[jira] [Updated] (FLINK-35452) MySQL CDC is not tracking newly added tables when a failure happens after the snapshot phase

["Sharon Xie (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/FLINK-35452?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sharon Xie updated FLINK-35452:
-------------------------------
    Summary: MySQL CDC is not tracking newly added tables when a failure happens after the snapshot phase  (was: MySQL CDC not tracking newly added tables after snapshot)

> MySQL CDC is not tracking newly added tables when a failure happens after the snapshot phase
> --------------------------------------------------------------------------------------------
>
>                 Key: FLINK-35452
>                 URL: https://issues.apache.org/jira/browse/FLINK-35452
>             Project: Flink
>          Issue Type: Bug
>          Components: Flink CDC
>    Affects Versions: 2.0.0
>            Reporter: Sharon Xie
>            Priority: Major
>         Attachments: image-2024-05-25-20-24-40-460.png
>
>
> We've seen some cases where a MySQL CDC job is restarted with some newly added tables, and the newly added tables are not getting change events after the snapshot. 
> From the log, it seems that a failure happened right after the snapshot phase, and when the job is restored from the last checkpoint, the newly added tables are not in the state. It looks like whenever a failure happens in the snapshot phase could cause this to happen in general.
> Using Flink 1.15.3 with MySQL CDC 2.4.2
> Some logs and analysis
> !image-2024-05-25-20-24-40-460.png|width=1360,height=837!



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