Reducing Checkpoint Count for Chain Operator

2023-01-31 Thread Talat Uyarer via user
Hi, We have a job that is reading from kafka and writing some endpoints. The job does not have any shuffling steps. I implement it with multiple steps. Flink chained those operators in one operator in submission time. However I see all operators are doing checkpointing. Is there any way to crea

Re: How to add permission validation? flink reads and writes hive table。

2023-01-31 Thread yuxia
HI, melin li. Could you please explain a bit more about unified access check in flink? Best regards, Yuxia 发件人: "melin li" 收件人: "User" 发送时间: 星期三, 2023年 2 月 01日 下午 2:39:15 主题: How to add permission validation? flink reads and writes hive table。 flink supports both sql and jar types.H

RE: Processing watermarks in a broadcast connected stream

2023-01-31 Thread Schwalbe Matthias
Good Morning Sajjad, I’ve once had a similar problem. As you’ve found out, directly using KeyedBroadcastProcessFunction is a little tricky. What I ended up with instead is to use the rather new @PublicEvolving MultipleInputStreamOperator. It allows you to connect and process any (reasonable) num

How to add permission validation? flink reads and writes hive table。

2023-01-31 Thread melin li
flink supports both sql and jar types.How can we implement a unified access check in flink? spark supports extensions; flink lacks extensions.

Re: Custom catalog implementation - getting table schema for computed columns

2023-01-31 Thread yuxia
HI, > about the question can I assume that ResolvedCatalogTable will be always a > runtime type. Sorry for I don't really understand your question , why do you have such assumtion? Best regards, Yuxia 发件人: "Krzysztof Chmielewski" 收件人: "User" 发送时间: 星期六, 2023年 1 月 21日 上午 3:13:12 主题: R

Re: Custom catalog implementation - getting table schema for computed columns

2023-01-31 Thread yuxia
Hi. Just FYI, I have seen some catalogs are still use deprecated TableSchema in flink hive, Iceberg, etc connector. But it's in Flink plan to drop the deprecated table schema [1]. In long term, seems use new schema api is a better choice. If it's for the case of Catalog's createTable method,

Extremely long startup time for exactly_once kafka sink

2023-01-31 Thread Bobby Richard
When enabling exactly_once on my kafka sink, I am seeing extremely long initialization times (over 1 hour), especially after restoring from a savepoint. In the logs I see the job constantly initializing thousands of kafka producers like this: 2023-01-31 14:39:58,150 INFO org.apache.kafka.clients.p

Re: beam + flink + k8

2023-01-31 Thread Jan Lukavský
The script looks good to me, did you run the SDK harness? External environment needs the SDK harness to be run externally, see [1]. Generally, the best option is DOCKER, but that usually does not work in k8s. For this, you might try PROCESS environment and build your own docker image for flink,

Re: Using pyflink from flink distribution

2023-01-31 Thread Andrew Otto
Great, thank you so much for your responses. It all makes sense now. :) On Mon, Jan 30, 2023 at 10:41 PM Dian Fu wrote: > >> What is the reason for including > opt/python/{pyflink.zip,cloudpickle.zip,py4j.zip} in the base > distribution then? Oh, a guess: to make it easier for TaskManagers to

Re: "Error while retrieving the leader gateway" when using Kubernetes HA

2023-01-31 Thread Anton Ippolitov via user
Makes sense, thank you! On Tue, Jan 31, 2023 at 10:48 AM Gyula Fóra wrote: > Thanks @Anton Ippolitov > At this stage I would highly recommend the native mode if you have the > liberty to try that. > I think that has better production characteristics and will work out of > the box with the autos

Re: beam + flink + k8

2023-01-31 Thread Jan Lukavský
Hi, can you please share the also the script itself? I'd say that the problem is that the flink jobmanager is not accessible through localhost:8081, because it runs inside the minikube. You need to expose it outside of the minikube via [1], or run the script from pod inside the minikube and a

Re: "Error while retrieving the leader gateway" when using Kubernetes HA

2023-01-31 Thread Gyula Fóra
Thanks @Anton Ippolitov At this stage I would highly recommend the native mode if you have the liberty to try that. I think that has better production characteristics and will work out of the box with the autoscaler. (the standalone mode won't) Gyula On Tue, Jan 31, 2023 at 10:41 AM Anton Ippoli

Re: "Error while retrieving the leader gateway" when using Kubernetes HA

2023-01-31 Thread Anton Ippolitov via user
I am using the Standalone Mode indeed, should've mentioned it right away. This fix looks exactly like what I need, thank you!! On Tue, Jan 31, 2023 at 9:16 AM Gyula Fóra wrote: > There is also a pending fix for the standalone + k8s HA case : > https://github.com/apache/flink-kubernetes-operator/

Flink 1.16 arm64/v8 docker images

2023-01-31 Thread Roberts, Ben (Senior Developer) via user
Hi, Is it possible for the arm64/v8 architecture images to be published for >1.16 flink-docker (apache/flink)? I’m aware that the official docker flink image is now published in the arm64 arch, but that image doesn’t include a JDK, so it’d be super helpful to have the apache/flink images publi

Re: "Error while retrieving the leader gateway" when using Kubernetes HA

2023-01-31 Thread Gyula Fóra
There is also a pending fix for the standalone + k8s HA case : https://github.com/apache/flink-kubernetes-operator/pull/518 You could maybe try and review the fix :) Gyula On Tue, Jan 31, 2023 at 8:36 AM Yang Wang wrote: > I assume you are using the standalone mode. Right? > > For the native K

Re: [ANNOUNCE] FRocksDB 6.20.3-ververica-2.0 released

2023-01-31 Thread Leonard Xu
Thanks Yanfei for driving the release ! ! Best, Leonard > On Jan 31, 2023, at 3:43 PM, Yun Tang wrote: > > Thanks Yuanfei for driving the frocksdb release! > > Best > Yun Tang > From: Yuan Mei > Sent: Tuesday, January 31, 2023 15:09 > To: Jing Ge > Cc: Yanfei Lei ; d...@flink.apache.org >