Panic: file smtp-client-connection.c: line 1142 (smtp_client_connection_established)
I'm using dovecot in a multi server setup. Two directors forward pop3, imap and
smtp to backend dovecots with their central storage on NFS.
Yesterday I stumbled upon these log lines:
2019-11-18 20:17:51 lmtp(17274): Info: Connect from 172.22.1.6
2019-11-18 20:17:51 lmtp(17274): Panic: file smtp-client-connection.c: line
1142 (smtp_client_connection_established): assertion failed:
(!conn->connect_succeeded)
2019-11-18 20:17:51 lmtp(17274): Fatal: master: service(lmtp): child 17274
killed with signal 6 (core not dumped -
https://dovecot.org/bugreport.html#coredumps - set /proc/sys/fs/suid_dumpable
to 2)
Is a core dump required to find the source of that error?
Thanks for any help!
$ doveconf -n
# 2.3.4.1 (f79e8e7e4): /etc/dovecot/dovecot.conf
# OS: Linux 4.9.0-8-amd64 x86_64
# Hostname: m03.company.com
auth_cache_negative_ttl = 65 secs
auth_cache_ttl = 3 mins
auth_cache_verify_password_with_worker = yes
auth_mechanisms = plain login
auth_verbose = yes
debug_log_path = /dev/stdout
default_client_limit = 10400
default_process_limit = 1024
default_vsz_limit = 1 G
deliver_log_format = msgid=%m: %s %{session_time} %$
director_mail_servers = 10.10.10.2 10.10.10.3
director_servers = 10.10.1.2 10.10.1.3
imap_id_log = *
imap_max_line_length = 2 M
import_environment = TZ CORE_OUTOFMEM CORE_ERROR
info_log_path = /dev/stdout
instance_name = dovecot_director
listen = 10.10.1.3,::
lmtp_proxy = yes
log_path = /dev/stderr
log_timestamp = "%Y-%m-%d %H:%M:%S "
login_log_format_elements = user=<%u> method=%m rip=%r lip=%l mpid=%e %c %k
session=<%{session}>
mail_max_userip_connections = 250
passdb {
args = /etc/dovecot/sql/dovecot-dict-sql-passdb.conf
driver = sql
result_failure = return-fail
result_internalfail = return-fail
result_success = continue-ok
}
passdb {
args = proxy=y starttls=any-cert master=direc...@m03.company.com
pass=#hidden_use-P_to_show# nopassword=y
driver = static
skip = unauthenticated
}
pop3_logout_format = top=%t/%p, retr=%r/%b, del=%d/%m, size=%s, uidl=%u
protocols = imap lmtp pop3
service auth-worker {
user = $default_internal_user
}
service director {
fifo_listener login/proxy-notify {
mode = 0600
user = $default_login_user
}
inet_listener {
port = 9090
}
unix_listener director-userdb {
mode = 0600
user = vmail
}
unix_listener login/director {
mode = 0666
}
}
service imap-login {
executable = imap-login director
process_min_avail = 2
vsz_limit = 1 G
}
service lmtp {
inet_listener lmtp {
port = 24
}
process_min_avail = 2
unix_listener lmtp {
mode = 00
}
user = vmail
vsz_limit = 1 G
}
service log {
user = dovenull
}
service pop3-login {
executable = pop3-login director
process_min_avail = 2
vsz_limit = 1 G
}
service pop3 {
user = vmail
}
ssl_cert =
Error: Mailbox INBOX: Broken file dovecot-uidlist: Invalid data
Hi list, today a user reported messages are disappearing from his INBOX and I discovered a two digit number of mailboxes with broken dovecot-uidlist files. The log says: Oct 29 08:51:34 mail02 dovecot: lmtp(i...@example.com)<5270>: Error: Mailbox INBOX: Broken file /var/vmail/example.com/info/dovecot-uidlist line 88: Invalid data: And the respective files do look broken: 4596 :1572344253.M529133P14861.mail03.example.com,S=53053,W=54233 ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@4597 :1572344797.M93833P15905.mail03.example.com,S=126866,W=129256 Setup : Behind a director I am running two dovecot backend servers sharing the vmail directory via NFS. Do you have an idea where this might come from? How can I debug this further? Cheers, Dirk
LMTP Proxy failure using master login via destuser field
Hi list,
I set up dovecot as IMAP proxy using a master user to login into the backend
and it works beautifully. Next I set out to do LMTP proxying and I struggled
(initially).
I got the log message:
Mar 17 11:23:44 mail postfix/lmtp[2665]: 99A3F219C8: to=,
relay=10.15.1.2[10.15.1.2]:24, delay=5.1, delays=4.9/0.05/0.01/0.07, dsn=5.1.1,
status=bounced (host 10.15.1.2[10.15.1.2] said: 550 5.1.1
<"joh...@doe.com*director"@mail.server.com> User doesn't exist:
"joh...@doe.com*director"@mail.server.com (in reply to RCPT TO command))
My director's passdb setting looked like this:
passdb {
driver = sql
args = /etc/dovecot/sql/my-passdb.conf
result_success = continue-ok
result_failure = return-fail
result_internalfail = return-fail
}
# Add director master user credentials if above succeeded
passdb {
driver = static
args = proxy=y starttls=any-cert \
destuser=%u*direc...@example.com \
pass=correcthorsebatterystaple \
nopassword=y
skip = unauthenticated
}
In the end I found that setting the `master` field instead of `destuser` solves
the issue. It looks like LMTP does not know how to handle the master user
separator although set. I this known/wanted behavior? Maybe this is something
to mention on the wiki page, it took a while to figure this out...
Kind regards,
Dirk
Mit freundlichen Grüßen
Dirk Weise
DevOps Engineer
--
bitloft GmbH | Schulstraße 2 | D-42551 Velbert
www.bitloft.de | we...@bitloft.de
Telefon: +49 2051 40514-32
Geschäftsführer: Robin Clemens, Franz-Josef Wieneke
Amtsgericht Wuppertal HRB 17297 | USt-IDNr. DE121540053
