For testing purpose, I am using NoSQLBench to load the data on Cassandra 5.x. Its a 3 nodes cluster. I’ve noticed that when rampup(load) phase completes, the size of Bytes unrepaired reaches to the amount of data loaded / written.

Same happened when I use the data folders from a backup created using Cassandra 4.x. The data there was around 167GB.

I want to know if it’s an unusual behaviour or the expected one. Is it normal to have bytes unrepaired every time after bulk writes or bulk loading data.