[ 
https://issues.apache.org/jira/browse/FLINK-28035?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Jane Chan updated FLINK-28035:
------------------------------
    Description: 
For an ordinary read-write job, the scan will check the numBuckets read from 
manifests against the current numBuckets, to avoid data corruption. See 
FLINK-27316.

 

However, this can be improved as follows.
 * If no new writes happen after changing the bucket number, the reads should 
not be blocked.
 * For rescale overwrite, we should support scan as the old bucket num, rescale 
and commit as the new bucket num.
 * The streaming job can be suspended and recovered from the rescaled data 
layout.

  was:
For an ordinary read-write job, the scan will check the numBuckets read from 
manifests against the current numBuckets, to avoid data corruption. See 
FLINK-27316.

 

However, this can be improved as follows.
 * If no new writes happen after changing the bucket number, the reads should 
not be blocked.
 * For rescale overwrite, we should support scan as the old bucket num, rescale 
and commit as the new bucket num.
 * The streaming job can be suspended, and recovered from the rescaled data 
layout.


> Support rescale overwrite
> -------------------------
>
>                 Key: FLINK-28035
>                 URL: https://issues.apache.org/jira/browse/FLINK-28035
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Table Store
>    Affects Versions: table-store-0.2.0
>            Reporter: Jane Chan
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: table-store-0.2.0
>
>
> For an ordinary read-write job, the scan will check the numBuckets read from 
> manifests against the current numBuckets, to avoid data corruption. See 
> FLINK-27316.
>  
> However, this can be improved as follows.
>  * If no new writes happen after changing the bucket number, the reads should 
> not be blocked.
>  * For rescale overwrite, we should support scan as the old bucket num, 
> rescale and commit as the new bucket num.
>  * The streaming job can be suspended and recovered from the rescaled data 
> layout.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

Reply via email to