Critical Description
It is very easy to crash the zeppelin web-interface by having a large text output in Spark. Then after restarting zeppelin, the notebook will often be too big to load, causing the user to have to manually edit the note.json.
Example:
%pyspark
for var in range(1000000):
print(var)
Potentially we could implement some sort of safeguard against this, like an optional config for byte/character limit on the output, after which the output is truncated.
Attachments
Issue Links
- links to
