Github user bbejeck commented on the pull request:
https://github.com/apache/spark/pull/2309#issuecomment-54865508
> Still think the fix should be made in the Utils class. This is not the
only place where this could cause trouble.
I agree there are other areas in the code that would benefit from the
memory check.
I originally handled the memory issue in the WorkerArguments class because
it has a precedence with respect to memory settings - default is using a
percentage of system memory, then use the SPARK_WORKER_MEMORY setting (if it
exists) and finally use the memory setting provided from the command line with
a "-m" switch.
My concern is If it's handled in the Utils class we could have a case where
the user is setting memory from the command line but is getting unexpected
errors from the mislabeled environment variable. At any rate I'm probably
over-optimizing and unless there are any objections I'll change the code to
handle memory set with no labels in the Utils class.
Thanks
---
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]
