Expected Behavior

On a running cluster, if one turns half of a particular shard's replicas into maintenance_mode, the cluster should keep working properly.

Current Behavior

On a running cluster, if one turns half of a particular shard's replicas into maintenance_mode, Error getting security objects for <<"affected_database_here">> : {error,no_majority} begin to appear.

Possible Solution

Allow the security objects sync internal process to push through maintenance mode.

Context

I'm building a CouchDB automatic administration toolset and I was trying out different scenarios to see it working, in this case I was simulating an operation where on a database with two replicas per shard and one of them being down. I add a third node, put it into maintenance mode and make it replicate the particular shard. At that moment, majority cannot be achieved and the errors start arising.

Your Environment

It is a Kubernetes managed cluster where the Docker image in use is https://hub.docker.com/r/klaemo/couchdb/

• Version used: 2.0.0
• Link to your project: https://github.com/cabify/couchdb-admin