[jira] [Created] (FLINK-7661) Add credit field in PartitionRequest message
zhijiang created FLINK-7661: --- Summary: Add credit field in PartitionRequest message Key: FLINK-7661 URL: https://issues.apache.org/jira/browse/FLINK-7661 Project: Flink Issue Type: Sub-task Components: Network Affects Versions: 1.4.0 Reporter: zhijiang Assignee: zhijiang Currently the {{PartitionRequest}} message contains {{ResultPartitionID}}|{{queueIndex}}|{{InputChannelID}} fields. We will add a new {credit} field indicating the initial credit of {{InputChannel}, and this info can be got from {{InputChannel}} directly after assigning exclusive buffers to it. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
Re: [DISCUSS] Dropping Scala 2.10
+1 @ Frederico: I think Aljoscha is right - Flink only executes Kafka client code, which is Scala independent from 0.9 on. Do you use Kafka 0.8 still? On Wed, Sep 20, 2017 at 10:00 PM, Aljoscha Krettek wrote: > Hi Federico, > > As far as I know, the Kafka client code has been rewritten in Java for > version 0.9, meaning there is no more Scala dependency in there. Only the > server (broker) code still contains Scala but it doesn't matter what Scala > version a client uses, if any. > > Best, > Aljoscha > > On 20. Sep 2017, at 14:32, Federico D'Ambrosio wrote: > > Hi, as far as I know some vendors like Hortonworks still use Kafka_2.10 as > part of their hadoop distribution. > Could the use of a different scala version cause issues with the Kafka > connector? I'm asking because we are using HDP 2.6 and we once already had > some issue with conflicting scala versions concerning Kafka (though, we > were using Storm, I still haven't tested the Flink connector in this > context). > > Regards, > Federico > > 2017-09-20 14:19 GMT+02:00 Ted Yu : > >> +1 >> >> Original message >> From: Hai Zhou >> Date: 9/20/17 12:44 AM (GMT-08:00) >> To: Aljoscha Krettek , dev@flink.apache.org, user < >> u...@flink.apache.org> >> Subject: Re: [DISCUSS] Dropping Scala 2.10 >> >> +1 >> >> > 在 2017年9月19日,17:56,Aljoscha Krettek 写道: >> > >> > Hi, >> > >> > Talking to some people I get the impression that Scala 2.10 is quite >> outdated by now. I would like to drop support for Scala 2.10 and my main >> motivation is that this would allow us to drop our custom Flakka build of >> Akka that we use because newer Akka versions only support Scala 2.11/2.12 >> and we need a backported feature. >> > >> > Are there any concerns about this? >> > >> > Best, >> > Aljoscha >> >> > > > -- > Federico D'Ambrosio > > >
[jira] [Created] (FLINK-7662) Remove unnecessary packaged licenses
Stephan Ewen created FLINK-7662: --- Summary: Remove unnecessary packaged licenses Key: FLINK-7662 URL: https://issues.apache.org/jira/browse/FLINK-7662 Project: Flink Issue Type: Improvement Components: Build System Reporter: Stephan Ewen Assignee: Stephan Ewen Fix For: 1.4.0 With the new shading approach, we no longer shade ASM into Flink artifacts, so we do not need to package the ASM license into those artifacts any more. Instead, a shaded ASM artifact already containing a packaged license is used in the distribution build. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
Re: [DISCUSS] Dropping Scala 2.10
Ok, thank you all for the clarification. @Stephan: I'm using Kafka 0.10, I guess the problem I had then was actually unrelated to specific Kafka version Federico D'Ambrosio Il 21 set 2017 16:30, "Stephan Ewen" ha scritto: > +1 > > @ Frederico: I think Aljoscha is right - Flink only executes Kafka client > code, which is Scala independent from 0.9 on. Do you use Kafka 0.8 still? > > On Wed, Sep 20, 2017 at 10:00 PM, Aljoscha Krettek > wrote: > >> Hi Federico, >> >> As far as I know, the Kafka client code has been rewritten in Java for >> version 0.9, meaning there is no more Scala dependency in there. Only the >> server (broker) code still contains Scala but it doesn't matter what Scala >> version a client uses, if any. >> >> Best, >> Aljoscha >> >> On 20. Sep 2017, at 14:32, Federico D'Ambrosio >> wrote: >> >> Hi, as far as I know some vendors like Hortonworks still use Kafka_2.10 >> as part of their hadoop distribution. >> Could the use of a different scala version cause issues with the Kafka >> connector? I'm asking because we are using HDP 2.6 and we once already had >> some issue with conflicting scala versions concerning Kafka (though, we >> were using Storm, I still haven't tested the Flink connector in this >> context). >> >> Regards, >> Federico >> >> 2017-09-20 14:19 GMT+02:00 Ted Yu : >> >>> +1 >>> >>> Original message >>> From: Hai Zhou >>> Date: 9/20/17 12:44 AM (GMT-08:00) >>> To: Aljoscha Krettek , dev@flink.apache.org, user < >>> u...@flink.apache.org> >>> Subject: Re: [DISCUSS] Dropping Scala 2.10 >>> >>> +1 >>> >>> > 在 2017年9月19日,17:56,Aljoscha Krettek 写道: >>> > >>> > Hi, >>> > >>> > Talking to some people I get the impression that Scala 2.10 is quite >>> outdated by now. I would like to drop support for Scala 2.10 and my main >>> motivation is that this would allow us to drop our custom Flakka build of >>> Akka that we use because newer Akka versions only support Scala 2.11/2.12 >>> and we need a backported feature. >>> > >>> > Are there any concerns about this? >>> > >>> > Best, >>> > Aljoscha >>> >>> >> >> >> -- >> Federico D'Ambrosio >> >> >> >
[jira] [Created] (FLINK-7663) Allow AbstractRestHandler to handle bad requests
Till Rohrmann created FLINK-7663: Summary: Allow AbstractRestHandler to handle bad requests Key: FLINK-7663 URL: https://issues.apache.org/jira/browse/FLINK-7663 Project: Flink Issue Type: Bug Components: REST Affects Versions: 1.4.0 Reporter: Till Rohrmann Assignee: Till Rohrmann The {{AbstractRestHandler}} parses the request and tries to generate a {{HandlerRequest}}. If this fails, then the server answers with an internal server error. Instead we should allow the {{AbstractRestHandler}} to be able to return a BAD_REQUEST status code. In order to do that, I would like to introduce a {{HandlerRequestException}} which can be thrown while creating the {{HandlerRequest}}. If this exception is thrown, then we return an error message with {{BAD_REQUEST}} {{HttpResponseStatus}}. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (FLINK-7664) Replace FlinkFutureException by CompletionException
Till Rohrmann created FLINK-7664: Summary: Replace FlinkFutureException by CompletionException Key: FLINK-7664 URL: https://issues.apache.org/jira/browse/FLINK-7664 Project: Flink Issue Type: Improvement Components: Core Affects Versions: 1.4.0 Reporter: Till Rohrmann Assignee: Till Rohrmann The {{FlinkFutureException}} was introduced to fail in a {{CompletableFuture}} callback. This can, however, better be done via the {{CompletionException}}. Therefore, we should remove the {{FlinkFutureException}} and replace it with the {{CompletionException}} instead. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (FLINK-7665) Use wait/notify in ContinuousFileReaderOperator
Ufuk Celebi created FLINK-7665: -- Summary: Use wait/notify in ContinuousFileReaderOperator Key: FLINK-7665 URL: https://issues.apache.org/jira/browse/FLINK-7665 Project: Flink Issue Type: Improvement Components: Streaming Connectors Affects Versions: 1.4.0 Reporter: Ufuk Celebi Priority: Minor {{ContinuousFileReaderOperator}} has the following loop to receive input splits: {code} synchronized (checkpointLock) { if (currentSplit == null) { currentSplit = this.pendingSplits.poll(); if (currentSplit == null) { if (this.shouldClose) { isRunning = false; } else { checkpointLock.wait(50); } continue; } } } {code} I think we can replace this with a {{wait()}} and {{notify()}} in {{addSplit}} and {{close}}. If there is a reason to keep the {{wait(50)}}, feel free to close this issue. :-) -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (FLINK-7666) ContinuousFileReaderOperator swallows chained watermarks
Ufuk Celebi created FLINK-7666: -- Summary: ContinuousFileReaderOperator swallows chained watermarks Key: FLINK-7666 URL: https://issues.apache.org/jira/browse/FLINK-7666 Project: Flink Issue Type: Improvement Components: Streaming Connectors Affects Versions: 1.3.2 Reporter: Ufuk Celebi I use event time and read from a (finite) file. I assign watermarks right after the {{ContinuousFileReaderOperator}} with parallelism 1. {code} env .readFile(new TextInputFormat(...), ...) .setParallelism(1) .assignTimestampsAndWatermarks(...) .setParallelism(1) .map()... {code} The watermarks I assign never progress through the pipeline. I can work around this by inserting a {{shuffle()}} after the file reader or starting a new chain at the assigner: {code} env .readFile(new TextInputFormat(...), ...) .setParallelism(1) .shuffle() .assignTimestampsAndWatermarks(...) .setParallelism(1) .map()... {code} -- This message was sent by Atlassian JIRA (v6.4.14#64029)