[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17336915#comment-17336915 ] Flink Jira Bot commented on FLINK-2313: --- This issue was labeled "stale-major" 7 ago and has not received any updates so it is being deprioritized. If this ticket is actually Major, please raise the priority and ask a committer to assign you the issue or revive the public discussion. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: API / DataStream >Reporter: Stephan Ewen >Priority: Major > Labels: stale-major > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17328979#comment-17328979 ] Flink Jira Bot commented on FLINK-2313: --- This major issue is unassigned and itself and all of its Sub-Tasks have not been updated for 30 days. So, it has been labeled "stale-major". If this ticket is indeed "major", please either assign yourself or give an update. Afterwards, please remove the label. In 7 days the issue will be deprioritized. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: API / DataStream >Reporter: Stephan Ewen >Priority: Major > Labels: stale-major > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15575392#comment-15575392 ] Maximilian Michels commented on FLINK-2313: --- +1 this also popped up recently on the Beam side: https://issues.apache.org/jira/browse/BEAM-593 This is partially also duplicated in https://issues.apache.org/jira/browse/FLINK-4272 > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15574877#comment-15574877 ] David Anderson commented on FLINK-2313: --- +1 This would be super helpful when working with streams inside notebook environments (like Jupyter and Zeppelin). In the case of a local flink, could we still keep stdout connected to the notebook? > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15342845#comment-15342845 ] Gyula Fora commented on FLINK-2313: --- Btw, do you have any thoughts on how should this behave in case of job failures/restarts etc? This is also related to Robert's question regarding the job id. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15342789#comment-15342789 ] Robert Metzger commented on FLINK-2313: --- Another important feature here would be to make the job id, and in the yarn case also the yarn application id available. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14631264#comment-14631264 ] Maximilian Michels commented on FLINK-2313: --- +1 Why should this only be changed for the {{StreamExecutionEnvironment}}? This could also go into the batch {{ExecutionEnvironment}}. Ultimately, it would be nice to join the two. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14631076#comment-14631076 ] Matthias J. Sax commented on FLINK-2313: Cool. I like this. Hope to I find some time to work on it soon. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611753#comment-14611753 ] Stephan Ewen commented on FLINK-2313: - Since this is API breaking, let's get at least the minimal change into the next release. Minimal would be to return the object with the {{wairFor()}} method. I also created a doc to collect breaking changes to the streaming API: https://docs.google.com/document/d/1itrALuaj_mmBMFqfEI1pOt8m8R1BIsmVFS_uuabsJIE > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611750#comment-14611750 ] Aljoscha Krettek commented on FLINK-2313: - +1 this should be super nice > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611730#comment-14611730 ] Robert Metzger commented on FLINK-2313: --- +1 I can actually imagine more possible methods ... for example getting metrics from the running application or even operator state. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2313) Change Streaming Driver Execution Model
[ https://issues.apache.org/jira/browse/FLINK-2313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14611729#comment-14611729 ] Gyula Fora commented on FLINK-2313: --- I think this would be a very cool change. I also don't like the blocking execute calls, once it is removed we can nicely integrate a collect() method to the data streams which will return the iterators over the streams like we have now in the DataStreamUtils. > Change Streaming Driver Execution Model > --- > > Key: FLINK-2313 > URL: https://issues.apache.org/jira/browse/FLINK-2313 > Project: Flink > Issue Type: Improvement > Components: Streaming >Reporter: Stephan Ewen > > I would propose the following change to the Streaming Execution Environment: > When calling {{env.execute()}}, it does not block, but return an > {{ExecutionContext}} object. This execution context has various methods, that > allow you to control the execution: > - waitFor() -> Blocks until the job terminates, or the connection to the > JobManager is lost. > - stop() -> initiates a clean stop > - cancel() -> cancels the program > - setParallelism() -> Later: adjusts the parallelism of the program > If this is voted as desirable, I would make a design how to do this... -- This message was sent by Atlassian JIRA (v6.3.4#6332)