flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aljoscha Krettek (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-4410) Split checkpoint times into synchronous and asynchronous part
Date Wed, 14 Sep 2016 08:24:20 GMT

    [ https://issues.apache.org/jira/browse/FLINK-4410?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15489788#comment-15489788
] 

Aljoscha Krettek commented on FLINK-4410:
-----------------------------------------

Hi,
there's actually three different durations that could be reported:

 - time from the checkpoint coordinator initiating a checkpoint to an operator acknowledging
that checkpoint
 - time that an operator spends in the synchronous part of the checkpoint
 - time that an operator spends in the asynchronous part of the checkpoint

About synchronous/asynchronous. For this you can look at {{StreamTask.performCheckpoint()}}.
At the end of the method a Thread is started that does the asynchronous work of the checkpoint
and the method returns. Thus, time until then would be the synchronous part and the time spend
in that thread would be the asynchronous part.

> Split checkpoint times into synchronous and asynchronous part
> -------------------------------------------------------------
>
>                 Key: FLINK-4410
>                 URL: https://issues.apache.org/jira/browse/FLINK-4410
>             Project: Flink
>          Issue Type: Improvement
>          Components: Webfrontend
>            Reporter: Ufuk Celebi
>            Priority: Minor
>
> Checkpoint statistics contain the duration of a checkpoint. We should split this time
into the synchronous and asynchronous part. This will give more insight into the inner workings
of the checkpointing mechanism and help users better understand what's going on.



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

Mime
View raw message