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-4810) Checkpoint Coordinator should fail ExecutionGraph after "n" unsuccessful checkpoints
Date Wed, 01 Mar 2017 03:48:45 GMT

    [ https://issues.apache.org/jira/browse/FLINK-4810?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15889445#comment-15889445
] 

ASF GitHub Bot commented on FLINK-4810:
---------------------------------------

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

    https://github.com/apache/flink/pull/3334#discussion_r103605271
  
    --- Diff: flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/CheckpointCoordinator.java
---
    @@ -537,12 +562,27 @@ else if (!props.forceCheckpoint()) {
     				if (!checkpoint.isDiscarded()) {
     					checkpoint.abortError(new Exception("Failed to trigger checkpoint"));
     				}
    +				if(numUnsuccessful > maxUnsuccessfulCheckpoints) {
    +					return failExecution(executions);
    +				}
     				return new CheckpointTriggerResult(CheckpointDeclineReason.EXCEPTION);
     			}
     
     		} // end trigger lock
     	}
     
    +	private CheckpointTriggerResult failExecution(Execution[] executions) {
    +		if (currentPeriodicTrigger != null) {
    +			currentPeriodicTrigger.cancel();
    +			currentPeriodicTrigger = null;
    +		}
    +		for (Execution execution : executions) {
    +			// fail the graph
    +			execution.fail(new Throwable("The number of max unsuccessful checkpoints attempts
exhausted"));
    --- End diff --
    
    I think it's not good here to fail the executions one by one. We should call `ExecutionGraph#fail`
to fail the execution graph.


> Checkpoint Coordinator should fail ExecutionGraph after "n" unsuccessful checkpoints
> ------------------------------------------------------------------------------------
>
>                 Key: FLINK-4810
>                 URL: https://issues.apache.org/jira/browse/FLINK-4810
>             Project: Flink
>          Issue Type: Sub-task
>          Components: State Backends, Checkpointing
>            Reporter: Stephan Ewen
>
> The Checkpoint coordinator should track the number of consecutive unsuccessful checkpoints.
> If more than {{n}} (configured value) checkpoints fail in a row, it should call {{fail()}}
on the execution graph to trigger a recovery.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message