Github user steveloughran commented on the issue:
https://github.com/apache/spark/pull/17364
I haven't reviewed that bit of code: make it a separate JIRA and assign to
me. This one I came across in the HADOOP-2.8.0 RC3 testing; the underlying fix
there is going in, but the spark code still should be made more resilient to
failure here. It's always those failure modes which get you âand working with
S3, that close() can be where the PUT is initiated, so P(fail) > 0.
Part of [HADOOP-13786]
(https://issues.apache.org/jira/browse/HADOOP-13786) is MAPREDUCE-6823: making
it straighforward to define a different implementation of the FileOutputFormat
committer. That should make it easier to do some fault injection in the commit
processes, especially all those bits that violate the state machine entirely.
I'll see what I can do about breaking things :)
---
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
