Github user davies commented on the pull request:
https://github.com/apache/spark/pull/3003#issuecomment-61186022
@kayousterhout Keep the details in the message of exception will help at
first, but becoming noisy later. In order to put the bytes in the message, it
needs to grab some details from TaskSetManager in TaskResultGetter, so I would
like to avoid it to keep the code more separated. Also there are many cases
that the final message in the exception did not carry the details of failure,
once these confusing behavior happens, user should look for logs( at least for
ERROR log).
In shared SparkContext (in our product), it's possible that multiple
collect may happen in the same time. But use maxResultSize for all concurrent
collect() will let it hard to understand, some small collect() will be aborted
by others. So I think it's better to do it in current approach.
Thanks for your comments!
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
