I am trying to train and optimize a random forest. At first the cluster handles the garbage collection fine, but after a couple of hours the cluster breaks down as Garbage Collection has gone up significantly.The train_df has a size of 6,365,018 reco...