You are tasked with building a managed Hadoop system to serve as your data lake, where the data transformation process consists of a series of Hadoop jobs executed in sequence. To achieve the architecture of separating storage from compute, you have opted to utilize the Cloud Storage connector for handling all input data, output data, and intermediary data. However, it has come to your attention that one specific Hadoop job executes very slowly on Cloud Dataproc when compared to an on-premises bare-metal Hadoop environment, which is equipped with 8-core nodes and 100-GB RAM. An analysis indicates that this particular Hadoop job is highly disk I/O intensive. Your goal is to resolve this performance issue. What steps should you take?