Databricks Certified Data Engineer Associate free question 1 / 10 · Domain 1: Databricks Intelligence Platform
A team creates all-purpose compute and leaves the access mode field at its default, Auto. Which of the following four setups makes Auto choose the Standard access mode?
- Runtime 16.4 LTS ML (machine learning), an instance type with no GPU, notebooks written in Scala
- Runtime 13.3 LTS (standard), an instance type with no GPU, notebooks written in R
- Runtime 16.4 LTS (standard), a GPU instance type, notebooks written in Python
- Runtime 15.4 LTS (standard), a memory-optimized instance type with no GPU, notebooks written in R
Show the answer and explanation
Answer: D
Why this is correct
Auto picks the access mode from the Databricks Runtime and the instance type you select. It defaults to Standard, and uses Dedicated only when a machine learning runtime, a GPU instance type, or a Databricks Runtime lower than 14.3 is selected. A standard 15.4 LTS runtime on a memory-optimized instance without a GPU matches none of those three conditions, so Auto picks Standard. The language the notebooks are written in is not one of the inputs to the decision.
Why the other options are wrong
"Runtime 16.4 LTS ML (machine learning), an instance type with no GPU, notebooks written in Scala" … Selecting a machine learning runtime (here 16.4 LTS ML) makes Auto choose Dedicated. Whether notebooks are written in Scala plays no part in what Auto decides.
"Runtime 16.4 LTS (standard), a GPU instance type, notebooks written in Python" … Selecting a GPU instance type makes Auto choose Dedicated, even when the runtime is the standard 16.4 LTS. The notebook language plays no part in the decision either.
"Runtime 13.3 LTS (standard), an instance type with no GPU, notebooks written in R" … When the Databricks Runtime is lower than 14.3, Auto chooses Dedicated, and 13.3 LTS is lower than 14.3. The fact that the notebooks use R does not enter the decision.
Common trap
Standard compute does not support R. Because Auto ignores the language, a team that plans to run R notebooks still gets Standard and cannot run R on it as it stands. If you know you need R, do not leave it to Auto; set the access mode explicitly when you create the compute.