Github user vanzin commented on the issue:
https://github.com/apache/spark/pull/17723
> For example, use of UGI for spark security in core.
UGI, as far as I understand, is being used mostly as a credential store and
wrapper around the JAAS / GSSAPI code. Replicating that code would be very hard
and error prone.
If at some point in the future someone decides they want to support some
different paradigm, I think that would be the time to think about abstracting
that stuff. Right now I don't really see why using UGI outside of YARN is bad
(aside from assumptions that the Hadoop API makes and Michael has run into
already).
> I was actually surprised the mesos is.
It's not that Mesos is compatible. It's that as far as Spark is concerned,
UGI is mostly a container for credentials, and that's how it's being used here.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
