[
https://issues.apache.org/jira/browse/FLUME-1019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13225923#comment-13225923
]
[email protected] commented on FLUME-1019:
------------------------------------------------------
bq. On 2012-03-08 00:32:59, Mike Percy wrote:
bq. > flume-ng-core/src/main/java/org/apache/flume/Sink.java, line 60
bq. > <https://reviews.apache.org/r/4175/diff/1/?file=88130#file88130line60>
bq. >
bq. > I think this part is a little vague. How about this?
bq. >
bq. > @return Returns READY if 1 or more events were successfully
delivered, or BACKOFF if no events could be retrieved from the channel feeding
this sink.
sounds good
bq. On 2012-03-08 00:32:59, Mike Percy wrote:
bq. > flume-ng-core/src/main/java/org/apache/flume/Sink.java, line 62
bq. > <https://reviews.apache.org/r/4175/diff/1/?file=88130#file88130line62>
bq. >
bq. > Maybe we should say:
bq. >
bq. > @throws EventDeliveryException if there is a failure delivering any
Event to the next-hop destination.
bq. >
bq. > (final destination is not guaranteed)
yeah, I see how that could be misunderstood, I had meant it as final iff the
sink was something like a file sink with no further hops. Taking that out
bq. On 2012-03-08 00:32:59, Mike Percy wrote:
bq. > flume-ng-core/src/main/java/org/apache/flume/SinkProcessor.java, line 31
bq. > <https://reviews.apache.org/r/4175/diff/1/?file=88131#file88131line31>
bq. >
bq. > Missing a component of the package: org.apache.flume.sink.SinkGroup
oops.
We may want to review which packages belong in org.apache.flume, I think some
of them don't belong there but in a sub-package. That's an issue for another
jira though
bq. On 2012-03-08 00:32:59, Mike Percy wrote:
bq. > flume-ng-core/src/main/java/org/apache/flume/SinkProcessor.java, line 37
bq. > <https://reviews.apache.org/r/4175/diff/1/?file=88131#file88131line37>
bq. >
bq. > That is an improvement but I think it kind of refers to the failover
sink processor's delivery policy (w.r.t. throwing the exception). Maybe
something like this would be more general?
bq. >
bq. > The processor is expected to call {@linkplain Sink#process()} on
whatever sink(s) appropriate, handling failures as appropriate and throwing
{@link EventDeliveryException} when there is a failure to deliver any events
according to the delivery policy defined by the sink processor implementation.
See specific implementations of this interface for delivery behavior and
policies.
bq. >
Much better, thanks
bq. On 2012-03-08 00:32:59, Mike Percy wrote:
bq. > flume-ng-core/src/main/java/org/apache/flume/SinkProcessor.java, line 41
bq. > <https://reviews.apache.org/r/4175/diff/1/?file=88131#file88131line41>
bq. >
bq. > These old constants should be corrected. How about:
bq. >
bq. > @return Returns {@code READY} if events were successfully consumed,
or {@code BACKOFF} if no events were available in the channel to consume.
bq. >
bq. > @throws EventDeliveryException if any failure to deliver events
occurred
Done. I changed the @throws a bit to refer to individual processor behaviors
- Juhani
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/4175/#review5698
-----------------------------------------------------------
On 2012-03-05 09:30:21, Juhani Connolly wrote:
bq.
bq. -----------------------------------------------------------
bq. This is an automatically generated e-mail. To reply, visit:
bq. https://reviews.apache.org/r/4175/
bq. -----------------------------------------------------------
bq.
bq. (Updated 2012-03-05 09:30:21)
bq.
bq.
bq. Review request for Flume.
bq.
bq.
bq. Summary
bq. -------
bq.
bq. An initial pass at documenting the interfaces.
bq. Let me know if I missed anything relevant, or if you feel that this does
not correctly represent our expected behaviors.
bq.
bq.
bq. This addresses bug FLUME-1019.
bq. https://issues.apache.org/jira/browse/FLUME-1019
bq.
bq.
bq. Diffs
bq. -----
bq.
bq. flume-ng-core/src/main/java/org/apache/flume/Sink.java 3abeeb6
bq. flume-ng-core/src/main/java/org/apache/flume/SinkProcessor.java 11651ad
bq. flume-ng-core/src/main/java/org/apache/flume/sink/SinkGroup.java 2e80a56
bq.
flume-ng-core/src/main/java/org/apache/flume/sink/SinkProcessorFactory.java
10f9f4e
bq.
bq. Diff: https://reviews.apache.org/r/4175/diff
bq.
bq.
bq. Testing
bq. -------
bq.
bq. No changes have been made to code
bq.
bq.
bq. Thanks,
bq.
bq. Juhani
bq.
bq.
> Document Sink and related interfaces, defining expected behaviors
> -----------------------------------------------------------------
>
> Key: FLUME-1019
> URL: https://issues.apache.org/jira/browse/FLUME-1019
> Project: Flume
> Issue Type: Improvement
> Components: Sinks+Sources
> Affects Versions: v1.0.0, v1.1.0
> Reporter: Juhani Connolly
> Assignee: Juhani Connolly
> Fix For: v1.1.0
>
> Attachments: FLUME-1019.patch
>
>
> Currently the Sink interface is undocumented and some of the other public
> interfaces that might be implemented by users for custom sinks/processors are
> poorly documented.
> In the process of documenting these interfaces, I hope to get input from
> others on what our expected behavior is.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
