This article explains where DB access to all the analytic nodes are impacted when a single node's cassandra is in DN state.

This issue is induced because "cassandra" user needs a consistently quorum check where in a 4 node cluster. 3 nodes have to respond to the consistency check not including the self node which leads to a situation of an access/login failure in case one of the nodes is in DN state. The solution is to create, and use, a different user such as  "versa".

The quorum check issue is restricted to “cassandra” user. The “versa” user, that we are creating, is not subject to this restriction and does not need a quorum consistency validation.

Please find the procedure to be followed is as below

Step 1: Go to the cqlsh prompt of any one of analytic personality nodes

Step 2: Execute the below cmd to create "versa" user and set it as superuser

exit from cqlsh

Step 3:  Update the following fields in /opt/versa/var/van-app/properties/application.properties on all the nodes in the cluster (Analytics + Search + Log forwarders):

db.analytics.username=versa

db.analytics.password=versa123

Step 4:  Restart services on all the nodes (vsh restart). you can do this one by one and check "vsh status" to confirm if all services are "running"