I was wondering how the closing of database work during replication. As far as we understand continuous replication is something we can start at anytime. If after this there is some new input in CBS bucket (remote) or CBL (local) it get picked up automatically and synced.
But when we start continous replication and want to acces a CBL database we get the error that the database is in use. We made sure the database isn’t left open elsewhere… .
Very strange, we don’t have the error anymore. We replicate the buckets in sequence. Every second time it hangs randomly at the replication of one bucket. Stat stays busy.
There is no way to force close all connections because they do not have a centralized way of knowing about each other. Just remember to close all database handlers and stop all replicators, as those are what keep those sorts of connections open. If that proves difficult then let us know about a scenario in which it is confusing so that we can potentially address it.