Hi,
If it is to extend the Context to pass more information between the stages of
processing a window (triggering -> process -> eviction), why not adding also a
"EvictionInfo"? I think this might actually help with the issues discussed in
the tread related to the eviction policy. I could
Do Le Quoc created FLINK-4205:
-
Summary: Implement stratified sampling for DataSet
Key: FLINK-4205
URL: https://issues.apache.org/jira/browse/FLINK-4205
Project: Flink
Issue Type: New Feature
I added a FLIP document in the wiki:
https://cwiki.apache.org/confluence/display/FLINK/FLIP-1+%3A+Fine+Grained+Recovery+from+Task+Failures
For now, this contains the link to the Google Doc and a link to this
discussion thread. Once a Jira is created for this it should also be added
there.
On
Hi Till,
I have created the JIRA: https://issues.apache.org/jira/browse/FLINK-4205
Thank you,
Do
On Tue, Jul 12, 2016 at 6:05 PM, Till Rohrmann wrote:
> Stratified sampling would also be beneficial for the DataSet API. I think
> it would be best if this method is also
Hi all,
I just noticed some scaladocs (Gelly Scala API, Streaming Scala API, and
FlinkML) are missing in scaladoc page but found in javadoc page, even though
the APIs are for Scala. Is this intentional?
I think we have to move some documentation to scaladoc.
Regards,
Chiwan Park
Hi Aljoscha,
I agree, the user will know exactly that they are creating an EventTime
based evictor or ProcessingTime based evictor looking at the code.
So do you think it will be ok to have multiple versions of TimeEvictor (one
for event time and one for processing time) and also a DeltaEvcitor
Sumit Chawla created FLINK-4214:
---
Summary: JobExceptionsHandler will return all exceptions
Key: FLINK-4214
URL: https://issues.apache.org/jira/browse/FLINK-4214
Project: Flink
Issue Type: Bug
Thanks for this very first proposal! Both the proposed functionality
and the way you explained it are super nice. :-)
I think that this has been long overdue in Flink. :-) Having worked on
both the ExecutionGraph and IntermediateResults before, I agree that
these are the relevant components for
Aljoscha Krettek created FLINK-4207:
---
Summary: WindowOperator becomes very slow with allowed lateness
Key: FLINK-4207
URL: https://issues.apache.org/jira/browse/FLINK-4207
Project: Flink
Chesnay Schepler created FLINK-4206:
---
Summary: Metric names should alle special characters
Key: FLINK-4206
URL: https://issues.apache.org/jira/browse/FLINK-4206
Project: Flink
Issue Type:
+1
I really like the re-organization you just did [1]! :-) Thanks!
Is there a way to reflect this in the left-side navigation as well by
having all User/Contributors/... child pages grouped together?
[1] https://cwiki.apache.org/confluence/display/FLINK/Apache+Flink+Home
On Mon, Jul 11, 2016
I'm not sure, I'll try to figure that out.
On Wed, 13 Jul 2016 at 12:15 Ufuk Celebi wrote:
> +1
>
> I really like the re-organization you just did [1]! :-) Thanks!
>
> Is there a way to reflect this in the left-side navigation as well by
> having all User/Contributors/... child
Hi,
this is a proposal to introduce a new interface for the window function to
make it more extensible for the future where we might want to provide
additional information about why a window fired to the user function:
Sure, I also thought about this but went for the "extreme" initially. If
no-one objects I'll update the doc in a bit.
On Wed, 13 Jul 2016 at 14:17 Stephan Ewen wrote:
> Thanks for opening this.
>
> I see the need for having an extensible context object for window function
>
I deliberately kept it initially in the Google Doc because it is easier to
comment.
I would move it to the wiki once we have passed the initial discussion and
have converged (or are close to convergence) on the design.
On Tue, Jul 12, 2016 at 8:11 PM, Chesnay Schepler
wrote:
Thanks for opening this.
I see the need for having an extensible context object for window function
invocations, but i think hiding every parameter in the context is a bit
unnatural.
How about having a function "apply(Key, Values, WindowContext, Collector)" ?
It should be possible to write the
Chesnay Schepler created FLINK-4210:
---
Summary: Move close()/isClosed() out of MetricGroup interface
Key: FLINK-4210
URL: https://issues.apache.org/jira/browse/FLINK-4210
Project: Flink
Ismaël Mejía created FLINK-4209:
---
Summary: Docker image breaks with multiple NICs
Key: FLINK-4209
URL: https://issues.apache.org/jira/browse/FLINK-4209
Project: Flink
Issue Type: Improvement
Hi,
I think the way to go here is to add both an EventTimeEvictor and a
ProcessingTimeEvictor. The problem is that "isEventTime" cannot really be
determined. That's also the reason why there is an EventTimeTrigger and a
ProcessingTimeTrigger. It was just an oversight that the TimeEvictor does
not
Hi Aljoscha,
I am thinking of adding a method boolean isEventTime(); in the
EvictorContext apart from
long getCurrentProcessingTime();
MetricGroup getMetricGroup();
long getCurrentWatermark();
This method can be used to make the Evictor not iterate through all the
elements in TimeEvictor. There
Ismaël Mejía created FLINK-4208:
---
Summary: Support Running Flink processes in foreground mode
Key: FLINK-4208
URL: https://issues.apache.org/jira/browse/FLINK-4208
Project: Flink
Issue Type:
I still think it should be explicit in the class. For example, if you have
this code:
input
.keyBy()
.window()
.trigger(EventTimeTrigger.create())
.evictor(TimeTrigger.create())
the time behavior of the trigger is explicitly specified while the evictor
would dynamically adapt based on
Thank you for writing the proposal!
I agree with Chesnay. I thought one of the main motivations of the new FLIP
process is to move the discussion onto ASF infra, to record the decision
making process on the mailing list and to make ongoing activities more
visible to the community.
If we want to
Hi,
I was hoping to use the isEventTime method in the WindowAssigner to set
that information in the EvictorContext.
What do you think?.
Thanks and Regards,
Vishnu Viswanath,
On Wed, Jul 13, 2016 at 10:09 AM, Aljoscha Krettek
wrote:
> Hi,
> I think the way to go here is to
Stefan Richter created FLINK-4211:
-
Summary: Dynamic Properties not working for jobs submitted to Yarn
session
Key: FLINK-4211
URL: https://issues.apache.org/jira/browse/FLINK-4211
Project: Flink
+1
This was my original motivation. I saw a lot of discussion on the design
docs I initiated but none of this was reflected on the ML.
On Wed, 13 Jul 2016 at 17:11 Robert Metzger wrote:
> Thank you for writing the proposal!
>
> I agree with Chesnay. I thought one of the
26 matches
Mail list logo