You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Piotr Nowojski (JIRA)" <ji...@apache.org> on 2019/07/10 08:05:00 UTC

[jira] [Closed] (FLINK-12816) set maxParallelism to 1 when DataStreamSource is not isParallel

     [ https://issues.apache.org/jira/browse/FLINK-12816?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Piotr Nowojski closed FLINK-12816.
----------------------------------
    Resolution: Won't Fix

Copied from the [github's pull request|[https://github.com/apache/flink/pull/8708#issuecomment-505777251]:]

{noformat}
However this change brakes a savepoint restoration compatibility. As far as I know, there is some check somewhere, that ensures (for good reasons) that maxParallelism of the current job and of the savepoint from which we are restoring is the same.

I'm pretty sure that we can not fix this that simply as this PR wants to do it. It also brings me a question, whether this change is worth the effort. Could you explain what exactly are you trying to improve/solve here?
{noformat}

For now, I'm closing this ticket because of the above mentioned issue. Please feel free to re-open it if you [~julien1987] (or anyone else) will come up with good motivating example and a proposed solution for the savepoint compatibility issue.

> set maxParallelism to 1 when DataStreamSource is not isParallel
> ---------------------------------------------------------------
>
>                 Key: FLINK-12816
>                 URL: https://issues.apache.org/jira/browse/FLINK-12816
>             Project: Flink
>          Issue Type: Task
>          Components: API / DataStream
>    Affects Versions: 1.8.0
>            Reporter: XuPingyong
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Now when DataStreamSource is not parallel, only parallelism is set to 1, maxParallelism should be also set to 1.
>  
> As when SingleOutputStreamOperator is not  parallel, it set not only parallelism but also maxParallelism to 1



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)