In what manner does nodetool garbagecollect operate on several data directories?

edit

Nodetool garbagecollect operates on one data directory at a time. You can specify the directory using the -Dcassandra.storagedir flag when running the command. If you have multiple data directories, you will need to run the nodetool garbagecollect command on each of them individually. This is because each data directory may have its own set of tombstones and obsolete data that need to be cleaned up. Running the command on each data directory will help ensure that your cluster is running at optimal performance by removing unnecessary data.