Have gone through datastax docs where the recommended capacity is 2TB disk size per node for DSE 6.7, however, are the same recommendations applicable for apache cassandra too?
if the data storage requirement is 100TB; can it be split between 5 nodes having 20TB each?
or does it have to be 50 nodes having 2TB each?
@shiji.cherian_137128 There are optimisations in DSE which are not available in OSS Apache Cassandra so your experience may vary.
I would suggest sticking with the 500GB to 1TB density per node if your data is mostly hot (meaning it is updated regularly and accessed constantly). Only go beyond to 2-4 TB per node IF your data is cold. Cold data means:
The same caveats apply for dense nodes -- repairs are problematic, recovery takes a long time, bootstrap/decommission tasks take days. Cheers!
5 People are following this question.
