Stephan Ewen created FLINK-3370:
---
Summary: Add an aligned version of the window operator
Key: FLINK-3370
URL: https://issues.apache.org/jira/browse/FLINK-3370
Project: Flink
Issue Type:
Stephan Ewen created FLINK-3371:
---
Summary: Move TriggerCotext and TriggerResult to their own classes
Key: FLINK-3371
URL: https://issues.apache.org/jira/browse/FLINK-3371
Project: Flink
Issue
I created FLINK-3366 for renaming @Experimental to @PublicEvolving and
FLINK-3367 to annotate all remaining API classes with @PublicEvolving.
2016-02-08 9:57 GMT+01:00 Maximilian Michels :
> Never thought about "Experimental" meaning unstable but I agree it
> sounds better to
Fabian Hueske created FLINK-3367:
Summary: Annotate all user-facing API classes with @Public or
@PublicEvolving
Key: FLINK-3367
URL: https://issues.apache.org/jira/browse/FLINK-3367
Project: Flink
+1 (binding)
- build against Scala 2.11
- start-local (CLI, Webclient, Dashboard, log files, examples with build-in
and external data)
- start-cluster, local (2 JVMs, CLI + Webclient, Dashboard, log files,
examples with build-in and external data)
2016-02-07 20:21 GMT+01:00 Ufuk Celebi
Fabian Hueske created FLINK-3366:
Summary: Rename @Experimental annotation to @PublicEvolving
Key: FLINK-3366
URL: https://issues.apache.org/jira/browse/FLINK-3366
Project: Flink
Issue Type:
Regarding FLINK-2237 (hash-based combiner), I think we need a bit more
time.
It is a fairly large contribution and touches/adds core functionality.
I started reviewing the PR and will suggest a few changes in the next days.
2016-02-08 11:48 GMT+01:00 Robert Metzger :
>
Ufuk Celebi created FLINK-3364:
--
Summary: Don't initialize SavepointStore in JobManager constructor
Key: FLINK-3364
URL: https://issues.apache.org/jira/browse/FLINK-3364
Project: Flink
Issue
Great, thank you very much!
On Monday, February 8, 2016, Maximilian Michels wrote:
> Hi Stefano,
>
> Thanks for the pull request.
>
> You may use the following to only run tests from flink-runtime:
> > mvn verify -pl flink-runtime
>
> Alternatively, change to the flink-runtime
Stephan Ewen created FLINK-3363:
---
Summary: JobManager does not shut down dedicated executor properly
Key: FLINK-3363
URL: https://issues.apache.org/jira/browse/FLINK-3363
Project: Flink
Issue
Okay, I didn't check the individual PRs closely. I agree that we should not
merge big core changes if we are not certain about their stability.
On Mon, Feb 8, 2016 at 12:04 PM, Fabian Hueske wrote:
> Regarding FLINK-2237 (hash-based combiner), I think we need a bit more
>
Hi Stefano,
1) Please open a pull request. If the String depends on the locale,
this looks like a bug.
2) "mvn clean install" is the way to go for the complete check. If you
only want to run certain tests, selecting and running them from
IntelliJ works pretty well. In addition, it is nice to
Hi Stefano,
Thanks for the pull request.
You may use the following to only run tests from flink-runtime:
> mvn verify -pl flink-runtime
Alternatively, change to the flink-runtime directory and run "mvn
verify" from there.
This assumes Flink has been built and installed locally before. If
not,
There are still some 8 open blockers for the 1.0 release:
https://issues.apache.org/jira/issues/?jql=project%20%3D%20FLINK%20AND%20resolution%20%3D%20Unresolved%20AND%20priority%20%3D%20Blocker%20ORDER%20BY%20key%20DESC
I also think that there are some pull requests which are almost ready to
Never thought about "Experimental" meaning unstable but I agree it
sounds better to use the term "Evolving".
On Sun, Feb 7, 2016 at 12:45 AM, Fabian Hueske wrote:
> I agree, Experimental rather suggests unstable behavior than potentially
> changing interfaces.
> +1 for
Stephan Ewen created FLINK-3365:
---
Summary: BlobLibraryCacheManager does not shutdown Timer thread
Key: FLINK-3365
URL: https://issues.apache.org/jira/browse/FLINK-3365
Project: Flink
Issue
I like the idea to support partial functions with Flinkās Scala API.
However, I think that breaking the API and making it inconsistent with
respect to the Java API is not the best option. I would rather be in favour
of the first proposal where we add a new method xxxWith via implicit
conversions.
-1 (binding)
the quickstart for 0.10.2 has the Flink version set to 0.10-SNAPSHOT, not
to 0.10.2.
https://repository.apache.org/content/repositories/orgapacheflink-1060/org/apache/flink/flink-quickstart-java/0.10.2/flink-quickstart-java-0.10.2.jar
The same applies to the 0.10.2-hadoop1 version:
Ufuk Celebi created FLINK-3369:
--
Summary: RemoteTransportException should be instance of
CancelTaskException
Key: FLINK-3369
URL: https://issues.apache.org/jira/browse/FLINK-3369
Project: Flink
Robert Metzger created FLINK-3368:
-
Summary: Kafka 0.8 consumer fails to recover from broker shutdowns
Key: FLINK-3368
URL: https://issues.apache.org/jira/browse/FLINK-3368
Project: Flink
> On 08 Feb 2016, at 15:13, Robert Metzger wrote:
>
> -1 (binding)
>
> the quickstart for 0.10.2 has the Flink version set to 0.10-SNAPSHOT, not
> to 0.10.2.
>
> On 08 Feb 2016, at 13:57, Chesnay Schepler wrote:
>
> Hello all,
>
> The final flag on the
> org.apache.flink.runtime.io.network.api.writer.ResultPartitionWriter class is
> causing issues for me.
>
> The flag requires me to run a test I'm working on with a
>
Dear Flink community,
Please vote on releasing the following candidate as Apache Flink version
0.10.2.
Please note that this vote has a slightly shorter voting period of 48
hours. Only a single change has been made since the last release
candidate. Since the community has already done extensive
This vote is cancelled in favor of RC2.
On Mon, Feb 8, 2016 at 3:18 PM, Ufuk Celebi wrote:
>
>> On 08 Feb 2016, at 15:13, Robert Metzger wrote:
>>
>> -1 (binding)
>>
>> the quickstart for 0.10.2 has the Flink version set to 0.10-SNAPSHOT, not
>> to 0.10.2.
Is it possible to force operator chaining to be disabled? Similar to how
object reuse can be enabled or disabled?
Greg
Nick Dimiduk created FLINK-3372:
---
Summary: Setting custom YARN application name is ignored
Key: FLINK-3372
URL: https://issues.apache.org/jira/browse/FLINK-3372
Project: Flink
Issue Type: Bug
When is this useful in streaming?
On Mon, Feb 8, 2016 at 1:46 PM, Nick Dimiduk wrote:
>
> https://ci.apache.org/projects/flink/flink-docs-release-0.10/api/java/org/apache/flink/streaming/api/environment/StreamExecutionEnvironment.html#disableOperatorChaining()
>
> On Mon,
There is no fundamental reason to not implement this for batch as well.
In Streaming, users seem to want more control about threads and resources
(given that these are often continuous pipelines), while in batch that was
not requested so far.
But I see that a non-chained function is more safe
https://ci.apache.org/projects/flink/flink-docs-release-0.10/api/java/org/apache/flink/streaming/api/environment/StreamExecutionEnvironment.html#disableOperatorChaining()
On Mon, Feb 8, 2016 at 10:34 AM, Greg Hogan wrote:
> Is it possible to force operator chaining to be
Perhaps too late for the RC, but I've backported FLINK-3293 to this branch
via FLINK-3372. Would be nice for those wanting to monitory yarn
application submissions.
On Mon, Feb 8, 2016 at 9:37 AM, Ufuk Celebi wrote:
> Dear Flink community,
>
> Please vote on releasing the
It may be useful for example when you have two MapFunctions and each does
something CPU intensive, or communicates with an external service.
Without chaining, you will have two threads and an elastic channel between
the functions to buffer some records, which may help in such a case.
On Mon,
31 matches
Mail list logo