flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-1977) Rework Stream Operators to always be push based
Date Thu, 07 May 2015 13:45:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-1977?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14532638#comment-14532638
] 

ASF GitHub Bot commented on FLINK-1977:
---------------------------------------

Github user rmetzger commented on a diff in the pull request:

    https://github.com/apache/flink/pull/659#discussion_r29852424
  
    --- Diff: flink-staging/flink-streaming/flink-streaming-core/src/main/java/org/apache/flink/streaming/api/functions/source/FileSourceFunction.java
---
    @@ -46,39 +51,25 @@ public FileSourceFunction(InputFormat<String, ?> format, TypeInformation<String>
     	}
     
     	@Override
    +	@SuppressWarnings("unchecked")
     	public void open(Configuration parameters) throws Exception {
     		StreamingRuntimeContext context = (StreamingRuntimeContext) getRuntimeContext();
     		this.provider = context.getInputSplitProvider();
     		inputFormat.configure(context.getTaskStubParameters());
    +		serializer = typeInfo.createSerializer(getRuntimeContext().getExecutionConfig());
    --- End diff --
    
    As far as I understand the streaming code, the execution config at runtime is always invalid.
    
    When running a streaming job, you'll get plenty of error messages like this:
    ```
    11:26:25,600 WARN  org.apache.flink.runtime.jobgraph.tasks.AbstractInvokable     - Environment
did not contain an ExecutionConfig - using a default config.
    ```
    I was told that Streaming doesn't need the EC at runtime because all serializers are created
at the client and then send over serialized.


> Rework Stream Operators to always be push based
> -----------------------------------------------
>
>                 Key: FLINK-1977
>                 URL: https://issues.apache.org/jira/browse/FLINK-1977
>             Project: Flink
>          Issue Type: Improvement
>            Reporter: Aljoscha Krettek
>            Assignee: Aljoscha Krettek
>
> This is a result of the discussion on the mailing list. This is an excerpt from the mailing
list that gives the basic idea of the change:
> I propose to change all streaming operators to be push based, with a
> slightly improved interface: In addition to collect(), which I would
> call receiveElement() I would add receivePunctuation() and
> receiveBarrier(). The first operator in the chain would also get data
> from the outside invokable that reads from the input iterator and
> calls receiveElement() for the first operator in a chain.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message