Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/629#issuecomment-42616622
So this means that the Java-serialized format of a Spark class has changed
(or at least, it may have; the serialVersionUID changed) between the versions
of two Spark components that you are using. I doubt it is related to this PR
per se, but is related to having two versions of Spark in play -- what you just
compiled, and whatever else may be on your machines? I see you distributed the
updated code to some machines, but maybe not all, or, are you sure the running
processes are all the same new version? At the least, this error suggests it
isn't.
If so the fix (which is a separate issue) is to really run the same version
of everything. I suppose it may be worth looking into whether the serialized
form of this class really changed or not and whether management of its
serialVersionUID might resolve what is really not a problem, but that's less
good than matching versions everywhere, which should ensure this can't happen.
---
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.
---
