Hi Abdul, There is no mechanism offered in Cassandra to bind a keyspace (when created) to specific filesystem or directory. If multiple filesystems or directories are specified in the data_file_directories property in the *cassandra.yaml* then Cassandra will attempt to evenly distribute data from all keyspaces across them.
Cassandra places table directories for each keyspace in a folder under the path(s) specified in the data_file_directories property. That is, if the data_file_directories property was set to */var/lib/cassandra/data* and keyspace "foo" was created, Cassandra would create the directory */var/lib/cassandra/data/foo*. One possible way bind a keyspace to a particular file system is create a custom mount point that has the same path as the keyspace. For example if you had a particular volume that you wanted to use for keyspace "foo", you could do something like: sudo mount /<my_device_volume_path> /var/lib/cassandra/data/foo Note that you would probably need to do this after the keyspace is created and before the tables are created. This setup would mean that all reads/writes for tables in keyspace "foo" would touch that volume. Regards, Anthony On Tue, 3 Jul 2018 at 07:02, Abdul Patel <abd786...@gmail.com> wrote: > Hi > > Can we bind or specify while creating keyspace to bind to specific > filesystem or directory for writing? > I see we can split data on multiple filesystems but can we decide while > fileystem a particular keyspace can read and write? >
