https://bugs.openldap.org/show_bug.cgi?id=10274
--- Comment #2 from falgon.comp@gmail.com --- We implemented this architecture according to our needs and performance optimization. We tested the mirror mode first but it was not suitable for our case. For the issue we are reporting, we would like to know why we have no log announcing a problem and why our instances freeze at the replication level. Our observation is not a behavior expected by OpenLDAP, how can we explain this case?