Hi,
I have been seeing the following messages in logs on all my Dovecot servers each running both a dovecot-director instance and a dovecot pop3/imap backend. These logs appear to relate to problems with the director.
Aug 28 16:09:52 dovecot01 dovecot: director: Error: director(10.0.0.2:9090/right): Output buffer full, disconnecting Aug 28 16:10:01 dovecot01 dovecot: director: Error: Ring SYNC appears to have got lost, resending Aug 28 16:10:02 dovecot01 dovecot: director: Error: Director 10.0.0.2:9090/right disconnected
Aug 28 16:09:51 dovecot02 dovecot: director: Error: director(10.0.0.3:9090/right): Output buffer full, disconnecting Aug 28 16:09:57 dovecot02 dovecot: director: Error: Ring SYNC appears to have got lost, resending Aug 28 16:10:01 dovecot02 dovecot: director: Error: Director 10.0.0.3:9090/right disconnected Aug 28 16:10:02 dovecot02 dovecot: director: Error: director(10.0.0.1:9090/left): Ping timed out, disconnecting Aug 28 16:13:15 dovecot02 dovecot: director: Warning: Delaying new user requests until ring is synced
Note dovecot01 has IP 10.0.0.1 and dovecot02 has IP 10.0.0.2 and so on.
This seems to show up when I enable the poolmon script (https://github.com/brandond/poolmon) on all the servers, and all servers start simultaneous logins to each others Dovecot backends. I did notice that the poolmon script has been flapping vhost weights a lot (I am looking at making the script handle transient errors better), which means that it is connecting to the dovecot-director socket every few minutes and changing vhost weights from 100 to 0 and vice-versa. Is there some sort of limit in the Dovecot Director that is being exceeded that should be tweaked?
# 2.1.5: /etc/dovecot-director/dovecot.conf # OS: Linux 2.6.32-279.el6.mccqfix.x86_64 x86_64 Red Hat Enterprise Linux Server release 6.3 (Santiago) auth_cache_size = 100 M auth_default_realm = example.com auth_failure_delay = 5 secs auth_mechanisms = plain login auth_verbose = yes auth_verbose_passwords = sha1 auth_worker_max_count = 600 base_dir = /var/run/dovecot-director/ director_mail_servers = 10.0.0.1 10.0.0.2 10.0.0.3 10.0.0.4 10.0.0.5 director_servers = 10.0.0.1 10.0.0.2 10.0.0.3 10.0.0.4 10.0.0.5 disable_plaintext_auth = no instance_name = dovecot-dir passdb { args = /etc/dovecot-director/dovecot-sql.conf driver = sql } service director { fifo_listener login/proxy-notify { mode = 0666 } inet_listener { address = 10.0.0.1 port = 9090 } unix_listener director-userdb { mode = 0600 } unix_listener login/director { mode = 0666 } } service imap-login { executable = imap-login director inet_listener imap { address = 10.0.0.1 port = 143 } inet_listener imaps { address = 10.0.0.1 port = 993 ssl = yes } process_min_avail = 100 } service pop3-login { executable = pop3-login director inet_listener pop3 { address = 10.0.0.1 port = 110 } inet_listener pop3s { address = 10.0.0.1 port = 995 ssl = yes } process_min_avail = 100 } verbose_proctitle = yes