Re: Decreasing Checkpoints & allowing for heavy load

[Adam D]

Is there some sort of gauge to determine a fair value for these based on some average runtime information such as;
Hourly; selects; 50,000, deletes: 15,000, updates: 15,000, inserts: 15,000

It probably highly depends on row size. But would you have some 'fair' calculation similar to the one used in the docs about memory requirements per datanode for settings like these. As its quite hard to predict what 'load' settings required in the cluster without actually having the cluster in production. Would be great if something like that was around.

Adam