Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/1387#issuecomment-49076858
@witgo you do know that the Oracle VM has 3 different GCs, configurable
using command line arguments, right? Have you tried all of them (and
combinations thereof, since you can run different gcs in the young and old
gens) in different memory pressure scenarios?
These kinds of questions are why you shouldn't be trying to do that. It's
sort of ok to have that code in a test, especially given what the test is
trying to do. The test environment is more controlled that the environment
where Spark apps themselves are deployed. So you can't just take that code, put
it into the main code base, and say the issue is fixed...
Also, did you take a look at
https://github.com/apache/spark/pull/1387#issuecomment-49072342? That might be
an easier way to achieve a "best effort" cleanup without relying on the gc API.
---
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.
---
