In CouchDB3 clusters provide horizontal scaling as well as durability, In attempting to take advantage of the durability, in a 3 node cluster if 2 of the 3 go down (when sharding is setup so that all 3 nodes have all the data) the remaining node will wait until a consistent 60s timeout has elapsed before it will return the data it has. Even if you anticipate this having been monitoring with _up, and attempt to use ?r=1 (or 0?) it still refuses to return any data until the timeout elapses.

in the past i’ve once deleted the _users database i don’t know why. I recreated the database using curl But now it seems like i can no longer add users using curl or node into _users database. e.g:

I have a CouchDB database with the following documents.