> On Oct. 6, 2013, 8 p.m., Swapnil Ghike wrote:
> > core/src/main/scala/kafka/admin/ReassignPartitionsCommand.scala, lines 61-65
> > <https://reviews.apache.org/r/14496/diff/2/?file=361683#file361683line61>
> >
> >     Instead of asking the user to store the output of a dryrun into a JSON 
> > file, should the tool compute the dryrun output and use it to perform 
> > validation ?

That would be a better experience and since only one reassignment operation can 
ever be in progress, this could've been possible by reading the reassignment 
zookeeper path. However, we can't reliably do that right now since we overwrite 
that path when reassignment is completed. Ideally, if we can avoid rewriting 
that and report progress for the reassignment operation, it would be a better 
tooling experience. Probably we can fix that in a separate JIRA?


- Neha


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/14496/#review26714
-----------------------------------------------------------


On Oct. 5, 2013, 6:02 p.m., Jun Rao wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/14496/
> -----------------------------------------------------------
> 
> (Updated Oct. 5, 2013, 6:02 p.m.)
> 
> 
> Review request for kafka.
> 
> 
> Bugs: KAFKA-1073
>     https://issues.apache.org/jira/browse/KAFKA-1073
> 
> 
> Repository: kafka
> 
> 
> Description
> -------
> 
> kafka-1017; incorporating review feedback
> 
> 
> kafka-1017
> 
> 
> Diffs
> -----
> 
>   bin/kafka-check-reassignment-status.sh 
> 1f218585cddddb8bf58d8a85af38f368a49b27e5 
>   core/src/main/scala/kafka/admin/CheckReassignmentStatus.scala 
> 7e85f87e96dbddf4fd8785ae3960e8fe4813e8e5 
>   core/src/main/scala/kafka/admin/ReassignPartitionsCommand.scala 
> f333d29bf36bb7fdc66b3bf9af16e7ee19ad7e48 
> 
> Diff: https://reviews.apache.org/r/14496/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Jun Rao
> 
>

Reply via email to