Github user sachingoel0101 commented on the pull request:
https://github.com/apache/flink/pull/1083#issuecomment-138868165
I apologize. Like I said before, the way I've implemented this, it seemed
pretty trivial. Foremost, I needed this for something else and decided to take
a shot at it. Missed the finer details, going by your comments.
You're right though. Back-tracking the execution graph will most certainly
be a better solution.
@StephanEwen , am I correct in assuming this functionality will be added
after (or along with) Job Session management?
[As for the memory issue though, if data is to be persisted across jobs,
we'll always have to keep the memory segments allocated even after the job
finishes. This is why I defined a separate owner, which can be later *garbage
collected* through a cache freeing mechanism.]
As far as merging it is considered, I wasn't sure myself from the
beginning. I had actually written an email with a link to my branch for
opinions, but ended up opening a PR instead of sending that. So no worries. I'm
closing this.
---
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.
---
