Github user tdas commented on the pull request:
https://github.com/apache/spark/pull/8998#issuecomment-146016637
I strongly think that we should not add that require. It can break existing
streaming applications very easily. People might be using dynamic allocation
with direct kafka or kinesis or with anything + WAL, in which case it is fine
to have dynamic allocation as the system will not loose data. Upgrading Spark
will immediately break their application, and it might be hard to figure out
the solution for disabling dynamic allocation in their setting.
We have made mistakes in the past by throwing errors like this for usecases
which we didnt think would be useful in production, and people have complained
that it broke their testing stuff. See
https://issues.apache.org/jira/browse/SPARK-8630 and discussion
https://www.mail-archive.com/[email protected]&q=subject:%22Re%5C%3A+QueueStream+Does+Not+Support+Checkpointing%22&o=newest&f=1
We must not break stuff. So I suggest this be turned into a warning. In
that case we dont need a new configuration.
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
