Github user aarondav commented on the pull request:
https://github.com/apache/spark/pull/607#issuecomment-41882098
A second downside, following Patrick's idea, is that all FileSystems
created during the execution of the same task will share the same UGI and thus
may reuse the cached version. This means we have to be very careful about
potentially sharing FileSystems across tasks (say in the future if we use
FileSystem for writing DiskStore blocks), and it's pretty hard to reason about
this given both the cache and the weird namespacing of the UGI doAs.
---
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.
---
