Thought to update this thread. Figured out my issue with forEachBatch and
structured streaming, I had an issue where I did a count() before write()
so my streaming query branched into 2. I am now using Trigger and
structured streaming to handle checkpointing instead of doing it myself.
Thanks all
Hi all,
We recently upgraded to our jobs to spark 2.4.4 from 2.3 and noticed that
some jobs are failing due to lack of resources - particularly lack of
executor memory causing some executors to fail. However, no code change was
made other than the upgrade. Does spark 2.4.4 require more executor