Every commonly seen error has been discussed multiple times.
Meaning, you can find related discussions / JIRAs using indexing services,
such as:
http://search-hadoop.com/
Here is one related talk:
http://www.slideshare.net/Hadoop_Summit/why-your-spark-job-is-failing
FYI
On Sat, Oct 3, 2015 at 6:
There are executor logs and driver logs. Most of them are not intuitive enough
to mean anything to us.
Are there any notes, documents, talks on how to decipher these logs and
troubleshoot our applications' performance as a result?
thanks
Jeff