Understanding neo4j replication : causal clustering

Hi team,

We have a neo4j causal cluster already running for a year which is ec2 instance-store backed. Which means, we lose the data once we restart the server.

In a cluster of 3 core servers, we restarted one server which had 44G before reboot. Now after it got added back to the cluster, it has only 7G of data.

We assumed that the servers would auto replicate the data and the server would go back to having 44G like pre-reboot.

Can you please help us understand what happens when a core server is added or removed, and if data gets replicated or not?

TIA.