[
https://issues.apache.org/jira/browse/FLINK-32859?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rui Fan updated FLINK-32859:
----------------------------
Description:
Currently, most of states of adaptive scheduler extends
StateWithExecutionGraph, and it can reduce a lots of repeated code.
We can define the State{color:red}Without{color}ExecutionGraph and the
corresponding Context as well, then the rest of states can extend the
State{color:red}Without{color}ExecutionGraph, such as: Created,
WaitingForResources, CreatingExecutionGraph.
This improvement can reduce a lot of code for these states, such as: cancel(),
suspend(), getJob(), handleGlobalFailure() and getLogger().
These methods of Created, WaitingForResources and CreatingExecutionGraph are
same.
was:
Currently, most of states of adaptive scheduler are extends
StateWithExecutionGraph, and it can reduce a lots of repeated code.
We can define the State{color:red}Without{color}ExecutionGraph and the
corresponding Context, then the rest of states can extend the
State{color:red}Without{color}ExecutionGraph, such as: Created,
WaitingForResources, CreatingExecutionGraph.
This improvement can reduce a lot of code for these states, such as: cancel(),
suspend(), getJob(), handleGlobalFailure() and getLogger().
These methods of Created, WaitingForResources and CreatingExecutionGraph are
same.
> Improve the state of adaptive sheduler with StateWithoutExecutionGraph
> ----------------------------------------------------------------------
>
> Key: FLINK-32859
> URL: https://issues.apache.org/jira/browse/FLINK-32859
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Coordination
> Reporter: Rui Fan
> Assignee: Rui Fan
> Priority: Major
>
> Currently, most of states of adaptive scheduler extends
> StateWithExecutionGraph, and it can reduce a lots of repeated code.
> We can define the State{color:red}Without{color}ExecutionGraph and the
> corresponding Context as well, then the rest of states can extend the
> State{color:red}Without{color}ExecutionGraph, such as: Created,
> WaitingForResources, CreatingExecutionGraph.
> This improvement can reduce a lot of code for these states, such as:
> cancel(), suspend(), getJob(), handleGlobalFailure() and getLogger().
> These methods of Created, WaitingForResources and CreatingExecutionGraph are
> same.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
