Debian 8.6 - Multiserver/Cluster - Dovecot dsync not syncing

Discussion in 'Server Operation' started by ZKool, Nov 9, 2016.

  1. ZKool

    ZKool New Member

    Hello,

    I have setup 2 servers running Debian 8.6 using this guide;
    https://www.howtoforge.com/tutorial...abase-cluster-on-debian-8.4-with-ispconfig-3/

    Everything seems to be working except for dovecot dsync.

    "doveadm replicator status '*'" shows the account trying to sync, but the fast sync + full sync time just always goes up.
    if i manually sync using "doveadm replicator replicate -f [email protected]" it works, and the fast sync time resets back to 0.


    dovecot.conf replicator IPs are correct.
    Firewall is currently off on both servers.

    I can't see/find any error messages in logs anywhere.

    Does anyone have any insight or ideas on what to check here ?

    Thanks.
     
    Last edited: Nov 9, 2016
  2. florian030

    florian030 Well-Known Member HowtoForge Supporter

    Do you have different hostnames (at least for) dovecot? dovecot --hostdomain

    You can change some protocols in dovecot.conf:
    Code:
    protocol imap {
      mail_plugins = quota imap_quota notify replication
    }
    protocol pop3 {
      pop3_uidl_format = %08Xu%08Xv
      mail_plugins = quota
    }
    protocol lda {
      mail_plugins = sieve quota notify replication
    [email protected]
    }
    
    on both servers and resart dovecot.
     
  3. ZKool

    ZKool New Member

    I did have different hostnames.

    Looks like its fixed now - I assume this did it;
    Code:
    protocol imap {
    mail_plugins = quota imap_quota notify replication
    }
    
    Thanks heaps!
     
  4. ZKool

    ZKool New Member

    Minor issues remaining;

    When I remove a mailbox from ispconfig, it still shows as trying to replicate with
    Code:
    doveadm replicator status '*'
    I was expecting it to automatically remove when mailbox deleted.

    Is this expected behaviour, or should it auto-remove ?

    Also, full sync's do not seem to be running even with
    Code:
    replication_full_sync_interval = 1 hours
    Any thoughts?
     

Share This Page