It seems that the use of tablespaces doesn't store everything in the tablespace directory and will complicate management during operation as the size and amount of the data files needs to be pre-defined. Having multiple instances seems to be a good solution and recommended when having standard mysql (https://dev.mysql.com/doc/refman/8.0/en/multiple-data-directories.html ), I couldn't find a good reference for this point for the ndb cluster. 

Added comments and added configurations in: https://confluence.lsstcorp.org/display/LTS/MySQL+cluster+setup

Currently reviewing with Andres Villalobos [X] for puppet deployment

To test ssh to 139.229.236.200 then do the following commands:

cd lsst_devops/nodes/

vagrant ssh ts-efd-data-01

sudo su -

cd /home/vagrant/

du -h --max-depth=1 /mnt

You will see the directory sizes, now you can fill each instance with data and see how they increase in size depending on the port:

python fillTables.py 3308 > NULL  # For TIER1

du -h --max-depth=1 /mnt**

python fillTables.py 3309 > NULL  # For TIER2**

du -h --max-depth=1 /mnt**

IP is 139.229.136.200

Tested with suggested commands. Tier storage reports increased as expected

Done

done