Herbert Wang created FLINK-39585:
------------------------------------

             Summary: Log AdaptiveScheduler ExecutionGraph creation failures at 
ERROR
                 Key: FLINK-39585
                 URL: https://issues.apache.org/jira/browse/FLINK-39585
             Project: Flink
          Issue Type: Improvement
          Components: Runtime / Coordination
    Affects Versions: 2.2.1
            Reporter: Herbert Wang


In {{CreatingExecutionGraph.handleExecutionGraphCreation}}, when the 
{{ExecutionGraphWithVertexParallelism}} future completes exceptionally, the 
AdaptiveScheduler logs the root-cause throwable at INFO before transitioning 
the job to {{FAILED}}:

{code:java}
// flink-runtime .../scheduler/adaptive/CreatingExecutionGraph.java             
                                                                                
                                                      
  if (throwable != null) {                                                      
                                                                                
                                                        
      getLogger()                                                               
                                                                                
                                                        
              .info(
                      "Failed to go from {} to {} because the ExecutionGraph 
creation failed.",                                                              
                                                           
                      CreatingExecutionGraph.class.getSimpleName(),             
                                                                                
                                                        
                      Executing.class.getSimpleName(),                          
                                                                                
                                                        
                      throwable);                                               
                                                                                
                                                        
      ...                                                                       
                                                                                
                                                        
      context.goToFinished(context.getArchivedExecutionGraph(JobStatus.FAILED, 
throwable));
  }    
{code}
  
Proposed change: promote the single failure log call from INFO to ERROR.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to