Github user pwendell commented on the pull request:
https://github.com/apache/spark/pull/158#issuecomment-38945893
@haoyuan but with Tachyon you create a new folder for each application.
Then within the application you have O(RDD blocks) files. This is way less than
O(shuffle blocks) files which happens withe the DiskStore - shuffle blocks tend
to be (RDD blocks)^2. What is the limitation of the number of files inside of a
Tachyon directory? How large can it be before performance starts to suffer?
---
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.
---