I have a setup of Ambari installation with 1 active namenode, 1 secondary namenode and 5 datanodes to use HBASE primarily. Since this is a brand new cluster setup, we are not sure to with the default setting or not.
We do have one use case, but for the data system setup what are the factors we should look upon to finalize the various parameters?
e.g.: Load testing -- How shall I proceed with it? What should be our compaction parameters (shortCompactions as well)? What should be the initial config for memstore flushes (periodic flusher as well)?
Ambari starts with good default values. But , some parameters could be based on your use cases, where second guessing the best configuration can be tough. If you have smartsense, that can be a good second set of checks to give you the correct parameter values.
If you are a hortonworks customer, I definitely recommend exporting your smartsense generated bundles.