What is a common cause of a Spark job failing with out of memory errors?
- A A wide transformation or skewed key causing a single executor to hold too much data
- B Too many small files
- C Insufficient disk space
- D Incorrect file format
Answer
A wide transformation or skewed key causing a single executor to hold too much data
Salting skewed keys, increasing partitions, or replacing a collect with a distributed write are the usual remedies.





