Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/6586#issuecomment-108893685
So, the problem here isn't heap memory right? because although this
allocates and caches buffers quite freely, they are in soft refs, which can be
removed if memory gets low. You are worried about off-heap memory right?
Because nothing understands when "off heap memory is low".
I think your change slows things down a bit and adds a little extra
complexity, but I don't know that we've shown there's a real problem here. Yes
Spark jobs already use a lot of off heap memory but this is not a problem per
se.
---
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]
