Hello, Today I started noticing a very odd Dovecot error in the mail logs across 2 replicated Dovecot hosts via dsync. I have searched the Dovecot archives, Google, back tracked, downgraded Dovecot to earlier versions from earlier this year and ruled everything out I could think of to no avail. There really are not any helpful reports with what I am specifically seeing. The strange thing is only vmail@domain.net is repeatedly shown in the log. The error looks like this. *Aug 24 22:46:02 orbitron dovecot: config: Error: BUG: Config client connection sent too much data****Aug 24 22:46:02 orbitron dovecot: doveadm(vmail@domain.net@domain.net@domain.net@domain.net@domain.net@domain.net@domain.net@domain.net@domain.net): Fatal: Error reading configuration: read(/var/run/dovecot/config) failed: read(size=8192) failed: Connection reset by peer*** $ dovecot --version 2.2.31 (65cde28) Dovecot #1: Linux kernel 4.9.36-x86_64 Dovecot #2: Linux kernel 4.12.8-x86_64 Here are the latest Dovecot updates performed this year. [2017-04-29 14:45] [ALPM] upgraded dovecot (2.2.28-2 -> 2.2.28-3) [2017-05-02 23:05] [ALPM] upgraded dovecot (2.2.28-3 -> 2.2.29.1-1) [2017-06-11 17:20] [ALPM] upgraded dovecot (2.2.29.1-1 -> 2.2.30.2-1) [2017-06-30 22:46] [ALPM] upgraded dovecot (2.2.30.2-1 -> 2.2.31-1) When Dovecot was last updated back on June 30, I did `systemctl daemon-reload` but never actually restarted the Dovecot daemon until what seems like today. The only thing I did today was edit /etc/dovecot/conf.d/10-auth.conf by simply enabling "*auth_username_format = %Lu*" which was commented out before I enabled it. That is the /*only*/ thing that I tested. After I was done testing I went back and commented out the same thing so it reads "*#auth_username_format = %Lu*" and of course restarted Dovecot. I really doubt this has anything to do with what I am reporting as auth is not the problem in this case. Thank You, Tony