xiaogang zhou created FLINK-21688:
-
Summary: we use setIncreaseParallelism function, can cause slow
flush in restore
Key: FLINK-21688
URL: https://issues.apache.org/jira/browse/FLINK-21688
Project:
Chesnay Schepler created FLINK-21689:
Summary: UnalignedCheckpointITCase does not terminate
Key: FLINK-21689
URL: https://issues.apache.org/jira/browse/FLINK-21689
Project: Flink
Issue
Matthias created FLINK-21687:
Summary: e2e "Run kubernetes session test (default input)" timed
out due to k8s issue
Key: FLINK-21687
URL: https://issues.apache.org/jira/browse/FLINK-21687
Project: Flink
Hello, fellows, long time no see ~
Here i want to fire a discussion about the SQL implicit type coercion, a
required feature for SQL production but missed for Flink SQL for a long
time ~
SQL implicit type coercion is very useful for these cases:
- the inter-operation within builtin SQL
It was originally designed that the framework vs. task memory are
distinguished by whether the memory owned by the slots. We have not
emphasized this in user documents because the dynamic slot allocation is
not yet available to the users, thus they do not need to worry about
whether the memory is
Hi, all
Thanks all for your suggestions and feedback.
I think it is a good idea that we increase the default size of the
separated pool by testing. I am fine with adding the suffix(".size") to the
config name, which makes it more clear to the user.
But I am a little worried about adding a
Igal Shilman created FLINK-21692:
Summary: Update the Python SDK to reflect the latest changes in
the request-reply protocol
Key: FLINK-21692
URL: https://issues.apache.org/jira/browse/FLINK-21692
Hi Jan,
if you run mvn install of your modified Flink, then you can simply use the
maven artefacts from the local repository. You should then either specify
--no-snapshot-updates or assign a different maven version to your Flink
build which you can reference from your test job.
Cheers,
Till
On
Thanks for creating the draft for the updated roadmap Stephan. It looks
good to me. I also like the feature radar. This will help our users to
navigate more safely in the growing Flink ecosystem.
Cheers,
Till
On Mon, Mar 8, 2021 at 8:32 PM Stephan Ewen wrote:
> Thanks, Robert!
>
> As far as I
Zhengqi Zhang created FLINK-21690:
-
Summary: remove redundant tolerableCheckpointFailureNumber setting
in CheckpointConfig
Key: FLINK-21690
URL: https://issues.apache.org/jira/browse/FLINK-21690
Till Rohrmann created FLINK-21691:
-
Summary: KafkaSource fails with NPE when setting it up
Key: FLINK-21691
URL: https://issues.apache.org/jira/browse/FLINK-21691
Project: Flink
Issue Type:
Till Rohrmann created FLINK-21693:
-
Summary: TestStreamEnvironment does not implement executeAsync
Key: FLINK-21693
URL: https://issues.apache.org/jira/browse/FLINK-21693
Project: Flink
Stephan Ewen created FLINK-21694:
Summary: Increase default value of
"state.backend.rocksdb.checkpoint.transfer.thread.num"
Key: FLINK-21694
URL: https://issues.apache.org/jira/browse/FLINK-21694
@Stephan: Indeed, we shouldn't add unreleased features there.
+1 to update after each release.
On Tue, Mar 9, 2021 at 3:05 PM Till Rohrmann wrote:
> Thanks for creating the draft for the updated roadmap Stephan. It looks
> good to me. I also like the feature radar. This will help our users to
>
Stephan Ewen created FLINK-21695:
Summary: Increase default value for number of KeyGroups
Key: FLINK-21695
URL: https://issues.apache.org/jira/browse/FLINK-21695
Project: Flink
Issue Type:
godfrey he created FLINK-21697:
--
Summary: Support StreamExecMiniBatchAssigner json
serialization/deserialization
Key: FLINK-21697
URL: https://issues.apache.org/jira/browse/FLINK-21697
Project: Flink
Hi Heng Du,
Thanks for starting the discussion.
>From the user's perspective, the integration between RocketMQ and Flink is
user-friendly. I agree with you the point that polish the existing
implementation and add new features, such as table connector
support. But before that, you need to
Thank you all for your positive feedbacks.
Considering that you’re binding votes for this FLIP and all of us agreed to
remove this option,
Thus I update this section of FLIP-162[1], I’ll start work on it next .
Best,
Leonard
[1]
Dear Flink community,
Thanks for @Nicholas's reply and bridge, I am a PMC member of RocketMQ, and
developers in the RocketMQ community have been concerned about integration
with Flink for a long time, and sincerely hope developers from both
communities will work closely together to make this
Apache RocketMQ[1], as a cloud-native messaging and streaming platform, has
a large number of users [2] all over the world, especially in China.
Recently, more and more users have begun to use RocketMQ to integrate with
Flink, which has resulted in a large number of issues[3] and PRs[4]. So,
Hi Statefun folk,
Now that we have a powerful framework to manage stateful functions, a part
of the algorithmic complexity has moved on the interaction between
functions side. But implementing an interaction protocol is complex and
error prone, so that could be great to provide an
Giacomo Gamba created FLINK-21696:
-
Summary: scala code in "Writing WatermarkGenerators" documentation
missing/deprecated
Key: FLINK-21696
URL: https://issues.apache.org/jira/browse/FLINK-21696
Wenlong Lyu created FLINK-21704:
---
Summary: Support StreamExecChangelogNormalize json serde
Key: FLINK-21704
URL: https://issues.apache.org/jira/browse/FLINK-21704
Project: Flink
Issue Type:
Tzu-Li (Gordon) Tai created FLINK-21705:
---
Summary: Add a SliceType to the StateFun Java SDK
Key: FLINK-21705
URL: https://issues.apache.org/jira/browse/FLINK-21705
Project: Flink
Issue
tonychan created FLINK-21699:
Summary: SecurityUtils JaasModule make JVM dirty
Key: FLINK-21699
URL: https://issues.apache.org/jira/browse/FLINK-21699
Project: Flink
Issue Type: Bug
Shengkai Fang created FLINK-21702:
-
Summary: Support option `sql-client.verbos` to print the exception
stack
Key: FLINK-21702
URL: https://issues.apache.org/jira/browse/FLINK-21702
Project: Flink
Shengkai Fang created FLINK-21703:
-
Summary: Add option `table.dml-sync`
Key: FLINK-21703
URL: https://issues.apache.org/jira/browse/FLINK-21703
Project: Flink
Issue Type: Sub-task
Leonard Xu created FLINK-21698:
--
Summary: Disable problematic cast conversion between NUMERIC type
and TIMESTAMP type
Key: FLINK-21698
URL: https://issues.apache.org/jira/browse/FLINK-21698
Project:
Junfan Zhang created FLINK-21700:
Summary: Allow to disable fetching Hadoop delegation token on Yarn
Key: FLINK-21700
URL: https://issues.apache.org/jira/browse/FLINK-21700
Project: Flink
Shengkai Fang created FLINK-21701:
-
Summary: Extend the `RESET` command in the sql client
Key: FLINK-21701
URL: https://issues.apache.org/jira/browse/FLINK-21701
Project: Flink
Issue Type:
Dawid Wysakowicz created FLINK-21706:
Summary:
FlinkKafkaProducerITCase.testMigrateFromAtExactlyOnceToAtLeastOnce fails with
"SchemaException: Error reading field 'api_keys': Error reading field
'api_key': java.nio.BufferUnderflowException"
Thank you Stephan for working on this. It's really nice to see the
roadmap updated. I also find the "Feature Radar" extremely helpful. It
looks good to be published from my side.
Best,
Dawid
On 02/03/2021 15:38, Stephan Ewen wrote:
> Hi all!
>
> The roadmap on the Flink website is quite
Zhu Zhu created FLINK-21707:
---
Summary: Job is possible to hang when restarting a FINISHED task
with POINTWISE BLOCKING consumers
Key: FLINK-21707
URL: https://issues.apache.org/jira/browse/FLINK-21707
humengyu created FLINK-21686:
Summary: Duplicate code in hive parser file should be abstracted
into functions
Key: FLINK-21686
URL: https://issues.apache.org/jira/browse/FLINK-21686
Project: Flink
Thanks Konstantin for driving this topic.
Generally +1 for the proposal, I went through the doc and have two concerns
here.
Will the robot send all notifications to assignee/reporter/watchers ?
I’m a little worried about too many push messages. Eg, I watched some
issues that I want to
Thanks Leonard,
I'm also +1 with not introducing this fallback option.
It's error-prone to mix the implementation of wrong behavior and correct
behavior.
And it's better to educate users the right way in one version instead of
spanning multiple versions.
Best,
Jark
On Tue, 9 Mar 2021 at 15:15,
Hi Leonard,
I'm fine with dropping the old buggy behavior immediatly. Users can
still implement a UDF with the old bavhior if needed. I hope the new
functions will be well-tested so that a fallback to the old functions is
not necessary as a workaround. It will definitely avoid confusion for
+1
Let's go straight to the right behavior. Drop the option for the wrong
behavior.
Best,
Jingsong
On Tue, Mar 9, 2021 at 4:29 PM Timo Walther wrote:
> Hi Leonard,
>
> I'm fine with dropping the old buggy behavior immediatly. Users can
> still implement a UDF with the old bavhior if needed.
38 matches
Mail list logo