Github user CodingCat commented on the pull request:
https://github.com/apache/spark/pull/228#issuecomment-40650644
Hi, @markhamstra @mridulm @kayousterhout @pwendell I was rethinking about
this PR
Sometimes user may use accumulator to measure the number of operations
which were performed, in this case, the accumulator value should be relevant
to the re-execution...and for the case, like K-means in mllib, we should not
make accumulator value change when the task is a resubmitted one...
So shall we provide a parameter for the user to indicate what they want
when they construct the accumulator?
---
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.
---
