Hi folks,
Wondering if this is expected behavior; we'd like to be able to pause replication for a particular account on our 2.2.32 active-active cluster, but delivering a message (LMTP) to the account seems to turn replication back on again:
(replication on) server1:~# doveadm replicator status rgiles username priority fast sync full sync success sync failed rgiles none 00:00:30 23:45:01 00:00:30 -
(remove manually) server:~# doveadm replicator remove rgiles
(confirm replication is off) server:~# doveadm replicator remove rgiles Error: Replicator failed: User not found
(confirm again that replication is still off) server1:~# doveadm replicator status rgiles username priority fast sync full sync success sync failed
(and one more time, still off) server1:~# doveadm replicator status rgiles username priority fast sync full sync success sync failed
(send an e-mail, and voila, replication is running again) server1:~# doveadm replicator status rgiles username priority fast sync full sync success sync failed rgiles none 00:00:01 00:00:03 00:00:01 -
I've tried quickly disabling replication on both systems, and they inevitably override the 'doveadm replicator remove' command and begin dsync'ing again with each other.
Thanks for any ideas out there -
Robert