specific countries)?‎‎ Anvar Kuchkartaev anvar(a)anvartay.com

3 2

Re: Core Dumps with dovecot-2.2 -git as of 2.2.devel (eaf276b33)
by Aki Tuomi 14 Oct '17

14 Oct '17

We'll take a look ---Aki TuomiDovecot oy -------- Original message --------From: Reuben Farrelly <reuben-dovecot(a)reub.net> Date: 12/10/2017 07:50 (GMT+02:00) To: Aki Tuomi <aki.tuomi(a)dovecot.fi> Subject: Re: Core Dumps with dovecot-2.2 -git as of 2.2.devel (eaf276b33) Hi Aki, I think I posted a bt full of this in the original email I sent to the list: https://www.dovecot.org/pipermail/dovecot/2017-October/109583.html I've got heaps of coredumps if we need to pick apart one specific one now though. Reuben On 12/10/2017 3:36 PM, Aki Tuomi wrote: > can you provide bt full output from core? > > > > --- > Aki Tuomi > Dovecot oy > > -------- Original message -------- > From: Reuben Farrelly <reuben-dovecot(a)reub.net> > Date: 12/10/2017 06:22 (GMT+02:00) > To: Aki Tuomi <aki.tuomi(a)dovecot.fi>, dovecot(a)dovecot.org > Subject: Re: Core Dumps with dovecot-2.2 -git as of 2.2.devel (eaf276b33) > > Hi, > > Just to confirm this is still a major problem for me with the 2.2.33.1 > release (did not exist in previous releases). > > Reuben > > > On 8/10/2017 6:30 AM, Aki Tuomi wrote: >> Thanks! We'll look into it. >> >> >> >> --- >> Aki Tuomi >> Dovecot oy >> >> -------- Original message -------- >> From: Reuben Farrelly <reuben-dovecot(a)reub.net> >> Date: 07/10/2017 13:08 (GMT+02:00) >> To: dovecot(a)dovecot.org >> Subject: Core Dumps with dovecot-2.2 -git as of 2.2.devel (eaf276b33) >> >> This appears to be breaking dsync at the moment and is caused by >> something committed in the last week or so. This is with the current >> -git of dovecot-2.2 (eaf276b33). >> >> The crash happens all the time - the logs are full of this message. >> >> Oct 7 20:31:24 thunderstorm.reub.net dovecot: doveadm: Warning: I/O >> leak: 0x7f8bb6581b4c (ostream-file.c:350, fd 14) >> Oct 7 20:31:24 thunderstorm.reub.net dovecot: doveadm: Panic: file >> ioloop.c: line 126 (io_remove_full): assertion failed: (io->callback != >> NULL) >> Oct 7 20:31:24 thunderstorm.reub.net dovecot: doveadm: Error: Raw >> backtrace: /usr/lib64/dovecot/libdovecot.so.0(+0xd48a6) [0x7f8bb654a8a6] >> -> /usr/lib64/dovecot/li >> bdovecot.so.0(+0xd5e26) [0x7f8bb654be26] -> >> /usr/lib64/dovecot/libdovecot.so.0(i_fatal+0) [0x7f8bb654ac39] -> >> /usr/lib64/dovecot/libdovecot.so.0(+0xf5cb0) [0x7f8bb6 >> 56bcb0] -> /usr/lib64/dovecot/libdovecot.so.0(io_remove+0x1d) >> [0x7f8bb656be31] -> /usr/lib64/dovecot/libdovecot.so.0(+0x10b5f8) >> [0x7f8bb65815f8] -> /usr/lib64/dovec >> ot/libdovecot.so.0(o_stream_cork+0x5a) [0x7f8bb657e289] -> >> /usr/lib64/dovecot/libssl_iostream_openssl.so(+0x6406) [0x7f8bb12d6406] >> -> /usr/lib64/dovecot/libssl_iost >> ream_openssl.so(openssl_iostream_bio_sync+0x18) [0x7f8bb12d6af6] -> >> /usr/lib64/dovecot/libssl_iostream_openssl.so(+0xa52e) >> [0x7f8bb12da52e] -> / >> usr/lib64/dovecot/libssl_iostream_openssl.so(+0xa6ef) [0x7f8bb12da6ef] >> -> /usr/lib64/dovecot/libdovecot.so.0(o_stream_sendv+0xc7) >> [0x7f8bb657e765] -> >> /usr/lib64/dovecot/libdovecot.so.0(o_stream_send+0x61) [0x7f8bb657e688] >> -> /usr/lib64/dovecot/libdovecot.so.0(+0x10dc2c) [0x7f8bb6583c2c] -> >> /usr/lib64/dovecot/libdovecot.so.0(+0x10ddd4) [0x7f8bb6583dd4] -> >> /usr/lib64/dovecot/libdovecot.so.0(o_stream_sendv+0xc7) [0x7f8bb657e765] >> -> /usr/lib64/dovecot/libdovecot.so.0(o_stream_nsendv+0x55) >> [0x7f8bb657e93f] -> >> /usr/lib64/dovecot/libdovecot.so.0(o_stream_nsend+0x61) [0x7f8bb657e8d3] >> -> dovecot/doveadm-server() [0x444a35] -> >> /usr/lib64/dovecot/libdovecot.so.0(i_warning+0xce) [0x7f8bb654afdc] -> >> /usr/lib64/dovecot/libdovecot.so.0(io_loop_destroy+0x2ca) >> [0x7f8bb656d8cd] -> dovecot/doveadm-server() [0x4454d1] -> >> dovecot/doveadm-server() [0x445793] -> dovecot/doveadm-server() >> [0x445dd6] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_call_io+0xd6) >> [0x7f8bb656d241] -> >> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x225) >> [0x7f8bb656fe58] -> >> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0x3e) >> [0x7f8bb656d422] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0xbe) >> [0x7f8bb656d345] >> Oct 7 20:31:24 thunderstorm.reub.net dovecot: doveadm: Fatal: master: >> service(doveadm): child 27186 killed with signal 6 (core dumped) >> >> >> thunderstorm reuben # file core >> core: ELF 64-bit LSB core file x86-64, version 1 (SYSV), SVR4-style, >> from 'dovecot/doveadm-server', real uid: 0, effective uid: 0, real gid: >> 0, effective gid: 0, execfn: '/usr/libexec/dovecot/doveadm-server', >> platform: 'x86_64' >> >> thunderstorm reuben # gdb /usr/libexec/dovecot/doveadm-server core >> GNU gdb (Gentoo 8.0.1 vanilla) 8.0.1 >> Copyright (C) 2017 Free Software Foundation, Inc. >> License GPLv3+: GNU GPL version 3 or later >> <http://gnu.org/licenses/gpl.html> >> This is free software: you are free to change and redistribute it. >> There is NO WARRANTY, to the extent permitted by law. Type "show copying" >> and "show warranty" for details. >> This GDB was configured as "x86_64-pc-linux-gnu". >> Type "show configuration" for configuration details. >> For bug reporting instructions, please see: >> <https://bugs.gentoo.org/>. >> Find the GDB manual and other documentation resources online at: >> <http://www.gnu.org/software/gdb/documentation/>. >> For help, type "help". >> Type "apropos word" to search for commands related to "word"... >> Reading symbols from /usr/libexec/dovecot/doveadm-server...done. >> [New LWP 27186] >> [Thread debugging using libthread_db enabled] >> Using host libthread_db library "/lib64/libthread_db.so.1". >> Core was generated by `dovecot/doveadm-server'. >> Program terminated with signal SIGABRT, Aborted. >> #0 0x00007f8bb60e7f70 in raise () from /lib64/libc.so.6 >> (gdb) bt full >> #0 0x00007f8bb60e7f70 in raise () from /lib64/libc.so.6 >> No symbol table info available. >> #1 0x00007f8bb60e9aa7 in abort () from /lib64/libc.so.6 >> No symbol table info available. >> #2 0x00007f8bb654a8f5 in default_fatal_finish (type=LOG_TYPE_PANIC, >> status=0) at failures.c:201 >> backtrace = 0x228e4d8 >> "/usr/lib64/dovecot/libdovecot.so.0(+0xd48a6) [0x7f8bb654a8a6] -> >> /usr/lib64/dovecot/libdovecot.so.0(+0xd5e26) [0x7f8bb654be26] -> >> /usr/lib64/dovecot/libdovecot.so.0(i_fatal+0) [0x7f8bb654ac39] -> /usr"... >> #3 0x00007f8bb654be26 in i_internal_fatal_handler (ctx=0x7ffef81daa50, >> format=0x7f8bb65ba898 "file %s: line %d (%s): assertion failed: (%s)", >> args=0x7ffef81daa70) >> at failures.c:670 >> status = 0 >> #4 0x00007f8bb654ac39 in i_panic (format=0x7f8bb65ba898 "file %s: line >> %d (%s): assertion failed: (%s)") at failures.c:275 >> ctx = {type = LOG_TYPE_PANIC, exit_status = 0, timestamp = 0x0, >> timestamp_usecs = 0} >> args = {{gp_offset = 40, fp_offset = 48, overflow_arg_area = >> 0x7ffef81dab50, reg_save_area = 0x7ffef81daa90}} >

3 6

Updating keywords on copy/move
by Daniele Ricci 14 Oct '17

14 Oct '17

Hello, I'm working on my virtual keywords plugin [1] to add an apparently simple feature: add a keyword to a message when it is moved or copied into a virtual folder (let's forget about what keyword to add, I'm starting with a hard-coded one to understand Dovecot plugin development). So I've added a mail_copy hook by the notify plugin [2] and did a simple call to mail_update_keywords [3]. The current source code shows a call to mail_get_backend_mail, this is me experimenting, I've also tried calling mail_update_keywords directly on "dst" mail. The result is an imap crash: Oct 14 21:22:26 notebook dovecot: imap(daniele): Panic: file mail-log-plugin.c: line 461 (mail_log_mail_transaction_commit): assertion failed: (msg->pretext == NULL) Oct 14 21:22:26 notebook dovecot: imap(daniele): Error: Raw backtrace: /usr/lib/dovecot/libdovecot.so.0(+0x9f0a2) [0x7f81af52f0a2] -> /usr/lib/dovecot/libdovecot.so.0(+0x9f19a) [0x7f81af52f19a] -> /usr/lib/dovecot/libdovecot.so.0(i_fatal+0) [0x7f81af4bfcf8] -> /usr/lib/dovecot/modules/lib20_mail_log_plugin.so(+0x19e8) [0x7f81aeae89e8] -> /usr/lib/dovecot/modules/lib15_notify_plugin.so(notify_contexts_mail_transaction_commit+0x58) [0x7f81aecec548] -> /usr/lib/dovecot/modules/lib15_notify_plugin.so(+0x2105) [0x7f81aeced105] -> /usr/lib/dovecot/libdovecot-storage.so.0(mailbox_transaction_commit_get_changes+0x52) [0x7f81af807ca2] -> /usr/lib/dovecot/libdovecot-storage.so.0(mailbox_transaction_commit+0x21) [0x7f81af807d71] -> /usr/lib/dovecot/modules/lib20_virtual_plugin.so(virtual_transaction_commit+0x78) [0x7f81ae8e2fb8] -> /usr/lib/dovecot/modules/lib15_notify_plugin.so(+0x20ec) [0x7f81aeced0ec] -> /usr/lib/dovecot/libdovecot-storage.so.0(mailbox_transaction_commit_get_changes+0x52) [0x7f81af807ca2] -> dovecot/imap(+0x10084) [0x563d0ff2b084] -> dovecot/imap(command_exec+0x5c) [0x563d0ff37d9c] -> dovecot/imap(+0x1b322) [0x563d0ff36322] -> dovecot/imap(+0x1b3bc) [0x563d0ff363bc] -> dovecot/imap(client_handle_input+0x18d) [0x563d0ff3677d] -> dovecot/imap(client_input+0xac) [0x563d0ff36ccc] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_call_io+0x52) [0x7f81af544742] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run_internal+0x12e) [0x7f81af545d8e] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_handler_run+0x36) [0x7f81af5447d6] -> /usr/lib/dovecot/libdovecot.so.0(io_loop_run+0x38) [0x7f81af544988] -> /usr/lib/dovecot/libdovecot.so.0(master_service_run+0x13) [0x7f81af4ca353] -> dovecot/imap(main+0x329) [0x563d0ff29139] -> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf1) [0x7f81af1132e1] -> dovecot/imap(_start+0x2a) [0x563d0ff292ca] This happens after my hook returns. Now, before activating core dumps and starting to do some serious debugging and reverse engineering, I ask: do I have to do something extra such as opening a transaction or something? Isn't the notify plugin supposed to do that already? Thanks for your help! [1] http://github.com/daniele-athome/dovecot-virtual-keywords-plugin [2] https://github.com/daniele-athome/dovecot-virtual-keywords-plugin/blob/keyw… [3] https://github.com/daniele-athome/dovecot-virtual-keywords-plugin/blob/keyw… -- Daniele

1 0