Re: Apache Flink - Reading data from Scylla DB

Seems you may need implement a custom connector for Scylla DB as I haven't found a connector on hand. Hope the doc[1][2] can help you implement your own connector. [1] https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/sourcessinks/ [2]

Flink operator deletes the FlinkDeplyoment after a while

after few hours of running job manager and task manager generated using the operator i get the following message in the operator log There really wasn't any traffic and the flink deployment is being delete === Finished metrics report === Deleting

Apache Flink - Reading data from Scylla DB

Team, I'm looking for a solution to Consume/Read data from Scylla DB into Apache Flink. If anyone can guide me or share pointers it will be helpful. Regards, Himanshu

Re: FW: FW: Issue Facing While Using EmbeddedRocksDbCheckpointing FlinkVersion(1.15.0)

Hi， >From the task manager's log, We can find the following exception stack trace, seems it was the operating system related problem with rocksDB. 2022-06-04 14:45:53,878 INFO > org.apache.flink.runtime.executiongraph.ExecutionGraph [] - KEYED > PROCESS, Map -> Sink: Print to Std. Out

Re:Flink - aggregated output with status progression

Hi, what about use "Top1 + Agg" or "UDAF" for your scene. The main idea I think is that when the event changed from status A to C, Flink needs to send a `DELETE` data to downstream to delete the old data and send a new one to downstream again. And `TOP1` will keep the newest one with same

Re:??????????

Hi?? user-zh-unsubscr...@flink.apache.org ?? 2022-06-14 07:43:29??"?I ?? ?? ?I" <877144...@qq.com.INVALID> ?? > > > > > >---- >??: >

Re: Exception: class java.sql.Timestamp cannot be cast to class java.time.LocalDateTime | pyflink 1.15.0

Hi Mark, Could you share an example which could reproduce this issue? Regards, Dian On Thu, Jun 9, 2022 at 9:22 PM Márk Bartos wrote: > Hi, > > I'd like to ask for help regarding the java exception: > Caused by: java.lang.ClassCastException: class java.sql.Timestamp cannot > be cast to class

Re: How to implement unnest() as udtf using Python?

I second Martijn, UNNEST should be supported. Besides, regarding the above exception, could you share an example which could reproduce this issue? Regards, Dian On Mon, Jun 13, 2022 at 8:21 PM Martijn Visser wrote: > Hi John, > > You're mentioning that Flink doesn't support UNNEST, but it

Re: custom table source, how to support json?

Hi Ivan, Is your question how to parse the JSON string in PyFlink? If so, maybe you could take a look at this [1]. Regards, Dian [1] https://stackoverflow.com/questions/71820015/how-to-reference-nested-json-within-pyflink-sql-when-json-schema-varies On Fri, Jun 10, 2022 at 7:40 PM

??????????

---- ??: "user-zh"

Re: Kafka Consumer commit error

Hi Christian, thanks for the reply. We use AT_LEAST_ONCE delivery semantics in this > application. Do you think this might still be related? No, in that case, Kafka transactions are not used, so it should not be relevant. Best, Alexander Fedulov On Mon, Jun 13, 2022 at 3:48 PM Christian

Flink - aggregated output with status progression

Hello, I have a stream of events that are coming over Kafka. Each event progresses through a series of statuses. I want to display aggregated output of how many events are in a particular status. If suppose an event has progressed from status A to Status C then that event needs to be only counted

Flink running same task on different Task Manager

Hi, I have one flink job which has two tasks Task1- Source some static data over https and keep it in memory, this keeps refreshing it every 1 hour Task2- Process some real-time events from Kafka and uses static data to validate something and transform, then forward to other Kafka topic. so far,

Re:Re: Flink k8s HA 手动删除作业deployment导致的异常

感谢两位大大回复！ 在 2022-06-13 10:09:39，"Yang Wang" 写道： >Zhanghao的回答已经非常全面了，我再补充小点，删除Deployment保留HA ConfigMap是预期内的行为，文档里面有说明[1] >之所以这样设计有两点原因： >(1.) 任务可能会被重启，但使用相同的cluster-id，并且希望从之前的checkpoint恢复 >(2.) 单纯的删除ConfigMap会导致存储在DFS(e.g. HDFS、S3、OSS)上面的HA数据泄露 > >[1].

Re: Apache flink doesn't work with avro kafka topic with multiple event types

Thanks, I'll check it out. On Mon, Jun 13, 2022 at 2:40 AM Qingsheng Ren wrote: > Hi Sucheth, > > If you are referring to Table / SQL API, I'm afraid it doesn't support > schema evolution or different types from one Kafka table. An > alternative way is to consume the topic with raw format [1]

Re: Kafka Consumer commit error

Hi Alexander, thanks for the reply. We use AT_LEAST_ONCE delivery semantics in this application. Do you think this might still be related? Best regards, Christian Von: Alexander Fedulov Datum: Montag, 13. Juni 2022 um 13:06 An: "user@flink.apache.org" Cc: Christian Lorenz Betreff: Re:

Re: NegativeArraySizeException trying to take a savepoint

Hi Mike, It would be worthwhile to check if this still occurs in Flink 1.14, since Flink bumped to a newer version of RocksDB in that version. Is that a possibility for you? Best regards, Martijn Op ma 13 jun. 2022 om 15:21 schreef Mike Barborak : > When trying to savepoint our job, we are

Re: Kafka Consumer commit error

Hi Martijn, thanks for replying. I would also expect the behavior you describe below. AFAICT it was also like this with Flink 1.14. I am aware that Flink is using checkpointing for fault tolerance, but for example the Kafka offsets are part of our monitoring and this will lead to alerts. Other

NegativeArraySizeException trying to take a savepoint

When trying to savepoint our job, we are getting the stack trace below. Is there a way to know more about this failure? Like which function in the job graph is associated with the problematic state and which key (assuming it is keyed state)? Or is there a fix for this exception? The only

Re:Re: Re: Flink 使用interval join数据丢失疑问

Hi, 1、理论上来说inner join关联的数据量应该比interval join更大吧。关于左右两边流速度不一致的情况，理论上应该问题不大，因为需要等到两边的watermark都到齐之后才会触发状态里过期数据的清除。 2、inner join没有水印的情况下，就是到了就发，完全根据这条数据进入这个算子的时间来算，也就是“处理时间”。默认数据是不会过期的，会存全量数据。如果定义了ttl，得看join两侧的表的pk和join

Re:unsubscribe

Hi, you can send any contents to user-unsubscr...@flink.apache.org to unsubscribe. 在 2022-06-12 11:41:27，"chenshu...@foxmail.com" 写道： unsubscribe 退订 chenshu...@foxmail.com

Re: How to implement unnest() as udtf using Python?

You’re a legend, thank you so much, I was looking on the internal functions docs page, not that one! John Sent from my iPhone On 13 Jun 2022, at 13:21, Martijn Visser wrote:  Hi John, You're mentioning that Flink doesn't support UNNEST, but it does [1]. Would this work for you? Best

Re: exception while activating calculateThroughp

Hi sigalit, It's a known bug that has been fixed in Flink 1.15.0. See [1] and [2] for details. [1] https://issues.apache.org/jira/browse/FLINK-27712 [2] https://issues.apache.org/jira/browse/FLINK-25454 Best, Lijie Sigalit Eliazov 于2022年6月13日周一 20:17写道： > Hi all > > > We are using the flink

Re: How to implement unnest() as udtf using Python?

Hi John, You're mentioning that Flink doesn't support UNNEST, but it does [1]. Would this work for you? Best regards, Martijn [1] https://nightlies.apache.org/flink/flink-docs-release-1.15/docs/dev/table/sql/queries/joins/#array-expansion Op ma 13 jun. 2022 om 13:55 schreef John Tipper : >

exception while activating calculateThroughp

Hi all We are using the flink k8s operator latest version with flink 1.14 in order to deploy our pipelines in application mode (one job per cluster, one job manager + one task manager) Once in a few minutes I receive the following error in the job manager and all the tasks are being restarted.

How to implement unnest() as udtf using Python?

Hi all, Flink doesn’t support the unnest() function, which takes an array and creates a row for each element in the array. I have column containing an array of map and I’d like to implement my own unnest. I try this as an initial do-nothing implementation: @udtf(result_types=Datatypes.MAP(

Re: Flink Shaded dependencies and extending Flink APIs

Can we find a more robust way to support this? Both flink-shaded, any relocation pattern and JsonRowDataSerializationSchema are Flink internals that users shouldn't use/rely on. On 13/06/2022 12:26, Qingsheng Ren wrote: Hi Andrew, This is indeed a tricky case since Flink doesn't provide

Re: Kafka Consumer commit error

Hi Christian, you should check if the exceptions that you see after the broker is back from maintenance are the same as the ones you posted here. If you are using EXACTLY_ONCE, it could be that the later errors are caused by Kafka purging transactions that Flink attempts to commit [1]. Best,

Re: Flink Shaded dependencies and extending Flink APIs

Hi Andrew, This is indeed a tricky case since Flink doesn't provide non-shaded JAR for flink-json. One hacky solution in my mind is like: 1. Create a module let's say "wikimedia-event-utilities-shaded" that relocates Jackson in the same way and uses the same Jackson version as

Re: Kafka Consumer commit error

Hi Christian, I would expect that after the broker comes back up and recovers completely, these error messages would disappear automagically. It should not require a restart (only time). Flink doesn't rely on Kafka's checkpointing mechanism for fault tolerance. Best regards, Martijn Op wo 8

Re: Flink SQL JDBC connector for Postgres can't handle UUID datatype

Hi Aaron, There's currently no support in Flink indeed to insert an UUID data type into Postgres. The Jira ticket you've included [1] is indeed the same issue. It's just that the solution is most likely not to map it as a RAW type, but use a STRING type instead. Is this something where you might

Fwd: Apache flink doesn't work with avro kafka topic with multiple event types

Hi Sucheth, If you are referring to Table / SQL API, I'm afraid it doesn't support schema evolution or different types from one Kafka table. An alternative way is to consume the topic with raw format [1] and do deserialization with a UDTF. If you are using the DataStream API, you can implement

Re: Unsubscribe

Hi, In order to unsubscribe, please send an email to user-unsubscr...@flink.apache.org Best regards, Martijn Op vr 10 jun. 2022 om 17:23 schreef : > Unsubscribe >

Re: Unsubscribe

Hi, In order to unsubscribe, please send an email to user-unsubscr...@flink.apache.org Best regards, Martijn Op za 11 jun. 2022 om 19:46 schreef tarun joshi <1985.ta...@gmail.com>: > Unsubscribe >

Re: Metrics for FileSource

Hi, I believe this is a case where for the FileSystem (both Source and Sink) the metrics that are defined as part of FLIP-33 [1] have not been implemented yet. I've created a ticket for that [2]. Best regards, Martijn [1]