We are implementing a HANA system with SLT replication from ECC. Our ECC is very large. Our HANA system is 4 TB and we have 8 nodes.
Firstly, we have read the Admin guidelines on possible partitioning approaches (Hash, round-robin, time based, multi-level etc) and various blogs, so the theory is well understood, however there seems to be a lack of any practical guidelines, eg if a table is expected to be >100milion records then partition.
Secondly, Hash partitioning is easy since you do not really need to think about it, Range partitioning has benefits but also requires more thought and maintenance. Should an initial guideline be just to always Hash partion and only look at Range if there are specific issues/problems?