Results of DecisionTreeClassifier differ between Spark 2 and Spark 3

Question

I trained a decision tree model on the same data using the same parameters with Spark 2 and Spark 3.

pyspark.ml.classification import DecisionTreeClassifier
tree_md = DecisionTreeClassifier(maxBins=500,maxDepth=6, seed=0)  

I obtained a different result of decision tree model. Could you tell me if this is normal?

Thanks,