Handle node decommission in cassandra Auto scaling

Question

Is anyone using Auto Scaling Group for Cassandra? if so how to handle node down events where a cassandra node goes down and immediately ASG launch a new node before running "nodetool removenode UUID" for the dead node. if anyone has done this, please share some inputs

Answer 1

Never, EVER, use autoscaling with Cassandra. Cassandra is not built to support autoscaling and any cluster topology change requires shuffling the data and other operational procedures.

There are procedures on how to add or remove nodes from the cluster but autoscaling ain't one of them.