Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/2471#issuecomment-58934284
inotify is certainly interesting for the HistoryServer, but kinda
orthogonal to this. The other bug is more related to this one, and they're
discussing exactly what I mentioned: that deleting things is not the
responsibility of the fs. The applications that manage the files are the ones
who have the most context about when it's safe to delete things and how to do
so.
This case is a simplified one in that "when" doesn't matter much, as long
as the list of things that exist is kept up-to-date in the application. But the
point remains that the fs seems like the wrong place for that; and having a
sweeper-like feature in HDFS without inotify would be kinda bad for
applications like the HistoryServer.
But I digress. Even if HDFS-6382 is implemented, it would only come in a
new version of HDFS, and Spark needs to work with existing versions...
---
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]
