Github user kayousterhout commented on the pull request:
https://github.com/apache/spark/pull/5403#issuecomment-111729666
@pwendell I've seen performance gains even when the data fits in the buffer
cache on the write side, because opening a file still requires writing metadata
to disk, which can become a point of contention when many tasks are opening
many files at the same time. I've also heard from a few folks who have
configurations with a lot of memory (and not much disk space / no local disks)
who would find this useful. That being said, there are other compelling
reasons to close this -- which I've summarized in the PR description.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
