replication Error with wrong file lock

18 Oct 2024


      Hi there,
I have setup a replication between two dovecot servers. It seems working:
# doveadm replicator status '*' && doveadm replicator dsync-status &&
doveadm replicator status
username priority fast sync full sync success sync failed
******************** none     00:59:22  18:50:17  18:21:25     -
********************low      01:54:26  18:57:58         -            y
********************none     21:42:55  21:47:55  21:36:24     -
********************           none     21:14:57  21:51:07  21:33:13     -
********************none     21:13:52  21:51:04  21:17:57     -
********************none     00:04:23  01:29:24        -            y
********************none     02:34:15  21:56:23  02:34:15     -
********************       none     00:03:56  18:55:17        -            y
********************none     01:18:42  01:18:42  19:18:46     -
********************none     21:20:04  21:24:46  21:43:49     -
********************none     00:05:13  00:09:36  21:47:55     -
********************none     00:16:50  18:55:17  18:45:57     -
********************none     00:04:23  18:55:17  38:07:24     y
username type   status
-      Not connected
-      Not connected
-      Not connected
********************normal Waiting for dsync to finish
-      Not connected
Queued 'sync' requests        0
Queued 'high' requests        0
Queued 'low' requests         1
Queued 'failed' requests      0
Queued 'full resync' requests 0
Waiting 'failed' requests     3
Total number of known users   13
But on the one server, I found this errors repeating and this users are
at the list above marked as failed:
Oct 18 09:41:08 dsync-local(1@tld.de): Error: Couldn't lock
/srv/mail/mail_storage/tld.de/1/.dovecot-sync.lock:
fcntl(/srv/mail/mail_storage/tld.de/1/.dovecot-sync.lock, write-lock,
F_SETLKW) locking failed: Timed out after 30 seconds (WRITE lock held by
pid 71613)
The directory shows:
# ls -la /srv/mail/mail_storage/celebrate.de/1/ | grep '\.dove'
drwxrwxrwx  5 vpostfix vpostfix     4096 Oct 18 08:55 .dovecot
-rw-------  2 vpostfix vpostfix        0 Oct 18 09:37 .dovecot-sync.lock
-rwxrwxrwx  1 vpostfix vpostfix      912 Mar 20  2023 .dovecot.sieve.log
-rwxrwxrwx  1 vpostfix vpostfix      275 Oct 11 00:00 .dovecot.svbin
So it seems, that there are more processes want accessing same locking
file. But normally this and other files should not exist there:
# ls -la /srv/mail/mail_storage/celebrate.de/fk/ | grep '\.index'
-rw-rw-rw-  1 vpostfix vpostfix   194008 Oct 17 12:26 dovecot.index
-rwxrwxrwx  1 vpostfix vpostfix 16890796 Oct 17 14:20 dovecot.index.cache
-rw-rw-rw-  1 vpostfix vpostfix     7420 Oct 17 14:04 dovecot.index.log
-rwxrwxrwx  1 vpostfix vpostfix    32824 Oct 17 11:46 dovecot.index.log.2
-rwxrwxrwx  1 vpostfix vpostfix     5752 Oct  9 05:14 dovecot.list.index
-rwxrwxrwx  1 vpostfix vpostfix     7556 Oct  9 09:32 dovecot.list.index.log
In my config, I have set:
mail_location =
maildir:~:VOLATILEDIR=/dev/shm/dovecot/%2.256Nu/%u:INDEX=/opt/dovecot/%2.256Nu/%u:ITERINDEX
But below /dev/shm/dovecot and /opt/dovecot, directories and files are
generated too
Following my config and hoping for some useful hints:
# dovecot -n
# 2.2.36 (1f10bfa63): /etc/dovecot/dovecot.conf
# Pigeonhole version 0.4.24 (124e06aa)
# OS: Linux 5.15.149-1-pve x86_64 CentOS Linux release 7.9.2009 (Core)
# Hostname: host.tld.de
disable_plaintext_auth = no
doveadm_password =  # hidden, use -P to show it
first_valid_uid = 1000
imap_id_send =
lda_mailbox_autocreate = yes
lda_mailbox_autosubscribe = yes
lmtp_save_to_detail_mailbox = yes
log_path = /var/log/dovecot.log
mail_fsync = always
mail_full_filesystem_access = yes
mail_gid = 1000
mail_home = /srv/mail/mail_storage/%d/%n
mail_location =
maildir:~:VOLATILEDIR=/dev/shm/dovecot/%2.256Nu/%u:INDEX=/opt/dovecot/%2.256Nu/%u:ITERINDEX
mail_plugins = " notify replication"
mail_privileged_group = vpostfix
mail_uid = 1000
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 index ihave duplicate mime foreverypart
extracttext vacation-seconds
mbox_write_locks = fcntl
mmap_disable = yes
namespace inbox {
inbox = yes
location =
mailbox Drafts {
special_use = \Drafts
}
mailbox Sent {
special_use = \Sent
}
mailbox "Sent Messages" {
special_use = \Sent
}
mailbox Spam {
auto = subscribe
special_use = \Junk
}
mailbox Trash {
special_use = \Trash
}
prefix =
}
passdb {
args = scheme=CRYPT username_format=%u /etc/dovecot/users
driver = passwd-file
}
plugin {
mail_replica = tcp:192.168.130.192:4711
recipient_delimiter = +
replication_full_sync_interval = 1 hour
replication_sync_timeout = 10
sieve = file:~/sieve;active=~/.dovecot.sieve
sieve_before = /etc/dovecot/spam-global.sieve
sieve_extensions = +vacation-seconds
sieve_vacation_default_period = 1d
sieve_vacation_max_period = 30d
sieve_vacation_min_period = 0
sieve_vacation_use_original_recipient = yes
}
protocols = imap pop3 lmtp sieve
service aggregator {
fifo_listener replication-notify-fifo {
mode = 0666
user = vpostfix
}
unix_listener replication-notify {
mode = 0666
user = vpostfix
}
}
service auth {
inet_listener {
address = 192.168.130.191
port = 3659
}
unix_listener /var/spool/postfix/private/auth {
group = vpostfix
mode = 0666
user = vpostfix
}
unix_listener auth-userdb {
group = vpostfix
mode = 0600
user = vpostfix
}
}
service config {
unix_listener config {
user = vpostfix
}
}
service doveadm {
inet_listener {
port = 4711
}
user = vpostfix
}
service imap-login {
process_min_avail = 1
service_count = 1
}
service lmtp {
inet_listener lmtp {
address = 192.168.130.191
port = 24
}
}
service managesieve-login {
inet_listener sieve {
port = 4190
}
}
service replicator {
process_min_avail = 1
unix_listener replicator-doveadm {
mode = 0666
user = vpostfix
}
}
ssl_cert =

Frank Kirschner

Frank Kirschner

Aki Tuomi

tags

participants (2)