Hi Till,
Thank you for your email! Please find my comments inline.
On Mon, Dec 28, 2020 at 5:50 AM Till Rohrmann wrote:
> Hi Jianyun,
>
> Thanks a lot for reviving this discussion. I think it would be great to
> have a well working Pulsar connector for Flink. Before diving into the
> detailed
Hi Arvid,
Very thanks for the deep thoughts !
> If this somehow works, we would not need to change much in the checkpoint
> coordinator. He would always inject into sources. We could also ignore the
> race conditions as long as the TM lives. Checkpointing times are also not
> worse as with the
Jark Wu created FLINK-20877:
---
Summary: Refactor BytesHashMap and BytesMultiMap to support window
key
Key: FLINK-20877
URL: https://issues.apache.org/jira/browse/FLINK-20877
Project: Flink
Issue
Wenlong Lyu created FLINK-20876:
---
Summary: Separate the implementation of StreamExecTemporalJoin
Key: FLINK-20876
URL: https://issues.apache.org/jira/browse/FLINK-20876
Project: Flink
Issue
Hi, there,
We would like to start a discussion thread on "FLIP-156: Runtime
Interfaces for Fine-Grained Resource Requirements"[1], where we
propose Slot Sharing Group (SSG) based runtime interfaces for
specifying fine-grained resource requirements.
In this FLIP:
- Expound the user story of
Wong Mulan created FLINK-20875:
--
Summary: Could patch CVE-2020-17518 to version 1.10
Key: FLINK-20875
URL: https://issues.apache.org/jira/browse/FLINK-20875
Project: Flink
Issue Type: Bug
Huang Xingbo created FLINK-20874:
Summary: Python DataStreamTests.test_key_by_on_connect_stream test
failed with "ArrayIndexOutOfBoundsException"
Key: FLINK-20874
URL:
Jark Wu created FLINK-20873:
---
Summary: Upgrade Calcite version to 1.27
Key: FLINK-20873
URL: https://issues.apache.org/jira/browse/FLINK-20873
Project: Flink
Issue Type: Improvement
Thanks for updating the design doc.
It looks good to me.
Best,
Jark
On Thu, 7 Jan 2021 at 10:16, Jingsong Li wrote:
> Sounds good to me.
>
> We don't have to worry about future changes, because it has covered all
> the capabilities of calcite aggregation.
>
> Best,
> Jingsong
>
> On Thu, Jan
Liu created FLINK-20872:
---
Summary: Job resume from history savepoint when failover if
checkpoint is disabled
Key: FLINK-20872
URL: https://issues.apache.org/jira/browse/FLINK-20872
Project: Flink
Sounds good to me.
We don't have to worry about future changes, because it has covered all the
capabilities of calcite aggregation.
Best,
Jingsong
On Thu, Jan 7, 2021 at 12:14 AM Sebastian Liu wrote:
> Hi Jark,
>
> Sounds good to me. For better scalability in the future, we could add the
>
Steven Zhen Wu created FLINK-20871:
--
Summary: Make DataStream#executeAndCollectWithClient public
Key: FLINK-20871
URL: https://issues.apache.org/jira/browse/FLINK-20871
Project: Flink
Issue
Hi Till,
Thank you for your reply. I found such a chain of method calls:
JobMaster#startScheduling -> SchedulerBase#startScheduling ->
DefaultScheduler#startSchedulingInternal ->
EagerSchedulingStrategy#startScheduling ->
EagerSchedulingStrategy#allocateSlotsAndDeploy ->
xx chai created FLINK-20870:
---
Summary: FlinkKafkaSink
Key: FLINK-20870
URL: https://issues.apache.org/jira/browse/FLINK-20870
Project: Flink
Issue Type: Improvement
Components: API /
+1 As Yupeng mentioned, we at Uber are also looking into the Pinot Sink. It
would be great to collaborate on this proposal.
Thanks,
Sanath
On Wed, Jan 6, 2021 at 9:23 AM Yupeng Fu wrote:
> Hi Mats,
>
> Glad to see this interest! We at Uber are also working on a Pinot sink (for
> BATCH
Hi,
I see that the last update to the roadmap for FlinkML was some time ago
(2016). Is development still active? If so, I would like to contribute some
unsupervised clustering algorithms like CLARANS. Would love some pointers.
Thanks,
Badrul
Timo Walther created FLINK-20869:
Summary: Support TIMESTAMP WITH TIME ZONE
Key: FLINK-20869
URL: https://issues.apache.org/jira/browse/FLINK-20869
Project: Flink
Issue Type: Sub-task
Hi Mats,
Glad to see this interest! We at Uber are also working on a Pinot sink (for
BATCH execution), with some help from the Pinot community on abstracting
Pinot interfaces for segment writes and catalog retrieval. Perhaps we can
collaborate on this proposal/POC.
Cheers,
Yupeng
On Wed, Jan
Okay then at least you guys are in sync ;) (Although I'm also not too far
away)
I hope I'm not super derailing but could we reiterate why it's good to get
rid of finished tasks (note: I'm also mostly in favor of that):
1. We can free all acquired resources including buffer pools, state
Piotr Nowojski created FLINK-20868:
--
Summary: Pause the idle/back pressure timers during processing
mailbox actions
Key: FLINK-20868
URL: https://issues.apache.org/jira/browse/FLINK-20868
Project:
Hi Jark,
Sounds good to me. For better scalability in the future, we could add the
AggregateExpression.
```
public class AggregateExpression implements ResolvedExpression {
private final FunctionDefinition functionDefinition;
private final List args;
private final @Nullable
Ruguo Yu created FLINK-20867:
Summary: support ship files from local/remote path to k8s pod
Key: FLINK-20867
URL: https://issues.apache.org/jira/browse/FLINK-20867
Project: Flink
Issue Type: New
On 2021/01/06 16:05, Arvid Heise wrote:
thanks for the detailed example. It feels like Aljoscha and you are also
not fully aligned yet. For me, it sounded as if Aljoscha would like to
avoid sending RPC to non-source subtasks.
No, I think we need the triggering of intermediate operators.
I was
On 2021/01/06 13:35, Arvid Heise wrote:
I was actually not thinking about concurrent checkpoints (and actually want
to get rid of them once UC is established, since they are addressing the
same thing).
I would give a yuge +1 to that. I don't see why we would need concurrent
checkpoints in
Hi Yun,
thanks for the detailed example. It feels like Aljoscha and you are also
not fully aligned yet. For me, it sounded as if Aljoscha would like to
avoid sending RPC to non-source subtasks.
I think we are still on the same page that we would like to trigger
> checkpoint periodically until
Hi Arvid,
Very thanks for the feedbacks! I'll try to answer the questions inline:
> I'm also concerned about the notion of a final checkpoint. What happens
> when this final checkpoint times out (checkpoint timeout > async timeout)
> or fails for a different reason? I'm currently more inclined
That's good to hear. I wasn't sure because the explanation focused a lot
on checkpoints and the details of it while with the new Sink interface
implementers don't need to be concerned with those. And in fact, when
the Sink is used in BATCH execution mode there will be no checkpoints.
Other
Hi Liu, Jingsong,
Regarding the agg with filter, I think in theory we can support pushing
such a pattern into source.
We don't need to support it in the first version, but in the long term, we
can support it.
The designed interface should be future proof.
Considering filter arg and distinct flag
Yang Wang created FLINK-20866:
-
Summary: Add how to list jobs in Yarn deployment documentation
when HA enabled
Key: FLINK-20866
URL: https://issues.apache.org/jira/browse/FLINK-20866
Project: Flink
I was actually not thinking about concurrent checkpoints (and actually want
to get rid of them once UC is established, since they are addressing the
same thing).
But your explanation definitely helped me to better understand the race
condition.
However, I have the impression that you think
Thanks for driving the docker image efforts, Robert.
+1 for canceling this RC. That should also give us the chance to fix
FLINK-20781.
I'll try to create the next release candidate as soon as the following
blockers are resolved.
* FLINK-20781: UnalignedCheckpointITCase failure caused by
Hi Jingsong, Jark,
Thx so much for our discussion, and the cases mentioned above are really
worthy for further discussion.
1. For aggregate with filter expressions: eg: select COUNT(1) FILTER(WHERE
cc_call_center_sk > 3) from call_center;
For the current Blink Planner, the optimized plan will
The Docker images for Flink 1.12.0 are still not on Docker Hub (the
official images), because they were not accepted by Docker.
To avoid such issues in the future, we've decided to additionally publish
the Docker images on GitHub as part of the release process (immediately)
[1]. We still intend to
Yes, we will use the latest sink interface.
Best,
Mats
> On 6. Jan 2021, at 11:05, Aljoscha Krettek wrote:
>
> It's great to see interest in this. Where you planning to use the new Sink
> interface that we recently introduced? [1]
>
> Best,
> Aljoscha
>
> [1] https://s.apache.org/FLIP-143
>
On 2021/01/06 11:30, Arvid Heise wrote:
I'm assuming that this is the normal case. In a A->B graph, as soon as A
finishes, B still has a couple of input buffers to process. If you add
backpressure or longer pipelines into the mix, it's quite likely that a
checkpoint may occur with B being the
Hi Till,
1) Who from the Flink community will mentor this effort and could take
> responsibility for it?
>
I'd be happy to mentor the transition. It remains to be seen who is doing
mainly the maintenance in the long run. If all fails, I can also take that
over but I was hoping that the
Yangze Guo created FLINK-20865:
--
Summary: Prevent potential resource deadlock in fine-grained
resource management
Key: FLINK-20865
URL: https://issues.apache.org/jira/browse/FLINK-20865
Project: Flink
>
> I was referring to the case where intermediate operators don't have any
> active upstream (input) operators. In that case, they basically become
> the "source" of that part of the graph. In your example, M1 is still
> connected to a "real" source.
I'm assuming that this is the normal case.
It's great to see interest in this. Where you planning to use the new
Sink interface that we recently introduced? [1]
Best,
Aljoscha
[1] https://s.apache.org/FLIP-143
On 2021/01/05 12:21, Poerschke, Mats wrote:
Hi all,
we want to contribute a sink connector for Apache Pinot. The following
On 2021/01/05 17:27, Arvid Heise wrote:
For your question: will there ever be intermediate operators that should be
running that are not connected to at least once source?
I think there are plenty of examples if you go beyond chained operators and
fully connected exchanges. Think of any fan-in,
Yangze Guo created FLINK-20864:
--
Summary: Apply exact matching rules in fulfilling resource
requirement with slot resource
Key: FLINK-20864
URL: https://issues.apache.org/jira/browse/FLINK-20864
Yangze Guo created FLINK-20863:
--
Summary: Exclude network memory from ResourceProfile
Key: FLINK-20863
URL: https://issues.apache.org/jira/browse/FLINK-20863
Project: Flink
Issue Type: Task
Hi Penguin,
What do you wanna do? If you want to change Flink's scheduling behaviour,
then you can take a look at the implementations of SchedulerNG.
Cheers,
Till
On Wed, Jan 6, 2021 at 6:58 AM penguin. wrote:
> Hello! Do you know how to modify the task scheduling method of Flink?
Timo Walther created FLINK-20862:
Summary: Add a converter from TypeInformation to DataType
Key: FLINK-20862
URL: https://issues.apache.org/jira/browse/FLINK-20862
Project: Flink
Issue Type:
I wanted to check in again to see if a discussion has started around
releasing 1.10.3.
There are a few patches[0] in 1.10.3 that we are very eager to pick up.
[0] https://issues.apache.org/jira/browse/FLINK-15467 and
https://issues.apache.org/jira/browse/FLINK-19237
--
Sent from:
Hi Arvid,
Very thanks for the feedbacks!
> For 2) the race condition, I was more thinking of still injecting the
> barrier at the source in all cases, but having some kind of
short-cut to
> immediately execute the RPC inside the respective taskmanager.
46 matches
Mail list logo