Hi, The replicator contacts the source and target on whichever port you specify, using the http protocol. port 4369 is erlang’s port daemon mapper, unrelated to replication.
As already noted, the _replicator database is the mechanism for defining a replication that will survive server reboots, whether you decide that the replication just gets the docs from source to target and then ends, or whether it runs indefinitely until you delete the replication. B. > On 12 Jun 2019, at 13:31, Sinan Gabel <sinan.ga...@gmail.com> wrote: > > You should use the _replicator database (and not one-off replications, > even if continuous), these also restart after reboot. > > As far as I know - someone else can correct me if I am wrong - the > _replicator database speaks through Erlang port, perhaps port 4369. > Thus you should check if the servers can speak to each other on the > necessary ports. > > On Wed, 12 Jun 2019 at 11:23, Rene Veerman <seductivea...@gmail.com> wrote: > >> i'm trying to set up replication of my databases, but i'm stuck in 3 ways : >> >> 1) i can't seem to be able to set up automated backup of all databases on a >> couchdb server to another server. >> >> 2) when i try to replicate a single database from couchdb servers which >> have no problem accessing databases via javascript or PHP, from the /_utils >> interface, the interface throws an error at me : "Failed to create the >> _replicator database." >> >> 3) i don't know, and the docs don't exactly spell this out either, >> whether or not a continuous replication job will survive a reboot of either >> or both of the machines involved in the replication. >> >> if someone more experienced could shed some light on this, i'd be very >> grateful. >>
